The Oxford Handbook of Moral Responsibility 9780190679309, 0190679301

Citation preview

T h e Ox f o r d H a n d b o o k o f

M OR A L R E SP ON SI B I L I T Y

The Oxford Handbook of

MORAL RESPONSIBILITY Edited by

DANA KAY NELKIN and

DERK PEREBOOM

1

3 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 certain other countries. Published in the United States of America by Oxford University Press 198 Madison Avenue, New York, NY 10016, United States of America. © Oxford University Press 2022 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 license, or under terms agreed with the appropriate reproduction rights organization. Inquiries 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. Library of Congress Control Number: 2021039940 cover art: Koi in the Rain by Lee Morais ISBN 978–​0–​19–​067930–9 DOI: 10.1093/oxfordhb/9780190679330.001.0001 1 3 5 7 9 8 6 4 2 Printed by Marquis, Canada

Contents

About the Contributors General Introduction Dana Kay Nelkin and Derk Pereboom

ix xvii

PA RT I   T H E OR I E S OF R E SP ON SI B I L I T Y 1. Instrumentalist Theories of moral Responsibility Manuel Vargas 2. Reasons-​Responsiveness, Frankfurt Examples, and the Free Will Ability Michael McKenna 3. Attributionist Theories of Moral Responsibility Matthew Talbert

3

27 53

PA RT I I   K I N D S OF R E SP ON SI B I L I T Y 4. Accountability, Answerability, and Attributability: On Different Kinds of Moral Responsibility Sofia Jeppsson

73

PA RT I I I   DI M E N SION S OF R E SP ON SI B I L I T Y 5. Responsibility for Acts and Omissions Randolph Clarke

91

6. Degrees of Responsibility D. Justin Coates

111

7. Group Responsibility Christian List

131

vi   Contents

PA RT I V   DE T E R M I N I SM A N D T H E A B I L I T Y TO D O OT H E RW I SE 8. Moral Responsibility, Alternative Possibilities, and Frankfurt Examples Derk Pereboom 9. Manipulation Arguments Against Compatibilism Derk Pereboom and Michael McKenna

157 179

PA RT V   SK E P T IC I SM 10. Illusionism Saul Smilansky

203

11. Free Will Skepticism and Criminal Justice: The Public Health-​Quarantine Model Gregg D. Caruso 12. Metaskepticism Tamler Sommers

222 247

PA RT V I   B L A M E 13. Blame and Holding Responsible Angela M. Smith

269

14. Responsibility and Reactive Attitudes R. Jay Wallace

287

15. Response-​Dependence Theories of Responsibility David Shoemaker

304

PA RT V I I   R E SP ON SI B I L I T Y, K N OW L E D G E , A N D C AU S AT ION 16. Ethics Is Hard! What Follows? On Moral Ignorance and Blame Elizabeth Harman

327

17. Responsibility and Causation Carolina Sartorio

348

Contents   vii

PA RT V I I I   R E SP ON SI B I L I T Y, L AW, A N D J U S T IC E 18. Responsibility, Punishment, and Predominant Retributivism David O. Brink

365

19. Legal Responsibility: Psychopathy, a Case Study Elizabeth Shaw

388

20. Responsibility and Distributive Justice Richard Arneson

412

PA RT I X   R E SP ON SI B I L I T Y, N E U RO S C I E N C E , A N D P SYC HOL O G Y 21. Moral Responsibility and Neuroscience Alfred R. Mele

433

22. Responsibility and Consciousness Matt King and Peter Carruthers

448

23. Responsibility and Situationism Brandon Warmke

468

24. Experimental Philosophy and Moral Responsibility Gunnar Björnsson

494

PA RT X   R E SP ON SI B I L I T Y, R E L AT ION SH I P S , AND MEANING IN LIFE 25. Moral Responsibility and Existential Attitudes Paul Russell

519

26. Relationships and Responsibility Dana Kay Nelkin

542

27. Responsibility, Personal Relationships, and the Significance of the Reactive Attitudes Seth Shabo 28. Forgiveness Per-​Erik Milam

566 590

viii   Contents

29. Reconciliation and the End of Responsibility Linda Radzik

609

30. Responsibility and Religion Daniel Speak

625

PA RT X I   C A SE ST U DI E S 31. Moral Responsibility in the Context of Addiction Doug McConnell 32. Moral Responsibility for Implicit Bias and the Impact of Social Categorization Maureen Sie

647

668

33. Skepticism about Evil: Atrocity and the Limits of Responsibility Dominic Murphy and John M. Doris

697

Index

727

About the Contributors

Richard Arneson (PhD UC Berkeley) is Distinguished Professor and holds the Valtz Family Chair in Philosophy at the University of California, San Diego. He works mainly in normative political and moral philosophy. Much of his writing is on theories of justice. He has been concerned to articulate a broadly egalitarian view on distributive justice joined with a sensible view on personal responsibility, a variant of so-​called “luck egalitarianism” that was then resuscitated as “desert-​catering prioritarianism.” He also has written on consequentialism versus nonconsequentialism and on some applied ethics topics. His current projects include writing a short book on prioritarianism and another short volume (co-​authored with Jason Brennan), Debating Capitalism. Gunnar Björnsson is Professor of Practical Philosophy at Stockholm University. He works on issues in metaethics, moral responsibility, collective moral agency, and philosophy of language. Recent work has concerned collective obligations, corporate moral agency, the explanatory and epistemic conditions on moral responsibility, and the nature of moral judgment and moral disagreement; it has appeared in, among other places, Mind, Noûs, Ethics, and Philosophy and Phenomenological Research. Björnsson founded and coordinated what is now the Lund Gothenburg Responsibility Project and has been the principal investigator of projects on individual and collective responsibility and an interdisciplinary project on moral motivation. David O. Brink (Cornell PhD, 1984) is Distinguished Professor of Philosophy at the University of California, San Diego. His research interests are in ethical theory, history of ethics, moral psychology, political philosophy, and jurisprudence. He is the author of Moral Realism and the Foundations of Ethics (Cambridge 1989), Perfectionism and the Common Good: Themes in the Philosophy of T.H. Green (Oxford 2003), Mill’s Progressive Principles (Oxford 2013), and Fair Opportunity and Responsibility (Oxford 2021); the editor of T.H. Green’s Prolegomena to Ethics (Oxford 2003); and a co-​editor of Virtue, Happiness, and Knowledge: Themes from the Work of Gail Fine and Terence Irwin (Oxford 2018). Peter Carruthers (DPhil, Oxford) is Distinguished University Professor of Philosophy at the University of Maryland. He works on various topics in philosophy of mind and cognitive science. Recent books include Human and Animal Minds: The Consciousness Questions Laid to Rest (Oxford University Press 2019) and The Centered Mind: What the Science of Working Memory Shows Us About the Nature of Human Thought (Oxford University Press 2015).

x   About the Contributors Gregg D. Caruso is Professor of Philosophy at SUNY Corning and Honorary Professor of Philosophy at Macquarie University. He is also the co-​director of the Justice Without Retribution Network at the University of Aberdeen School of Law. His books include Rejecting Retributivism: Free Will, Punishment, and Criminal Justice (Cambridge University Press 2021), Just Deserts: Debating Free Will (Polity 2021, co-​authored with Daniel Dennett), and Free Will and Consciousness: A Determinist Account of the Illusion of Free Will (Lexington Books 2012). His areas of research include free will, agency, and responsibility (both moral and legal), as well as philosophy of mind, cognitive science, neuroethics, moral psychology, criminal law, punishment, and public policy. Randolph Clarke (PhD Princeton) is Professor of Philosophy at Florida State University. His research interests include agency, free will, responsibility, dispositions, and causation. He is the author of Libertarian Accounts of Free Will (Oxford University Press 2003) and Omissions: Agency, Metaphysics, and Responsibility (Oxford University Press 2014) and co-​editor of The Nature of Moral Responsibility: New Essays (Oxford University Press 2015). D. Justin Coates is Associate Professor of Philosophy at the University of Houston. His areas of research include philosophy of action, moral psychology, and ethics. He is the co-​editor of Blame: Its Nature and Norms (Oxford University Press 2012) and has authored articles on topics including moral responsibility, ambivalence, blame, love, and practical reasons. John M. Doris is Peter L. Dyson Professor of Ethics in Organizations and Life, Charles H. Dyson School of Applied Economics and Management; and Professor, Sage School of Philosophy, Cornell University. He works at the intersection of cognitive science, moral psychology, and philosophical ethics. He is a winner of the Society for Philosophy and Psychology’s Stanton Prize for excellence in interdisciplinary research, among other honors. He is author of Lack of Character: Personality and Moral Behavior (Cambridge University Press 2002); Talking to Our Selves: Reflection, Ignorance, and Agency (Oxford University Press 2015); and Character Trouble: Undisciplined Essays on Moral Agency and Personality (Oxford University Press 2021). With his colleagues in the Moral Psychology Research Group, he wrote and edited The Moral Psychology Handbook (Oxford University Press 2010); and with Manuel Vargas, he edited The Oxford Handbook of Moral Psychology (Oxford University Press 2017). Elizabeth Harman is Laurance S. Rockefeller Professor of Philosophy and Human Values at Princeton University; she is also Director of Early-​Career Research at the University Center for Human Values. Her publications include “Creation Ethics” (Philosophy and Public Affairs), “ ‘I’ll Be Glad I Did It’ Reasoning and the Significance of Future Desires” (Philosophical Perspectives), “The Irrelevance of Moral Uncertainty” (Oxford Studies in Metaethics), and “Morally Permissible Moral Mistakes” (Ethics). She is co-​editor of Norton Introduction to Philosophy, Second Edition (2018), and Norton Introduction to Ethics (forthcoming in 2021).

About the Contributors    xi Sofia Jeppsson (PhD Stockholm University) is Associate Professor of Philosophy at Umeå University, Umeå, Sweden. She has published numerous articles on animal ethics, criminal justice, free will, and moral responsibility. Her current research concerns the intersection of moral responsibility, respect, and the philosophy of psychiatry. Matt King is Associate Professor of Philosophy at the University of Alabama at Birmingham, where he conducts research in ethics, moral psychology, and the law. He is the author of numerous articles on topics in responsibility and agency theory, including desert, manipulation, mental illness, negligence, the reasonable person standard, and the standing to blame. He is also co-editor of Agency in Mental Disorder: Philosophical Perspectives (Oxford University Press forthcoming). Christian List is Professor of Philosophy and Decision Theory at LMU Munich, a Visiting Professor of Philosophy at the London School of Economics, and a Fellow of the British Academy. He works at the intersection of philosophy, economics, and political science, with a particular focus on individual and collective decision-​making and the nature of intentional agency. In recent years, a growing part of his work has addressed metaphysical questions, especially about free will, causation, and emergent phenomena. He is the author of Group Agency: The Possibility, Design, and Status of Corporate Agents (with Philip Pettit, Oxford University Press 2011) and Why Free Will is Real (Harvard University Press 2019). Doug McConnell (PhD Macquarie University) is a Research Fellow at the Uehiro Centre for Practical Ethics at the University of Oxford. His areas of research include bioethics, philosophical psychology, and philosophy of psychiatry. He has published articles on a broad range of topics including the role of self-​narration in recovery from addiction, conscientious objection in healthcare, and the biopsychosocial model of psychiatry. Michael McKenna (PhD Virginia) is Professor of Philosophy at the University of Arizona. His areas of research are mostly devoted to free will and moral responsibility, but also include issues in moral psychology, action theory, ethics, and metaphysics. He is the author of Conversation and Responsibility (Oxford University Press 2011), the co-​author with Derk Pereboom of Free Will: A Contemporary Introduction (Routledge 2014), and has also written numerous articles, most of which would impress you if you were to read them. He is also co-​editor of “Moral Responsibility and Alternative Possibilities,” “Free Will and Reactive Attitudes,” and “The Nature of Moral Responsibility.” He has a boundless lust for life, and he often drinks to excess. Alfred R. Mele is the William H. and Lucyle T. Werkmeister Professor of Philosophy at Florida State University. He is the author of 12 books and over 200 articles and an editor of 7 books. He is the past director of two multimillion dollar, interdisciplinary projects: the Big Questions in Free Will project (2010–​2013) and the Philosophy and Science of Self-​Control project (2014–​2017).

xii   About the Contributors Per-​Erik Milam is a Researcher in Practical Philosophy at the University of Gothenburg. He writes on agency and responsibility, focusing especially on practices like blame, apology, and forgiveness that individuals use to address and resolve moral conflicts. He is currently working on a project, funded by the Swedish Research Council, about the limits of forgiveness. Dominic Murphy is Professor of History and Philosophy of Science at the University of Sydney. He is the author of Psychiatry in the Scientific image (MIT Press, 2006) and joint editor of Stich and His Critics (Blackwell, 2008). He has published articles and chapters in philosophy of psychiatry, philosophy of mind, and moral psychology. Dana Kay Nelkin (PhD UCLA) is Professor of Philosophy at the University of California, San Diego, and an Affiliate Professor at the University of San Diego School of Law. Her areas of research include moral psychology, ethics, bioethics, and philosophy of law. She is the author of Making Sense of Freedom and Responsibility (Oxford University Press 2011), a co-​editor of the Ethics and Law of Omissions (Oxford University Press 2017), and a number of articles on a variety of topics, including self-​deception, friendship, the lottery paradox, psychopathy, forgiveness, and praise and blame. Her work in moral psychology includes participation in an interdisciplinary research collaboration of philosophers and psychologists, The Moral Judgements Project. Derk Pereboom (PhD UCLA) is the Susan Linn Sage Professor in the Philosophy Department at Cornell University and Senior Associate Dean for Arts and Humanities in Cornell’s College of Arts and Sciences. His areas of research include free will and moral responsibility, philosophy of mind, early modern philosophy, especially Kant, and philosophy of religion. He is the author of Living without Free Will (Cambridge University Press 2001), Consciousness and the Prospects of Physicalism (Oxford University Press 2011), Free Will, Agency, and Meaning in Life (Oxford University Press 2014), Wrongdoing and the Moral Emotions (Oxford University Press 2021), and he is co-​ author with Michael McKenna of Free Will: A Contemporary Introduction (Routledge 2016). He has published articles on free will and moral responsibility, consciousness and physicalism, nonreductive materialism, divine providence, the problem of evil, and on Kant’s metaphysics and epistemology. Linda Radzik (PhD Arizona) is Professor of Philosophy at Texas A&M University. Her work centers on the issues that arise in the aftermath of wrongdoing. She is the author of The Ethics of Social Punishment: The Enforcement of Morality in Everyday Life (Cambridge University Press, forthcoming), Making Amends: Atonement in Morality, Law and Politics (Oxford University Press 2009), as well as articles in moral, political, and legal philosophy. Paul Russell is Professor of Philosophy at Lund University (Sweden) the University of British Columbia (Canada). At Lund he is also the Director of the Lund | Gothenburg Responsibility project. Among his books are Freedom and Moral Sentiment: Hume’s Way

About the Contributors    xiii of Naturalizing Responsibility (Oxford University Press 1995) and The Limits of Free Will: Selected Essays (Oxford University Press 2017). Carolina Sartorio (PhD MIT, 2003) is Professor of Philosophy at the University of Arizona. She works at the intersection of metaphysics, the philosophy of action, and moral theory. She is the author of Causation and Free Will (Oxford University Press 2016), and of articles on causation, agency, free will, and moral responsibility, among other topics. Seth Shabo (PhD Syracuse University) is Associate Professor of Philosophy at the University of Delaware. His main areas of research are free will and moral responsibility, including luck, desert, and ethics and moral psychology of blame. He is the author of a number of articles on these topics. Elizabeth Shaw (PhD Edinburgh University, LLM Aberdeen University) is a Lecturer in the School of Law at Aberdeen University. Her research interests are interdisciplinary, involving criminal law, philosophy and neuroscience. She is a co-​editor of Free Will Skepticism, Law and Society: Challenging Retributive Justice (Cambridge University Press 2019). David Shoemaker is currently Professor in the Department of Philosophy & Murphy Institute of Tulane University, but as of fall 2021 he will be a Professor in the Department of Philosophy at Cornell University. He has done research primarily on agency and responsibility, personal identity and ethics, moral psychology, and social/​political philosophy, publishing many articles on these and other topics. He is the author of Responsibility from the Margins (Oxford University Press 2015), the editor of the general series Oxford Studies in Agency and Responsibility, an associate editor of Ethics, and a co-​ founder and co-​editor of the long-​running ethics blog PEA Soup. Maureen Sie is Full Professor Philosophy of Moral Agency at Tilburg School of Humanities and Digital Science and director of the Tilburg Center for Moral Philosophy, Epistemology and Philosophy of Science (TiLPS) at Tilburg University, the Netherlands. Her areas of research include moral psychology, philosophy of action, and (meta)ethics with a special attention to the developments in the behavioral, cognitive, and neurosciences. She is the author of a number of articles on moral responsibility and free will defending a pragmatic sentimentalist view on our moral practices. In the past decade she also published on hypocrisy, implicit bias, the traffic participation model of human agency, and love as the source of morality. Saul Smilansky (DPhil, Oxford) is a Professor at the Department of Philosophy, University of Haifa, Israel. He works primarily on normative and applied ethics, the free will problem, and the meaning of life. His particular topics of interest include moral paradoxes and other forms of “crazy ethics,” moral theory, justice, the role of self-​deception and illusion in our lives, population ethics, punishment, egalitarianism, moral complaint and hypocrisy, gratitude, and the significance of contribution. He is

xiv   About the Contributors the author of Free Will and Illusion (Oxford University Press 2000), 10 Moral Paradoxes (Blackwell 2007), and over 90 papers in philosophical journals and edited collections. Angela M. Smith (PhD Harvard) is Professor of Philosophy at Washington and Lee University. Her research interests concern the connections between morality, moral agency, and moral responsibility. She is co-​editor of the Oxford University Press book The Nature of Moral Responsibility (2015), and has published numerous articles on topics such as responsibility for attitudes, unconscious omissions, the nature of blame and forgiveness, and the value of tolerance. Tamler Sommers is Professor of Philosophy at the University of Houston. His research focuses on issues relating to freedom, responsibility, moral psychology, and the philosophy of punishment. He is the author of several books including Why Honor Matters (Basic Books 2018) and Relative Justice (Princeton University Press 2012), and he is the co-​host of the Very Bad Wizards podcast. Daniel Speak (PhD UC Riverside) is Professor of Philosophy at Loyola Marymount University in Los Angeles, California. His areas of research include the metaphysics, ethics, and epistemology of free will; and topics in the philosophy of religion. He is the author of The Problem of Evil (Polity Press 2014) and co-​editor with Kevin Timpe of Free Will and Theism (Oxford University Press 2016 ). His articles and chapters have appeared in The Oxford Handbook of Free Will, The Philosophical Quarterly, Faith and Philosophy, Religious Studies, Res Philosophica, The Routledge Companion to Free Will, and elsewhere. Matthew Talbert (PhD UC Riverside) is Associate Professor of Philosophy at West Virginia University and a Senior Researcher with the Lund-​Gothenburg Responsibility Project at Lund University. His research focuses on moral philosophy, moral psychology, and the philosophy of agency. He is the author of Moral Responsibility: An Introduction (Polity 2016), co-​author of War Crimes: Causes, Excuses, and Blame, (Oxford University Press 2019) and the author of a number of articles on the psychological, epistemic, social, and metaphysical conditions on moral responsibility and blameworthiness. Manuel Vargas (PhD Stanford) is Professor of Philosophy and affiliated faculty in Latin American Studies and Chicanx & Latinx Studies at the University of California, San Diego. His areas of research include moral psychology, ethics, the history of Mexican philosophy, and Latinx philosophy. He is the author of Building Better Beings: A Theory of Moral Responsibility, (Oxford University Press 2013) co-​author of Four Views on Free Will, (Wiley-Blackwell 2007) and co-​editor of The Oxford Handbook of Moral Psychology and Rational and Social Agency: Essays on the Philosophy of Michael Bratman. (Oxford University Press 2014) R. Jay Wallace (PhD Princeton) is Professor of Philosophy at the University of California, Berkeley, where he holds the Judy Chandler Webb Distinguished Chair. His research and teaching interests extend to all parts of moral philosophy. Recently he has written on promising, normativity, constructivism, resentment, hypocrisy, love, regret,

About the Contributors    xv affirmation, obligation, anger, and the relational structure of the moral domain. He is the author of numerous papers on these and other topics, and four books: Responsibility and the Moral Sentiments (Harvard University Press, 1994), Normativity and the Will (Oxford University Press 2006), The View from Here (Oxford University Press 2013), and The Moral Nexus (Princeton University Press 2019). Brandon Warmke (PhD Arizona) is Assistant Professor of Philosophy at Bowling Green State University. He works in ethics, social philosophy, moral psychology, and political philosophy. He is the author of several essays on public discourse, moral responsibility, and forgiveness. With Justin Tosi, he is the author of Grandstanding: The Use and Abuse of Moral Talk (Oxford University Press 2020). His work has been featured in The Atlantic, HuffPost, Scientific American, The Guardian, Slate, and Vox.

General Introduction

For a significant period until about two decades ago, moral responsibility was often seen as a kind of addendum to the free will literature rather than as a topic of its own. An interesting illustration is provided by Peter Strawson’s seminal 1962 essay, “Freedom and Resentment,” which laid the foundation for much of the current discussion of moral responsibility and our surrounding practices of praise and blame. It gained a wide readership that continues to this day, partly as a result of its inclusion in Gary Watson’s influential Oxford Readings in Free Will. Note that the title of this essay did not have the word “responsibility.” At the time, moral responsibility was often framed as an issue at stake in the free will debate, which was largely focused on the controversy concerning the compatibility of free will and determinism. The topic of moral responsibility came into its own after an initial transitional period. Susan Wolf ’s (1990) Freedom Within Reason offered a theory of freedom that was itself essentially normative in character; and Gary Watson’s response, “Two Faces of Responsibility,” set the stage for much future exploration of just what the concepts of responsibility are that we care about. John Martin Fischer and Mark Ravizza’s Responsibility and Control (1998), in part developing out of a response to Harry Frankfurt’s work on both free will and moral responsibility, provided another turning point in presenting a view according to which agents could be morally responsible without having free will. Robert Kane’s The Significance of Free Will (1996) laid out in greater detail than was usual what he saw as at stake in the free will debate from the libertarian perspective, and this included moral responsibility and desert for accomplishments, among several other phenomena. Currently, many aspects of the discussion of moral responsibility have taken on a life of their own, so that many articles and books on the subject either do not even mention free will or else quickly set that topic aside and go on to examine moral responsibility in its own right. At the same time, interesting new challenges, unrelated to determinism, to both free will and moral responsibility arose in neuroscience and psychology. These challenges served to shift the discussion away from a focus on determinism and indeterminism and thereby contributed to the wide-​ranging debate we see today. As is evident in many of the articles in this volume, P. F. Strawson’s 1962 article “Freedom and Resentment” is singularly important for the development of discussion of moral responsibility in this period. The participants in the preceding debate largely divide into two camps. One affirms that our practice of holding morally responsible is justified on consequentialist grounds, and that the appropriateness of blaming and praising practices and any notion of desert they presuppose depends on the consequences produced. The other camp holds that the appropriateness of these practices requires

xviii   General Introduction a notion of free will that is incompatible with causal determination. On both types of view, our practice of holding responsible is open to challenge, either from doubts about the consequences it allegedly produces, or from deterministic conceptions of nature. Strawson sought to liberate our practice of holding morally responsible from these threats. The conception he proposed is now widely accepted, both in the way that he sees the practice as justified, and in the role that the reactive attitudes, such as moral resentment, indignation, and guilt have in that practice. This proposal has proved to be extraordinarily fertile ground for philosophical exploration. It has also given rise to challenges, in part from more sophisticated versions of the preceding views, for example, from innovative consequentialist theories of responsibility; and from new incompatibilist perspectives of the both libertarian and free will skeptical varieties. Our own view of the matter is that a complete understanding of the debates surrounding moral responsibility requires seeing potential connections to free will and to metaphysical questions concerning determinism and indeterminism as threats to free will. At the same time, there are many questions that can be investigated without first settling these issues. Thus, it is an exciting time in which there is special focus on questions about the relationships among different notions of moral responsibility, about what it is and when it is appropriate to hold people morally responsible, about the relationship of moral responsibility to the law and to distributive justice, among others. Further, new empirical work on issues such as psychopathy, addiction, and implicit bias provide opportunities for moral responsibility theorists to contribute to pressing concerns and at the same time test and hone theoretical frameworks for responsible agency. We now turn to an overview of the articles in this volume.

I.  Theories of Responsibility The recent focus on moral responsibility has given rise to the several distinct theories about what moral responsibility is. One division among these theories is between backward-​looking and forward-​looking accounts. The paradigm case of a backward-​ looking view is the one that regards the practice of holding morally responsible as aiming to secure the satisfaction of basic moral desert. On the basic form of moral desert, an agent who has acted wrongly deserves to be blamed and perhaps punished just because she has acted for morally bad reasons or with a bad quality of will, and someone who has acted rightly deserves credit or praise and perhaps reward just because she has acted for morally good reasons or with good will. Such desert is basic because the desert claims are fundamental in their justification. They are thus not justified by virtue of further considerations, such as the anticipated good consequences of implementing these claims. Forward-​looking theories, by contrast, view the practice of holding morally responsible as aiming to secure goods such as moral formation, protection from harm, and reconciliation in relationships. A second division concerns the conditions on an agent’s moral responsibility or on an agent’s properly being held morally responsible.

General Introduction   xix Particularly influential in recent decades are views on which the leading condition is a rationality provision, one according to which moral responsibility requires adequate sensitivity or responsiveness to reasons. Other views deny this, and test cases involve agents who are defective in their reasons-​responsiveness, such as, arguably, psychopaths: Can they be blameworthy, or can’t they? In “Instrumentalist Theories of Responsibility” Manuel Vargas discusses forward-​ looking views and contrasts them with their backward-​looking counterparts. Of particular interest to him are theories that invoke non-​basic desert. John Rawls, in his 1955 article “Two Concepts of Rules,” presents a two-​tiered theory in which legal justifications for punishment are entirely backward-​looking, in that lawyers, judges, and juries appeal only to backward-​looking reasons for punishment, while at the higher tier that practice is justified on forward-​looking, consequentialist grounds. In a similar vein, Daniel Dennett, Manuel Vargas, and Victoria McGeer advocate views on which the practice-​ level justifications for blame and punishment invoke backward-​looking considerations of desert, while that desert is not basic because at a higher tier the practice is justified by the anticipated effects, such as fostering and refining the ability to recognize and respond to moral considerations, and protecting people from the dangers wrongdoers pose. On the two-​tier accounts developed by Dennett, Vargas, and McGeer, our practice of holding agents morally responsible in a desert sense should be retained because doing so would have the best overall effects relative to alternative practices. Such views have the virtue of leaving our ordinary intuitions about deserved blame and praise, punishment and reward, in place, while conceiving their grounding in a way that avoids the contested notion of basic desert. In “Reasons-​Responsiveness, Frankfurt Examples, and the Free Will Ability” Michael McKenna provides a searching discussion of the view, widely held especially among compatibilists, that reasons-​responsiveness is the core criterion for moral responsibility. Many contemporary compatibilists also accept the intended lesson of Frankfurt examples (see section IV), that robust alternative possibilities for action are not required for moral responsibility. The paradigmatic compatibilist and Frankfurt-​defending view was developed by John Fischer in the 1980s and 1990s, at times with Mark Ravizza as a collaborator. It affirms that moral responsibility requires an agent to be receptive to reasons, and to moral reasons in particular, and to be able to act on them. Accordingly, both reasons-​receptivity and reasons-​reactivity, each in some degree, is required for moral responsibility. The interaction between the reasons-​responsiveness criterion and the lesson of Frankfurt examples gives rise to complex and interesting views, species of which have been developed by Fischer and Ravizza, McKenna, Dana Nelkin, and Carolina Sartorio, among others. In “Attributionist Accounts of Responsibility” Matthew Talbert discusses another currently influential theory of moral responsibility, one that originates in the work of T. M. Scanlon, and was developed and defended by Pamela Hieronymi, Angela Smith, and Talbert himself. On the attributionist view, key to determining moral responsibility are an agent’s evaluative judgments, revealed through actions and omissions, and through directly expressed beliefs and attitudes. Evaluative judgments have this central

xx   General Introduction role because they concern how agents relate morally with one another; they indicate whether the agent has a morally appropriate other-​regarding attitude, such as respect or moral concern, or an inappropriate attitude, such as moral indifference or cruelty. A distinguishing feature of attributionism is that it takes the evaluative judgment and the attitude it reveals to settle whether an agent is blameworthy, even if she is not morally responsible for bringing it about that she has that attitude, and even if she could not have avoided having it. As Talbert puts it, an agent’s being at fault for having a morally defective attitude is not required for the agent to be blameworthy on account of it. Thus, on Smith’s view, an agent can be blameworthy for forgetting a birthday, even if he is not at fault for forgetting. On Talbert’s account, psychopaths can be blameworthy for their lack of moral concern despite this lack being a result of their psychopathic neural condition, over which they have no control. Unlike free will skeptics, attributionists contend that in such cases the appropriate sort of blameworthiness standardly involves moral resentment or indignation. In sum, proper assessments of blameworthiness are concerned with judgments and attitudes attributable to an agent, and considerations of control over those judgments and attitudes do not have the role they have in other views.

II.  Kinds of Responsibility Our actual practice of holding morally responsible involves a number of different aims, and a range of responses justified by those aims. Some have argued that this complexity can be regimented, and that there is ultimately a single sense of moral responsibility that serves to unify the practice. Such monists about moral responsibility include R. Jay Wallace, George Sher, T. M. Scanlon, and Angela Smith. Others contend that such unitary views misrepresent the practice, and they include Gary Watson, Dana Nelkin, Derk Pereboom, David Shoemaker, Gregg Caruso, and Michael Zimmerman. Such pluralists argue that it’s mistaken to claim that there is a single, unifying sense of moral responsibility; rather, there are multiple senses corresponding to multiple aims and attendant justifications, each of which is an aspect of our actual practice of holding morally responsible. In “Attributability, Answerability, and Accountability” Sofia Jeppsson points out that pluralists about moral responsibility embrace their position for one of two reasons. First, agents with various impairments, such as harsh childhoods affect the way in which they are morally responsible, without rendering them fully non-​responsible, and as a result they are morally responsible in some senses but not in others. David Shoemaker, for example, divides moral responsibility into three progressively more demanding senses: attributability, susceptibility to character evaluation; answerability, eligibility for engagement with moral reasons; and accountability, liability to responses with a deserved confrontational element. Second, skeptics about desert-​invoking moral responsibility such as Derk Pereboom and Gregg Caruso argue that there is a forward-​ looking, non-​desert-​involving kind of moral responsibility that we nonetheless have, and which suffices to ground our most important moral responsibility practices. This

General Introduction   xxi position requires differentiating between backward-​looking, desert-​invoking moral responsibility and the forward-​looking sort. Jeppsson argues that pluralism is crucial for the skeptics, since monism combined with skepticism stands to undermine our current moral life, while it is less crucial, but nonetheless reasonable, for non-​skeptics.

III.  Dimensions of Responsibility The philosophical literature on moral responsibility often takes for granted that there is a set of objects for which one can be responsible, though the focus is typically on the single category of actions. Further, it is frequently simply assumed that there are degrees of responsibility, and that only individuals can be responsible agents. But interesting questions arise about each of these dimensions of responsibility, namely, subjects, objects, and scalarity. A complete version of each theory of responsibility should say something about how it approaches each of these dimensions. In section X, we have three articles that explore these dimensions. Many examples in the literature on responsibility are cases of action, such as killing or harming another person. But we often hold people morally responsible for omissions as well. The fact that a friend didn’t show up when she said she would or that a parent forgot to pick up his child from school seem to be paradigm cases in which we hold each other responsible. Further, while some omissions are deliberate and therefore seem to involve at least a mental action of some sort, others are “unwitting” such as when it simply leaves our minds that we promised to be somewhere. We don’t do anything at all. It is not obvious how theories of responsibility can account for these being legitimate cases of responsible agency. In “Responsibility for Acts and Omissions,” Randolph Clarke takes on this question, among other related issues. He distinguishes between direct and indirect responsibility, where one might be indirectly responsible for something by at least partly being responsible for something else, such as an earlier decision not to write the appointment in the calendar. There are important cases, though, in which it seems there was no prior failure for which we are responsible that explains the omission for which we are culpable now. Attributionist theories that take agents to be blameworthy when they exhibit bad quality of will face difficulties including that these cases often don’t seem to involve any sort of ill will. Control theories face objections including that someone who is unaware that they ought to be meeting their friend can have the requisite control at that moment. While sensitive to these objections, Clarke provides the best case for how an attributionist theory of responsibility and for how a control theory that appeals to abilities and opportunities might account for such cases. Another dimension of responsibility that arises in numerous contexts is its scalarity. Moral responsibility and its forms, blameworthiness and praiseworthiness, seem to come in degrees. Interestingly, although this point is often taken for granted, it turns out that it is not easy to explain it on a variety of accounts of moral responsibility. In “Degrees of Responsibility,” Justin Coates suggests that degrees of moral responsibility and of

xxii   General Introduction blameworthiness in particular are a function of the difficulty of complying with the relevant norms, and he shows how a quality of will view, a control view, and an independent epistemic condition can best offer an account of difficulty and thus of degrees of responsibility. In the end, he suggests that we might not need to choose from among these options. A third dimension of responsibility concerns its subjects. Must responsible agents be individuals, or can groups be responsible? In “Group Responsibility,” Christian List takes up this pressing question, beginning with the challenge that we seem to want to assign moral responsibility in various cases in which assignments of responsibility to individuals do not seem adequate. When we’re asked to determine moral responsibility for the sinking of a ship, it might seem to be the operating company whose standards of safety were sub-​par. But when we think about what it takes to be morally responsible on any account we have seen, it is not obvious that groups meet the relevant conditions. List presents and assesses an argument that groups can be morally responsible by first showing that groups can be intentional agents. He subsequently considers the interesting question of the relationship between group and individual moral responsibility.

IV.  Determinism and the Ability to Do Otherwise This section features two articles on metaphysical challenges to moral responsibility, each of which addresses a challenge that arises from the threat of causal determination. In the first, “Moral Responsibility, Alternative Possibilities, and Frankfurt Examples,” Derk Pereboom engages the concern that the control in action required for moral responsibility requires that the agent could have refrained from performing the action at issue, which may be precluded by causal determination. The second, “Manipulation Arguments Against Compatibilism,” by Derk Pereboom and Michael McKenna, surveys argument that claim causal determination to be relevantly similar to intentional deterministic manipulation, and reasons that because such manipulation precludes moral responsibility, so does causal determination. The topic of the first article is the assumption that moral responsibility for an action requires that the agent could have avoided the action, and thus had an alternative possibility available to her. This intuition that blameworthiness, in particular, requires that the agent could have done something that would have precluded her blameworthiness is widespread and for a long time went unchallenged. Then in a landmark paper published in 1969, Harry Frankfurt, using an intriguing type of example, argued that an agent’s moral responsibility for an action does not require that she could have done otherwise. Here is Carolina Sartorio’s (2016) version of Frankfurt’s original example: A neuroscientist, Black, wants Jones to perform a certain action. Black is prepared to go to considerable lengths to get his way, but he prefers to avoid showing his hand unnecessarily. So he waits until Jones is about to make up his mind what to do, and

General Introduction   xxiii he does nothing unless it is clear to him (Black is an excellent judge of such things) that Jones is going to decide to do something other than what he wants him to do. If it were to become clear that Jones is going to decide to do something else, Black would take effective steps to ensure that Jones decides to do what he wants him to do, by directly manipulating the relevant processes in Jones’s brain. As it turns out, Black never has to show his hand because Jones, for reasons of his own, decides to perform the very action Black wants him to perform.

The idea is that even though Jones could not have avoided acting as he did, it’s intuitive that he is morally responsible, and blameworthy in particular, for the action: He did it on his own and Black never intervened. Since the publication of Frankfurt’s article, there has been a heated and voluminous debate about this issue. In response to the challenge from Frankfurt examples, defenders of the alternative-​possibilities requirement have advanced a number of counterstrategies. One involves contending that despite initial impressions, the example does feature a relevant alternative possibility after all; John Fischer calls this the “flicker of freedom strategy.” Another is to argue that despite initial impressions, the example requires causal determinism in the actual causal history of the action in order to secure the absence of alternatives, and this counts against the intuition that the agent in the example is morally responsible. These objections have given rise to Frankfurt examples designed to address the objections. The first article focuses on two Frankfurt examples, in particular, one due to Al Mele and David Robb, and another advanced by Derk Pereboom and David Hunt. Objections to each type of example are addressed, and tentative assessments are provided. A widespread concern for our being morally responsible for our actions is the prospect that they are naturalistically causally determined, that is, rendered inevitable by virtue of the remote past and the laws of nature, factors beyond the agent’s control. However, many compatibilists about naturalistic causal determination and moral responsibility are not strongly moved by this concern. In the face of this challenge, incompatibilists have drawn on an analogy between naturalistic causal determination and intentional deterministic manipulation by other agents. Derk Pereboom and Michael McKenna survey these kinds of arguments in “Manipulation Arguments Against Compatibilism.” A manipulation argument begins with an imaginary case in which an agent is intentionally deterministically manipulated by another agent to perform an action, typically an immoral one. The story specifies that the conditions compatibilists would take to be sufficient for moral responsibility are satisfied. The case is meant to elicit the intuition that, due to the nature of the manipulation, the agent is not morally responsible for what she does. It is then argued that this manipulated agent differs in no responsibility-​relevant respect from a counterpart who is naturalistically determined so to act. Because there is no responsibility-​relevant difference between the intentionally manipulated agent and the counterpart who is naturalistically determined, the naturalistically determined agent isn’t morally responsible for the action either. The conclusions drawn include any normal, non-​manipulated agent who is naturalistically determined to perform an action is not morally responsible for it even if she satisfies all the prominent compatibilist

xxiv   General Introduction conditions on moral responsibility; the compatibilist conditions are insufficient for moral responsibility; and naturalistic causal determination in general is incompatible with moral responsibility. Compatibilist responses to manipulation arguments can be divided into two categories. Soft-​line replies do not resist the intuition that the manipulated agent is not morally responsible, and instead aim to show that an agent who is merely naturalistically determined differs from the manipulated agent in a way relevant to moral responsibility. Hard-​line replies, by contrast, resist the intuition, essential to the incompatibilist’s case, that the intentionally and deterministically manipulated agent is not morally responsible (the terminology is due to McKenna). The article, co-​authored by a proponent of manipulation arguments (Pereboom), and a detractor (McKenna), surveys the types of manipulation argument, the nature of the dialectic, and the prominent replies.

V.  Skepticism Contemporary free will skeptics argue that sort of free will required desert-​invoking moral responsibility is incompatible with causal determination by factors beyond the agent’s control and also with the kind of indeterminacy in action required by the most plausible versions of libertarianism. Against the view that this sort of free will is compatible with causal determinism, some skeptics, such as Derk Pereboom, invoke a manipulation argument, while others, such as Neil Levy, adduce an argument according to which luck precludes the sort of control in action required for moral responsibility. Against event causal and non-​causal libertarianism, skeptics tend to advance another version of a luck objection. Some argue that while agent-​causal libertarianism could, in theory, supply this sort of control, it cannot be reconciled with our best physical theories. Other free will skeptics, such as Gideon Rosen in particular, argue that the epistemic conditions on blameworthiness are almost never satisfied. Galen Strawson and others influenced by him, such as Saul Smilansky and Tamler Sommers, invoke the basic argument against the sort of moral responsibility at issue in the debate. The basic argument, in outline, proceeds as follows. When an agent acts, she acts because of the way she is. To be responsible for acting, she must then be responsible for the way she is, at least in key mental respects. But if an agent is to be responsible for the way she is in those mental respects, she must be responsible for the way she is that resulted in those respects. This reasoning generates a regress, which indicates that finite beings like us cannot satisfy the conditions on the sort of responsibility at issue. Saul Smilanksy, in “Illusionism,” describes and defends the illusionist position about free will he has developed over the past 25 years. This position combines the following claims: (1) that we human beings lack the free will required for moral responsibility in the desert-​invoking sense; (2) that we are generally under the illusion that we have such free will and moral responsibility; (3) that such moral responsibility is central to our understanding and practice of morality, justice, interpersonal relations, self-​respect and respect for others; and (4) that the losses of not maintaining the illusion are too great to

General Introduction   xxv dispel it. In Smilansky’s view, the illusion of free will is an indispensable enabling condition for many of the most important elements of our view of ourselves, for our ability to maintain moral seriousness, a sense of value, and indeed our very sense of self. We ought to maintain this illusion to avoid the damaging consequences that would ensue if we did not. Gregg Caruso, in “Free Will Skepticism and Criminal Justice: The Public Health-​ Quarantine Model,” represents a more optimistic approach to the denial of desert-​ invoking moral responsibility and the sort of free will it requires. He develops such an optimistic outlook specifically in the area of criminal justice. In his view, free will skepticism is incompatible with retributive justification of punishment, but it leaves intact other, more benevolent ways to respond to criminal behavior, such as incapacitation, rehabilitation, and alteration of relevant social conditions. Caruso argues that these methods are both morally justifiable and sufficient for good social policy. The argument for incapacitation has as its starting point Derk Pereboom’s quarantine analogy, defended on the ground of the right to self-​defense, and subsequently embeds it within a broader justificatory framework drawn from public health ethics. The resulting perspective, which Caruso calls the public health-​quarantine model, provides a framework for treatment of criminals that he contends is more humane than what retributivism justifies, and is preferable to other non-​retributive alternatives. In “Metaskepticism” Tamler Sommers describes and defends a position on which disputes about theories of moral responsibility do not amount to factual disagreements, but rather to non-​factual disagreements rooted in distinct attitudes. In his view, the term “moral responsibility” is used in many ways, but the desert-​invoking sense is at the center of the contemporary philosophical debate. Sommers notes that philosophers who propose and defend theories of moral responsibility understood in this way propose criteria for moral responsibility that are meant to apply universally—​for all agents, for all societies—​and they appeal specifically to intuitions to justify these conditions. These theories thus rely on empirical assumptions concerning the universality of certain attitudes and on our intuitions. Against this tradition, Sommers first offers evidence of radical differences in attitudes and resulting intuitions about moral responsibility across cultures and throughout history. He then argues that this variation in intuitions cannot be explained away as the product of irrationality or disagreement over nonmoral facts, and he concludes that we have no principled way of establishing universal factual conditions of moral responsibility. Sommers calls this position “metaskepticism” about moral responsibility, because it would undermine not only positive theories of responsibility, but also skeptical theories according to which human beings at all times and places do not deserve praise or blame for their actions.

VI.  Blame As noted, recent decades have witnessed the strong influence of P. F. Strawson’s “Freedom and Resentment” on the philosophical discussion of moral responsibility. This has

xxvi   General Introduction especially been the case for the discussion of blame, which has notably flourished in this time. In “Blame and Holding Responsible,” Angela Smith provides an overview of the developments in accounts of blame over this period. As she points out, prior to this time the overt-​sanction account of blame was widely accepted, while it has now fallen decisively out of favor. A key objection that she notes to such an account is that even when blame is overt, its force seems to reside not in the outward blaming response, but rather in the inner reactive attitude expressed by this outward conduct. Smith divides current accounts of blame into four categories: (1) Appraisal Views, on which to blame a person for an action is to judge that action to reveal something negative about that person’s character (as discussed above; for Gary Watson appraisal/​attributability is one of the two kinds of blame, accountability is the other); (2) Reactive Attitude Views, developed by Strawson, Wallace, and many others, which take blame to be a response involving reactive attitudes such as resentment, indignation, and guilt to bad quality of will manifested in action or omission, (3) Hybrid Views, whose proponents include George Sher and T. M. Scanlon, which seek to accommodate the objection that Reactive Attitude Views fail to capture the variety in our blaming practices, while yet providing a unifying rationale that can explain what that variety has in common; and (4) Functional Views, which aim to specify the functional role blame plays in our moral lives; its advocates include Victoria McGeer and Manuel Vargas. Smith then recounts the current interest in the issue of standing to blame; the proposed conditions on standing include non-​hypocrisy, non-​complicity, and whether it is one’s business to blame. Finally, she notes the current interest in issues regarding the procedure of blame; and in this context she addresses the timing, place, and manner of blame. R. Jay Wallace, in “Responsibility and Reactive Attitudes,” explores the Strawsonian proposal that resentment, indignation, and guilt have the central role in an account of moral responsibility. As Wallace explains, these attitudes are reactive insofar as they are reactions to the quality of will expressed in actions or omissions of others. In Strawson’s view, these reactive attitudes are associated with demands that we make of each other for consideration or regard. Our susceptibility to these reactive attitudes is connected to our adopting a stance of demanding that others adopt attitudes of goodwill or consideration. When these demands are flouted, the result amounts to a kind of injury or offense by one party against another. Holding someone to such a demand is constitutively connected to the reactive attitudes: The making of the demand is the proneness to such attitudes. Wallace considers the objection that participation in personal relationships may require a susceptibility to reactive attitudes in only a more relaxed sense, in which the reactive attitudes include sorrow and disappointment. Wallace contends that it isn’t clear that a susceptibility to attitudes of this kind suffices for holding the target of those attitudes accountable. For this reason, one may doubt whether personal relationships and what Strawson calls the participant stance constitutively involve a disposition to hold people accountable for what they do. In “Response-​Dependence Theories of Responsibility” David Shoemaker sets out, refines, and defends a response-​dependent theory of responsibility. The article begins with P. F. Strawson’s response-​dependent account, and as Shoemaker explains, what

General Introduction   xxvii attracted Strawson to such view is the prospect of establishing that there is no threat to our practices from external considerations, from demands for justification for the practices as a whole from outside of the framework of those practices. But, Shoemaker argues, any theory of these practices would be vulnerable to demands for external justification unless moral responsibility itself is taken to be ultimately response-​dependent. Response-​independent views have our practice of holding responsible presupposing the antecedent fact that their targets are morally responsible, and so the justification for the practices would need to be grounded in external facts about agential responsibility. A problem for response-​dependent views is that our morally angry blaming responses may be mistaken. People are often angry with the mentally ill and with children for wrongdoing even though they are not morally responsible. But a concern for the response-​independent alternative is that there is no response-​independent way to unify the properties that make blaming anger appropriate. Shoemaker contends that what does unify all of the properties is just that they merit blaming anger, and that there is no better account. In his view, this is analogous with the funny; what unifies all of the properties that make the amusement response appropriate is just that they merit this response.

VII.  Responsibility, Knowledge, and Causation If an agent is responsible for some act, omission, or state, it is generally thought that she must meet certain epistemic conditions—​for example,, she must know what she is doing—​as well as certain causal conditions—​for example, she must be causally related to a state to be responsible for it. But getting clear on what exactly the content of the relevant conditions should be, or even whether they should be accepted is a matter of ongoing debate. In “Ethics Is Hard! What Follows? On Moral Ignorance and Blame” Elizabeth Harman takes up the question of the conditions under which ignorance excuses. In some cases, it is uncontroversial that ignorance excuses, as in Judith Thomson’s case in which one had taken all due care but was ignorant that one’s light switch was attached to the trigger for a bomb next door and turns on the light. Recently, there has been much discussion about whether moral ignorance excuses. For example, what should we say in the case where someone mistakenly thinks it is morally permissible to take revenge on uninvolved family members of those who have harmed others? Depending on what theory of moral responsibility we adopt, we are likely to arrive at different answers. Harman lays out arguments for the position that such ignorance does excuse when one has not violated any norms of belief management, but ultimately rejects these. In doing so, she takes there to be an asymmetry between moral and nonmoral ignorance. In some cases of moral ignorance, she argues, one’s ignorance manifests inadequate care, and this by

xxviii   General Introduction itself grounds blameworthiness. In this way, she appeals to a theory in the attributionist or quality of will family of theories. In “Responsibility and Causation” Carolina Sartorio examines the role of causation in cases of indirect responsibility for outcomes and in cases of direct (or as she calls it, “basic”) responsibility. As she notes, it is initially plausible that “an agent is responsible for an outcome if and only if there is some behavior of hers such that: she is responsible for that behavior, the behavior caused the outcome, and she foresaw that the behavior would cause (or was likely to cause) that outcome.” However, there are a number of challenges to this thesis she explores, including the existence of unwitting omissions which arguably are not instances of behavior at all and do not cause anything, while, intuitively, we can be responsible for states of affairs that seem to depend on them. When it comes to direct responsibility, Sartorio considers the thesis, now familiar as a lesson from Frankfurt cases, that responsibility depends only on the actual causal sequence and not on the availability to the agent of alternative possibilities. She canvasses objections to this view, some based on the idea that responsible agency requires reasons-​ responsiveness, which, in turn depends on more than the actual causal sequence. Finally, Sartorio proposes an intriguing way of rejecting the need for more than the actual causal sequence, in part by providing an expansive account of the nature of causal sequences.

VIII. Responsibility, Law, and Justice Questions concerning the nature of moral responsibility arise in a number of key debates in legal and political theory. First, there is a serious debate in the legal literature about the relationship between moral and criminal responsibility. This leads naturally to questions about what, if anything, can rationalize state punishment. Skepticism about basic desert moral responsibility challenges retributivist justifications for state punishment, and motivates alternative accounts, as argued in Gregg Caruso’s article in section V. To the extent that responsible agency is required for legitimate criminal conviction, interesting questions arise about where the burden of proof lies for showing that defendants have or lack it. Outside of criminal justice, moral responsibility is a much-​debated candidate factor in determining a just system of distribution of societal goods and burdens. s Brink, in “Responsibility, Punishment, and Predominant Retributivism,” explores possible rationales for state punishment, including purely consequentialist ones that take account only of the future good that can come from punishing such as a more protected and peaceful society, and a view that is retributive in taking appropriate punishment to depend on desert which in turn depends on agents’ culpable, or responsible wrongdoing. Brink surveys a number of objections to this latter approach, including that it implies that the felt suffering of offenders is good in itself. He argues that subtle versions of retributivism can avoid many of them by, for example, taking the object of desert to be the deprivation of certain rights rather than felt suffering. He also presents

General Introduction   xxix advantages for adopting a version in which retributivism is “predominant” as a constraint on the justification of punishment, but other factors such as good consequences for a peaceful society are also relevant. In the process of exploring these questions, Brink brings out the ways that a plausible theory of moral responsibility and one of legal responsibility can inform and complement each other. In “Legal Responsibility: Psychopathy, a Case Study” Elizabeth Shaw addresses the interesting question of whether evidence for responsible agency ought to be treated similarly or differently from other sorts of evidence of criminal guilt such as whether the person engaged in the relevant behavior or had an alibi. This question arises naturally in discussions of the “insanity defense.” There are many versions of the insanity defense across jurisdictions, but one common element is the ability to grasp reasons for avoiding wrongdoing. Shaw points out that while the burden of proof is typically on prosecutors to show each element of a crime, the burden is typically on defendants to show that they meet the excusing conditions provided by the insanity defense. Is this asymmetry justified? Using psychopathy as a case study, Shaw enters into the debate about this question and ultimately concludes that the current system is not well-​supported. She also argues for an approach that requires individualized evidence and not simply appeal to participation in a diagnostic category to show whether defendants possess the relevant responsible agency. These points generalize to defendants diagnosed with other conditions that might undermine responsible agency. Moving from criminal justice to distributive justice, in “Responsibility and Distributive Justice” Richard Arneson takes up the question of how responsible agency should factor into a just state system of distribution. One prominent view is Luck Egalitarianism, which endorses an equal distribution of goods (for example, resources or opportunities) but which relaxes that distribution in cases in which inequalities are the result of people’s own responsible and blameworthy choices. For example, if a fully responsible agent voluntarily takes great risks in mountain climbing, ignoring warnings, the luck egalitarian can point to the intuition that others who did not take such unreasonable risks should not then have their fair shares of society’s goods significantly decreased in order to support an expensive rescue. In evaluating this view and others, Arneson brings out a number of subtle ways in which responsible agency might be implicated in distributive justice.

IX.  Responsibility, Neuroscience, and Psychology As discussed earlier, in recent decades there has been a widening of the focus when it comes to threats to responsible agency beyond that of determinism or mechanism. Some of these come directly from recent developments in neuroscientific and psychological research. For example, a large body of research is often interpreted as indicating

xxx   General Introduction that many of our actions are not guided at all by conscious deliberation; and another reveals that situations, by contrast with stable character traits such as honesty and generosity, affect our choices much more than we might have expected. In very different ways, these findings have been thought to undermine our assumption that people are generally responsible for what they do and don’t do. Research in psychology has also given responsibility theorists a very different reason to reassess widespread assumptions via questioning of the use of intuitions about responsible agency in philosophical theorizing. Work in experimental psychology that is based on eliciting intuitions from participants about specific scenarios has provided reason for philosophers to take great care not to mistakenly generalize on the basis of what might be confounds or biases. Alfred Mele, in “Moral Responsibility and Neuroscience,” takes on the influential challenge to free will from the neuroscientific studies taken to show that much of human decision-​making is unconscious and that free will and moral responsibility are ruled out in such cases. In earlier work, Mele has taken a close look at the particular design of the experiments to raise questions about whether they really show that decisions are unconscious in the experimental conditions. Here Mele continues the project of questioning the conclusion of the experimenters by raising questions about whether the experimental results would generalize even granting the experimenters’ claims, and, further, whether the assumption that consciousness is required for morally responsible agency is plausible in the first place. In “Responsibility and Consciousness,” Matt King and Peter Carruthers also question this assumption in the course of exploring the relationship between consciousness and responsibility more generally. They consider three cases of unconscious phenomena: automatism, forgetting, and implicit bias. In each case, it seems that we have a choice of rejecting either the assumption that responsibility requires consciousness or the claim that we can be responsible in these cases. King and Carruthers offer reasons to deny that consciousness is required for responsible agency, although they note that it can still be highly relevant in many cases. It is important to recognize how the choice between different accounts of responsible agency (e.g., accounts that emphasize control or ones that emphasize a reflection of the agent’s motivations or true self) affect our judgments about the relevance of consciousness to responsibility. King and Carruthers conclude by noting connections between the question of how consciousness is related to responsibility and questions in legal theory concerning mens rea. Social psychology has also provided a serious challenge to the idea that we are responsible for our actions, or at least that we are responsible nearly as much as we typically assume. So-​called “situationist” experiments, such as that of Milgram’s obedience studies and the Stanford Prison Experiment, have been thought to show that the situations in which we find ourselves have a far greater impact on our choices and actions than any sort of individual and distinctive personality traits we have. If the vast majority of participants are willing to shock a person they believe to be an innocent participant in a learning experiment to levels that approach the danger of death, this suggests that many people whom we would have antecedently described as generous or kind people are in fact acting because of the luck of the situations in which they find themselves in. And

General Introduction   xxxi this seems to undermine our intuition that people are free and responsible agents. In “Responsibility and Situationism” Brandon Warmke surveys the psychological literature that brought about this challenge, and asks why these results seem so threatening. Is it because they seem to suggest that agents’ actions are determined? Or that they are victims of luck? Or that their reasons-​responsive capacities are impaired? Warmke goes on to propose ways one might address the challenge and still preserve the assumption that we are largely responsible for our actions, despite initial appearances. He argues that even if that challenge can be answered, we might still learn important things about the limits of our capacities that are relevant to moral responsibility. In “Experimental Philosophy and Moral Responsibility” Gunnar Björnsson surveys a large and rapidly expanding interdisciplinary area of research that elicits and analyzes reactions from nonspecialist participants to presentations of scenarios with philosophical relevance. These studies might be thought of as systematic large-​scale studies of intuitions in response to philosophers’ thought experiments. There have been surprising results that pose challenges for philosophers who appeal to intuitions about cases to support theoretical theses like compatibilism or incompatibilism. For example, some experiments appear to indicate that vast majorities of participants, when faced with the gory details of a killing and are told it was determined, believe that the killer is blameworthy and responsible. But others show that when asked in the abstract whether compatibilism about determinism and blameworthiness is plausible, a vast majority say no. Björnsson explores a number of hypotheses that attempt to explain how these seemingly inconsistent responses can be made consistent, or how one set or the other can be discredited as due to an unsupported bias. He then offers a rich critical examination of the value such experimental philosophy might have for philosophical theorizing, concluding that whatever value it has, it must be combined with other bodies of evidence and philosophical theorizing to have a legitimate place in the debate.

X.  Responsibility, Relationships, and Meaning in Life P. F. Strawson’s “Freedom and Resentment” has inspired much reflection concerning how the issues of free will, moral responsibility, personal relationships, and meaning in life interact. In his conception, the reactive attitudes are essential to the practice of holding responsible, and susceptibility to these reactive attitudes is required for good personal relationships. It’s widely agreed upon that good personal relationships are a core component of meaning in life for most of us, and this forges a connection between meaning in life and the preceding set of issues. While this set of connections has been contested, Strawson’s perspective enjoys widespread acceptance. Strawson did not comment on a possible connection of these considerations to religion, but historically this link is very prominent, in both theistic and non-​theistic conceptions.

xxxii   General Introduction In “Moral Responsibility and Metaphysical Attitudes” Paul Russell discusses various positions in the debate about free will and moral responsibility with the use of the notions of “optimism” and “pessimism” that Strawson introduced to characterize certain opposing parties in this debate. As Russell points out, Strawson categorized as pessimists those who hold that if the thesis of determinism is true, the concepts of moral obligation and responsibility have no application, and the practices of punishing and blaming, of expressing moral condemnation and approval are unjustified. Pessimists are thus incompatibilists of a certain sort. Optimists, by contrast, “hold that these concepts and practices in no way lose their raison d'être if the thesis of determinism is true; they “emphasize “the efficacy of our practices of punishment, and of moral condemnation and approval, in regulating behaviour in socially desirable ways.” Optimists are a certain kind of compatibilist, prevalent in the first half of the 20th century, who reject basic desert and affirm a consequentialist account of moral responsibility practices. Strawson maintains that such optimists leave out the role of the reactive attitudes. Russell discusses Daniel Dennett as a contemporary optimist and then contrasts his view with optimistic skepticism about free will, as well as other less optimistic skeptical positions. The article concludes with an account of Bernard Williams’s position and how it is situated in this landscape. Dana Nelkin, in “Relationships and Responsibility,” explores the ways in which free will and moral responsibility appear linked with personal relationships. She examines three discussions in which the connection between moral responsibility and relationships is at stake: the debate about whether valuable human relationships such as friendship in some way presuppose responsible agency on the part of their participants, the debate about whether the conditions of responsible agency must make essential reference to relationships, and the debate about the ethics of blame. The third of these discussions receives most of her attention, and on this issue she has made a distinctive contribution. Nelkin focuses on her argument from the premise that friendship requires special obligations or duties, and the principle that to be subject to an obligation requires free agency, to the conclusion that friendship requires free agency. The Kantian “ought implies can” principle has a key role in the argument. If one is obligated to meet a certain demand in the context of friendship, then it must be that one could have met the demand even if one hasn’t actually met it. Accordingly, this obligation requires the freedom to do what one hasn’t done. Nelkin provides a careful critical discussion of this line of reasoning. In “Responsibility, Personal Relationships, and the Significance of the Reactive Attitudes” Seth Shabo defends Strawson’s claim that relinquishing moral resentment, and the negative reactive attitudes more generally, is not a genuine practical possibility for us, even supposing that determinism’s truth provides theoretical grounds for doing so. The reason is that susceptibility to these reactive attitudes is inseparable from our commitment to personal relationships, which we can’t practically relinquish. Shabo argues that the only reliable strategy for immunizing ourselves against resentment is to be significantly less emotionally attuned than we normally are to others’ attitudes toward us. For when we are so attuned, our control over whether we take displays of disregard

General Introduction   xxxiii personally is at best indirect and limited. Because we cannot control how attuned we are to others’ attitudes without limiting our emotional investment in how they see us, there is an irresolvable practical tension between having good personal relationships and avoidance of taking maltreatment personally. In “Reconciliation and the End of Responsibility” Linda Radzik explores the nature of reconciliation, which she conceives as restoration of good relations properly understood after wrongdoing has impaired those relations. On her view, reconciliation is one way of ceasing to hold responsible, that is, ceasing to consider blameworthy. Although reconciliation might often be paired with forgiveness, she contends that reconciliation necessarily involves interaction between parties, while forgiveness can be private and non-​interactive. Radzik considers arguments against the value of reconciliation based on the concern that it amounts to condoning the wrongdoing that impaired the relationship. In response, Radzik shows how a subtle account of reconciliation might effectively respond to this objection, and she argues that reconciliation can be a desirable state despite this concern. Per-​Erik Milam, in “Forgiveness,” explores the connection between forgiveness and moral responsibility. He begins by introducing a standard model of forgiveness and several recent challenges to it, each of which highlights potential connections between forgiveness and responsibility. On this standard model, forgiveness is essentially the overcoming of blame, usually understood as involving resentment, in response to the wrongdoer’s change of heart and contrition. This model has recently been challenged on a number of grounds. For instance, Dana Nelkin argues that one might forgive while retaining blame and resentment, where forgiveness consists in releasing the wrongdoer from the obligation to apologize or compensate. In consequence, Nelkin and Brandon Warmke have defended an alternative norm-​changing view of forgiveness, according to which forgiveness changes the normative significance of the offense from a state that warrants various negative responses to one in which at least some of these responses are no longer justified. Milam provides a detailed discussion of the various interesting moves that have been made in this very current and lively debate. In “Responsibility and Religion” Dan Speak’s focus is on contemporary theistic religion, and his leading question is what it might be about a commitment to theism that shapes attitudes about free will and moral responsibility. The attitude that predominates among Christians in particular is libertarianism, which affirms that we have free will of the sort required for desert-​invoking moral responsibility, and that such free will is incompatible with determinism. In this view, not only is naturalistic determinism false, but theological determinism is ruled out as well. The aversion to theological determinism sets contemporary theism apart from a substantial portion of the historical monotheistic tradition. Speak suggests five general answers for why contemporary theism tends strongly toward libertarianism: (1) theism has emphasized human likeness to God, (2) theism proposes divine purposes for our agency, (3) theism invites eternal expectations for our moral lives, (4) theism finds itself with internal tensions that it is under philosophical pressure to resolve, and (5) theism faces relevant external challenges to its philosophical coherence and plausibility to which it is at pains to respond.

xxxiv   General Introduction

XI.  Case Studies Ideally, the best work on moral responsibility will have implications for real-​life hard cases about which we are often unsure at least at first what to say about whether, or to what extent, people are responsible for their actions. Putting this into practice requires bringing together subtle theoretical work on moral responsibility with a sophisticated understanding of the empirical facts. The influence might work in both directions: Not only would independently plausible theories have implications for particular cases, but our intuitions about particular cases would in turn make us rethink or revise our theoretical commitments when there is abiding conflict. Thus, work that brings together well-​worked out theory and an accurate understanding of the empirical facts can be valuable in multiple ways. In recent years, we have seen more and more work of this kind. Here we have three entries that consider moral responsibility in connection to three different kinds of cases: addiction, implicit bias, and atrocity in the form of war crimes and psychopathy. In “Moral Responsibility in the Context of Addiction” Doug McConnell sets out cutting-​edge scientific work on addiction and shows how various theories of moral responsibility can be deployed to explain how addiction can mitigate responsibility. He identifies at least three ways this can happen: Dopamine dysregulation can result in increased vulnerability to temptation via abnormally large judgment shifts and akratic motivation; environmental conditions that make living without drugs very difficult; and fatalistic narratives that become entrenched. McConnell draws subtle conclusions about just how much these factors mitigate and urge us to consider the question of how it is most effective to treat those who are addicted when it comes to our responsibility practices. In “Atrocity, Evil, and Responsibility” Dominic Murphy and John Doris draw our attention to cases of War Crimes and Psychopathology. They consider the most extreme cases of wrongdoing, which we naturally label “evil.” Yet, after examining the cases in more detail, they conclude that in these sorts of cases there is good reason to presume that the agents are not evil because, surprisingly, they are not responsible for their actions. In drawing this conclusion, Murphy and Doris focus on establishing an accurate understanding of the cases, and on common assumptions about what sorts of conditions are excusing, including both ones that are internal to agents and ones that are external or situational. Once we have an accurate understanding of the cases, they argue, we will find excusing conditions in both places. Another challenging case is that of implicit bias, or, in other words, cases in which we have implicit, non-​conscious, and typically negative attitudes toward people of particular groups. We can see how different theories explored in the preceding articles might apply to this case. Given that the attitudes are not conscious, are we nevertheless able to see them as in some way responsive to reasons, under our control, and as embodying evaluative judgments? Further, if we are responsible for omissions, then

General Introduction   xxxv even if the attitudes lack any or all of these features, we might be indirectly responsible for them in virtue of having omitted to take measures to eliminate them or mitigate their effects. Some theories can more easily treat these as objects of responsibility than others, but to apply any theory, we first need an accurate understanding of just how implicit bias operates. In “Moral Responsibility for Implicit Bias and the Impact of Social Categorization” Maureen Sie describes in some detail how implicit bias works and also focuses on its significant social impact, bringing out just how much is at stake. Sie then considers how some of the theories explored in earlier chapters apply, and she encourages us to consider the forward-​looking question of what ways of treating people might be most effective in cultivating better future behavior. In principle, it might be that treating people as responsible even if they do not meet the conditions identified by our most plausible theories is most effective in this regard.

References Fischer, J. M., and M. Ravizza. (1998). Responsibility and Control: A Theory of Moral Responsibility. Cambridge: Cambridge University Press. Frankfurt, H. G. (1969). “Alternate Possibilities and Moral Responsibility.” Journal of Philosophy 66: 829-39. Kane, R.. (1996). The Significance of Free Will. New York: Oxford University Press. Sartorio, C. (2017). “Frankfurt-Style Examples.” The Routledge Companion to Free Will (1st ed.). K. Timpe, M. Griffith, and N. Levy, eds., London: Routledge. https://doi. org/10.4324/9781315758206 Strawson, P. F. (1962). “Freedom and Resentment.” Proceedings of the British Academy 48, pp. 187-211. Watson, G., ed. (1982). Free Will. New York: Oxford University Press. Wolf, S. (1990). Freedom within Reason. Oxford: Oxford University Press.

PA RT I

T H E OR I E S OF R E SP ON SI B I L I T Y

CHAPTER 1

Inst rum entali st T h e ori e s of Moral Resp onsi bi l i t y Manuel Vargas

1.  Instrumentalism vs. Consequentialism There is no single, substantive commitment shared by instrumentalist approaches to responsibility. Instead, instrumentalist theories are best understood as a family of views that employ instrumentalist or consequentialist considerations in a variety of often distinct theoretical roles. For example, some accounts have maintained that individual instances of holding responsible are simply forward-​looking attempts to modify behavior (Nowell-​Smith 1948; Smart 1961); others have argued that what responsibility is just is a general practice of prosocial influence (Schlick 1939; Dennett 1984); still others have argued that whatever the right account of the nature of responsibility and the practice of holding people responsible, the justification of responsibility is properly understood in instrumentalist terms (Vargas 2008, 2013a; McGeer 2013, 2015; Jefferson 2019). It is common to refer to this variegated family of theories as “consequentialist.” The description is informative and misleading in equal measure. In the 20th century, most instrumentalist theories of responsibility were held by people who were also consequentialists about normative ethics and practical reasoning, and who tended to see the development of an instrumentalist account of responsibility as an application of a broader commitment to consequentialism. Thus, consequentialist was an apt descriptor of those views. Over the past few decades, instrumentalist theories of responsibility have come to be held by people who are either not committed to consequentialism, or who see the commitment to instrumentalism about responsibility as at least in principle separable from a commitment to consequentialism about normative ethics. To characterize these (at least in principle) independently motivated instrumentalist accounts as consequentialist can

4   Manuel Vargas misleadingly suggest that these accounts necessarily involve a commitment to consequentialism about practical reason or normative ethics. Although the label consequentialist is faithful to the tradition, for the sake of clarity, I will follow Jefferson (2019) in using instrumentalist to refer to the general class of theories and consequentialist to refer to instrumentalist theories that are in some notable way connected to consequentialism about practical reason or normative ethics.1

2.  A Provisional Sketch As noted before, perhaps the most important traditional motivation for adopting an instrumentalist theory of responsibility was the (formerly) widespread acceptance of consequentialism in normative ethics. However, for those unpersuaded by consequentialism, this was no appeal at all. Still, there are reasons connected to debates about free will that seem to have made instrumentalist approaches appealing to some philosophers. This section offers a brief overview of the current state of play, and its connection to issues in debates about free will. The nature of free will is a famously vexed matter, and the most traditional way of characterizing the partisans involves distinguishing between those who think free will is incompatible with determinism (incompatibilists) and those who do not (compatibilists).2 Many incompatibilists have thought that free will’s requirements are relatively robust, metaphysically speaking. For example, free will has been thought to involve “contra-​causal” powers, causa sui, indeterministic causation, agent causation, or high-​level emergent causes. Such esoterica—​“panicky metaphysics,” in Peter Strawson’s (1962) memorable phrase—​has struck many observers as either undermotivated, ad hoc, or at odds with independently plausible pictures of the world. The general upshot is that responsibility skepticism can seem an appealing view for those who think that these putatively unattainable forms of agency characterize the correct metaphysics of freedom (G. Strawson 1994; Pereboom 2001; Caruso 2012). Enter instrumentalist theories. The metaphysics of agency required to make sense of instrumentalist theories tend to be comparatively unproblematic. On the simplest versions (ones on which some observers will add, “by which you mean simplistic”), it is enough that agents be influenceable by praise and blame. Influenceability is a low metaphysical bar to cross, but it gets many of the cases right: Where praise and blame have no 1 

Although it raises interesting questions about the taxonomy of theories of responsibility, this chapter takes no stand on the relationship of instrumentalist accounts to recent functionalist accounts (e.g., in Shoemaker and Vargas 2021; and arguably McKenna 2012 and Smith 2013). 2  This is a traditional way of construing the terrain. In the current literature, “determinism” is usually meant as a metaphysical thesis, but it is sometimes treated as a placeholder for a range of freedom-​ threatening things, which can include, in various formulations, physicalism; causal explanations in general; the causal closure of the physical; mechanism; naturalism; and determinism of more and less particular forms (such as psychological, neurological, biological, and physical determinism).

Instrumentalist Theories of Moral Responsibility    5 hope of altering the dispositions of agents, praise and blame can seem to lose their luster. The appeal of instrumentalist accounts is that they involve a relatively uncontroversial metaphysics of agency while appearing to preserve the practice more or less as we find it, while also explaining why the practice is normatively appealing (about which more later in this chapter). The putative appeal of the instrumentalist’s metaphysics of agency can strike some incompatibilists as an instance of philosophical sour grapes: The instrumentalist denies that we ever wanted what (to the incompatibilist) we manifestly do want of a metaphysics of agency. The dialectical issues here are tricky, and familiar to wider debates between compatibilists and incompatibilists. Partly, the issue turns on what one thinks of as the task of philosophy, whether it is methodologically conservative—​for example, a matter of descriptive metaphysics (Strawson 1962); or vindicatory, showing how our ordinary beliefs might be true (Nozick 1981)—​or instead a matter of potential revisionary or even eliminativist possibilities. These reforming adjustments, whether revisionist or eliminativist, can in turn be structured by different considerations about the relative importance of our practices versus our conceptual conjectures in fixing meaning and reference (Double 1996; Vargas 2011, 2013b; Nichols 2015; Caruso 2015; McCormick 2016; in the broader literature, see Rawls 1971; Daniels 1979; Jackson 1998). Methodological concerns aside, simple instrumentalist approaches are vulnerable to challenges about their extensional adequacy. In actual practice, we find that some agents may be influenceable by blame and condemnation without being responsible (for example, pets and infants). Classical consequentialist theories seldom had a clear way of addressing the challenge. Contemporary instrumentalist theories of responsibility have tended to make use of explicit conditions on who is properly subject to praise and blame. Important candidates have included reasons-​responsiveness, including socially sensitive reactivity to reasons (Vargas 2008, 2013a; McGeer 2014; Fricker 2016; Jefferson 2019). Like traditional consequentialist accounts, these newer accounts rely on powers that few doubt are compatible with determinism, but they provide more resources for getting the cases right: Animals might be influenceable but not suitably reasons-​ responsive, and thus not blameworthy. Different strategies are sometimes available for capturing cases where agents seem blameworthy but not influenceable. On “one-​tiered” or direct approaches, there is the possibility of appealing to various indirect effects, and the instrumentalist may also insist that the phenomenology of blame may come apart from its instrumentally best function (Miller 2014a; McGeer 2019) For “two-tiered” or indirect versions of instrumentalism, the idea is that blameworthiness is a status in a justified practice, where the practice is instrumentally justified but the statuses within it make no appeal to instrumental considerations (Vargas 2013a). A different family of motivations for instrumentalist theories of responsibility turns on instrumentalism’s relatively ready answer to doubts about the rational or normative grounds for responsibility practices. On these approaches, responsibility practices have some claim on us to the extent to which they generate morally desirable outcomes. Philosophers can and have readily disputed the appeal of, for example, retributive practices, or practices of blame that depend on metaphysically dubious notions of desert.

6   Manuel Vargas However, the appeal of encouraging good actions, dispositions, or forms of agency and of discouraging bad instances of the same seems straightforward. Instrumentalists try to extend this appeal to a more general account of why we have reason to encourage and sustain practices of moralized praising and blaming. Whether such extensions must necessarily fail—​whether they confuse the efficacy of holding responsible for the truth of one’s being responsible, for example—​is a matter we will return to later. If an antecedent commitment to consequentialism was once the main basis for traditional instrumentalist theories, it is one of the many ironies in the rehabilitation of this family of views that they tend to make relatively thin demands on one’s theory of normative ethics. The reasoning is as follows. Given that instrumentalist considerations are allowed on most theories of normative ethics—​although they may be subject to side constraints or principled restrictions—​a theory of moral responsibility that employs those considerations is not invoking elements that are automatically excluded from many theories of normative ethics. Instrumentalist approaches can therefore promise a relatively “thin” or nonpartisan way of accounting for moral responsibility in the sense of being, compatible with a range of normative ethical theories (Arneson 2003; Vargas 2013a). If one goes on to accept some contractualist or deontological account of normative ethics, then it is conceivable that one’s theory of normative ethics may provide additional side constraints that, for example, prohibit scapegoating the non-​responsible. In contrast, if one accepts a normative ethics that is purely consequentialist, then there is no basis for objecting that an instrumentalist theory of responsibility introduces special worries about scapegoating the non-​responsible.3 Independence from consequentialism about practical reasoning and normative ethics brings with it a number of advantages. Even so, instrumentalist approaches to responsibility that are not grounded in an antecedent commitment to consequentialism raise new puzzles. First, holding that there is a potentially autonomous normative domain of responsibility that is discreet from questions about normative ethics invites questions about the relationship between those normative domains, and whether one trumps or constrains the other. Second, the introduction of a free-​standing consequentialist account does not obviously block the possibility that one could develop a distinct account of the normative foundations of responsibility grounded in some other view of normative ethics. How these potentially competing and non-​overlapping accounts might interact raises difficult questions for the non-​consequentialist instrumentalist about responsibility. To better understand the current state of play—​and to better see what progress remains to be made—​it will be helpful to first canvass some of the historical origins of the approach.

3  To be sure, important versions of contemporary instrumentalist accounts are explicitly wedded to a broader consequentialism about normative ethics (e.g., McGeer 2015; Miller 2014a, 2014b). The point here is about the resources for instrumentalist accounts that aren’t wedded to consequentialism.

Instrumentalist Theories of Moral Responsibility    7

3.  Classical Consequentialist Theories If one thinks of Strawson’s “Freedom and Resentment” as ushering in the contemporary era of work on moral responsibility, it is perhaps fair to say that for most of that post-​Strawsonian history consequentialist theories of responsibility have been treated as bogeymen—​something seldom clearly seen, but universally regarded as bad. One source of this reputation was rooted in a prominent reading of Strawson’s “Freedom and Resentment.” Strawson’s so-​called optimist was committed to a “one-​eyed consequentialism” that understood responsibility as a matter of the efficacy of moral influence. Consequentialists missed our concern for the quality of will evinced in action, and failed to recognize the central role of the reactive attitudes in shaping our practices. Given the formidable grip that consequentialism exercised over the imagination of some philosophers in the first half of 20th-​century analytic philosophy, it is tempting to portray the main theoretical options in the theory of responsibility as a choice between some or another version of consequentialism about responsibility, on the one hand, and metaphysical libertarianism on the other (Cf. Moore 1903; Schlick 1939; Nowell-​Smith 1948). This is accurate enough, but it also overplays the role of consequentialism in that period. The bulk of the pre-​Strawsonian literature tended to focus less on the normative foundations of responsibility and more on the meaning and metaphysics of ability talk. Consequentialist considerations were oftentimes in the background. However, in many of the discussions of that era, the philosophical heavy-​lifting of compatibilism was performed by invocations of the supposed ordinary meanings of “can,” “action,” and “volition” and by appeal to various broadly Humean pictures of freedom and responsibility (for examples, see Hook 1958; Pears 1963; and Dworkin 1970). There were exceptions where consequentialism, rather than Humeanism about freedom and action, figured more prominently in the explanatory approach. For example, Schlick (1939) claimed that clarity about responsibility can be gotten by looking to the actual role our concept plays in ordinary life. On his account, our interest in responsibility and punishment is, at bottom, about shaping dispositions. We blame and punish to shape motives in ways that are designed to prevent or call forth certain acts. Thus, we don’t punish agents for whom those punishments have no plausible influence on that agent’s motives (1939, 151–​154). Where some saw the consequentialist motive as built into the explicit aspirations or intentions of a blamer or punisher (e.g., Nowell-​Smith 1948), Schlick’s analysis suggests that the consequentialism is at the level of something like the functional role of responsibility attributions. For Schlick, we can understand the nature of responsibility by the role of the concept in our practices. Individual cases of blaming and punishing, it seems, might not be undertaken with a teleological end. Indeed, Schlick gives an error theory for how that might be so (154–​158). Other theories held that individual tokens of blaming were always aiming at character modification. A striking version of this is Nowell-​Smith’s (1948) view. Nowell-​Smith

8   Manuel Vargas was the famous “optimist” of Strawson’s “Freedom and Resentment.”4 Echoing some of the ideas in Schlick, Nowell-​Smith holds that the crucial idea for responsibility is that “the class of actions generally agreed to be voluntary coincides roughly with the class of actions that are caused by characteristics that can be strengthened or inhibited by praise and blame” (1948, 56). And, like Schlick, he quickly moves from the loose fit between blame and character modification to talk of punishment. On his view, this picture of the domain of the voluntary and its susceptibility to influence entails a utilitarian theory of punishment. However, unlike Schlick, Nowell-​Smith holds that the intention of the punisher (and, it seems, the blamer) in so punishing just is to bring about particular effects. Failures to achieve those results don’t show that punishment lacks instrumental intent, but rather, those failures simply speak to the “lack of skill in the practitioner” (56). A third version of classical consequentialism about responsibility worth highlighting is J.J.C. Smart’s “Free Will, Praise, and Blame” (1961). Given that its publication date is close to the publication of “Freedom and Resentment,” it seems unlikely that Smart was one of Strawson’s targets, although his account is vulnerable to some of Strawson’s objections. Even so, Smart’s proposal has continued to enjoy some visibility, in part because of its distinctive character and its curious ambiguity about whether it is a theory of moral responsibility at all. Akin to Schlick, Smart presents his view as getting at the conceptual role of responsibility practices. However, unlike Schlick, Smart finds that ordinary practices are a confused mix of tenable and untenable features. In particular, moral blame involves a picture of desert-​entailing judgments that Smart regards as metaphysically confused. So, he advocates a relatively radical recasting of responsibility practices, jettisoning desert-​entailing judgments of blame for a notion of “dispraise” that amounts to a kind of “grading” of the moral quality of the action. Smart’s account can be championed as both a compatibilist and incompatibilist account of responsibility.5 Which way we read it partly depends on what we think fixes the 4 

Strawson’s presentation of his optimist reads as though it was supposed to be about a general class of views, rather than any particular figure. And, indeed, in that time period the kind of view he was gesturing at would have been recognizable to his audience. However, it is unclear how much he was really gesturing at a class of views and how much he was simply pointing to Nowell-​Smith. After all, he twice footnotes Nowell-​Smith (1948, 1954) as a proponent of the view. Given the stingy citation practices of the era, this is notable both in its repeated specificity but also because those footnotes count for a plurality of the total citations in the piece, and a majority of citations if we discount a passing reference to Hume. 5  Arneson calls Smart’s view “hard soft determinism,” and reads him as a revisionist compatibilist (2003, 233). Talbert (2016) and McKenna and Pereboom (2016), among others, read Smart as a compatibilist as well. However, the analysis that McKenna and Pereboom offer suggests one reason for reading Smart as an incompatibilist. McKenna and Pereboom claim that “Incompatibilists would not regard the control required for moral responsibility in [the sense employed by Smart and others] to be incompatible with determinism, and thus it is open to free will skeptics to endorse these senses” (2016, 263; Cf. Pereboom 2014, 125–​138). Notice that this suggests that, by McKenna and Pereboom’s reading of incompatibilism (and contrary to how they label him), Smart is more plausibly read as an incompatibilist. This should be unsurprising: Smart’s interest in an exclusively forward-​looking account of responsibility is introduced as a consequence of an antecedent conclusion that desert-​invoking responsibility is incompatible with determinism.

Instrumentalist Theories of Moral Responsibility    9 target for something being a theory of moral responsibility. One could hold that what makes a theory about responsibility is that it does a satisfactory job of capturing our concepts, thoughts, or beliefs about the term “responsibility.” (The notion of “capturing” admits of variation, and we needn’t fix a specific conception. For some, capturing involves descriptive accuracy; for others, it will emphasize rational or normative appeal.) This conceptualist view animates many incompatibilist views, and at least some compatibilist accounts as well.6 On a conceptualist picture of responsibility, one according to which desert talk is one of the elements essential to something being a theory of moral responsibility, Smart’s positive account is no theory of responsibility at all.7 In contrast, a conceptualist who thinks that desert is not essential to responsibility (McKenna 2012), can agree that Smart’s account may be read as a form of compatibilism. His rejection of desert is no barrier to capturing whatever other theoretical features figure in the demarcation of the theory being about moral responsibility. However, there is another way of thinking about the proper target of a theory of moral responsibility, according to which Smart’s account may be read as a compatibilist one. On a phenomenalist view about responsibility, what makes something a theory of responsibility is not its capturing of some privileged concepts, thoughts, or beliefs, but instead, its capture of the phenomena out in the world (and in our psychologies) that we take to figure prominently in our moralized praising and blaming.8 On such accounts, what makes something a theory of moral responsibility is that it is about the psychology and social economy of blame, and in general, the stuff of our holding one another responsible.9 According to the phenomenalist about responsibility, the compatibility debate is really a debate about whether our existing practices—​things which are typically demonstratively specifiable—​retain their normative and rational integrity if determinism is true. To the extent to which Smart captures those things—​the phenomena out 6  It is especially evident, for example, in views that hold that what makes a theory of responsibility a theory of responsibility is that it captures some (inevitably theoretically and not demonstratively specified) notion of desert (e.g., Pereboom 2001, 2017; Caruso and Morris 2017) or self-​creation (G. Strawson, 1994), properties that do not readily submit to ostension. 7  Critical reactions to Smart often seem to presuppose an antecedent commitment to a conceptualist construal of responsibility. Watson reads the general reaction to Smart’s proposal as it being “so off-​ centre that they in effect change the subject” (Watson 2003, 24). Similarly, Pereboom maintains that any view of responsibility that does not include desert at the center is simply not at issue in the free will debate (2017, 260). To my mind, these claims are most naturally read as expressing a conceptualist understanding of the target notion in the debate, although this reading is not incontestable. 8  For arguments in favor of what I am here calling the phenomenalist view about responsibility, see Vargas (2004, 2015, 2017, forthcoming). This approach builds on insights from referentialist semantics, including the idea that we can isolate some reference-​fixing thing—​something that does not work exclusively through our thoughts—​to fix reference. A proposal about that thing—​e.g., the “work of the concept”—​provides an independent specification of a conceptual role or functional characterization that captures the stakes in debates about free will and responsibility. Individual theories about responsibility are thus construed as candidates for what best fits that role or characterization. 9  The phenomenalist view received its canonical statement in Strawson’s (1962) “Freedom and Resentment,” and in his admonition to avoid “over-​intellectualization” of the phenomenon, and the corresponding injunction to attend to the moral psychology and social economy of praise and blame.

10   Manuel Vargas in the world that one takes to be the subject of a theory of responsibility—​then Smart’s account would count as a compatibilist theory of moral responsibility. However, Smart’s shift to dispraise, a shift away from blame, means that his theory is no orthodox version of compatibilism according to which our practices remain largely intact. Rather, it amounts to a revisionist compatibilism that seeks to alter our practices.10 Although the interpretive issues surrounding Smart’s account are interesting in their own right, it also prefigures a number of debates that recur in the contemporary literature. Among those issues are the putative centrality of desert, the basis on which we count something as a theory of responsibility, and how we should understand what is central to debates about moral responsibility. This section has described some of the main features of consequentialist approaches to responsibility that dominated the pre-​Strawsonian Anglophone philosophical literature on moral responsibility. However, any account of this period would be remiss without a note about Strawson’s “Freedom and Resentment.” More than anything else, Strawson’s essay undermined the credibility of consequentialist approaches to moral responsibility in the Anglophone philosophical literature. It is perhaps no small irony that a number of recent readers of Strawson’s work—​all members of the recent resurgence of instrumentalist approaches—​have suggested that core features of Strawson’s account can be accommodated within a consequentialist framework (Vargas 2008, 2013a,b), or even that Strawson’s original account, properly understood, just is a defense of a more nuanced form of consequentialism about responsibility (McGeer 2014; Miller 2014a). If the consequentialist-​friendly reading of Strawson is correct, then Strawson’s legacy has been, in some sense, an undoing of the very kind of theory he may have set out to defend.

4.  Contemporary Approaches At the start of the 21st century, consequentialist approaches were mostly regarded as nonstarters. As Arneson put it, “This is the position everyone loves to hate” (2003, 233). Consequentialist theories were held to be (1) extensionally defective, unable to distinguish between responsible agents and non-​responsible agents (for example, adults, children, and dogs might all be influenceable by blame); (2) unable to distinguish between blame and other means of influencing agents; (3) products of confusing our judgments of responsibility with the appropriateness of their expression (Scanlon 1988); (4) unable to accommodate the distinctive role of blame (Strawson 1962, Bennett 1980); (5) unable to account for backward-​looking blame (Dworkin 1986); and (6) infelicitously saddled with consequentialism about normative ethics.

10 

On Smart’s relationship to revisionist approaches to free will and moral responsibility, see Vargas (2011); for more on revisionism in general, see McCormick (2016) and Vargas (forthcoming).

Instrumentalist Theories of Moral Responsibility    11 The rehabilitation of instrumentalist approaches to responsibility has happened in no small part because proponents of these views have tended to find those objections surmountable. Moreover, the recent trend of rereading Strawson’s own account as friendly to instrumentalist approaches (as in McGeer 2014 and Miller 2014a) has cast some of the traditional concerns about these approaches in new light. Rather than recapitulating the back and forth of those historical objections—​for discussions more in that vein, see Arneson (2003); Vargas (2008, 2013) and Miller (2014b)—​this section focuses on five post-​Strawsonian developments that have loomed large in different parts of the instrumentalist literature: the separation from consequentialist normative ethics; the idea of blame as building out or enhancing the moral or rational powers of agents; a more elaborate theory of exemptions; the development of revisionism as an explicit methodological position; and the invocation of a distinction between questions internal and external to a theory of moral responsibility. First, because it figures prominently in what follows, it bears repeating that contemporary instrumentalist accounts have tended to emphasize that they can be (and often are) disentangled from consequentialist theories of normative ethics (e.g., Arneson 2003, 243–​246; Vargas 2008, 2013a). Even so, instrumentalist accounts have benefited from making use of the considerable resources of contemporary consequentialist theorizing. For example, instrumentalist rehabilitations have appealed to both act and rule-​consequentialist formulations for the justification of responsibility practices, to indirect consequentialism, and to familiar consequentialist replies to scapegoating worries. On Arneson’s (2003) account, the appeal of responsibility practices is found in it generally being the case that holding people responsible (even in backward-​looking ways) plausibly tends to produce desirable results. These practices plausibly gain greater efficacy from us restricting our reactive responses—​our judging responses, as he puts it—​to wrongdoing of the relevant sort, by the relevant kinds of agents. According to Vargas’s (2013) account, whether someone is responsible or not is settled by rules internal to what he calls “the responsibility system”—​those judgments, practices, and attitudes concerned with the worthiness of moralized praise and blame. That practice can make extensive use of backward-​looking assessments, and it may employ notions of desert. What makes that set of practices normatively appealing or justified, however, is that it is conducive to securing the (putatively valuable) goal of fostering and refining our ability to recognize and respond to moral considerations. Victoria McGeer’s (2013, 2015, 2019) formulation of instrumentalism holds that our practices of moralized praising and blaming are normatively grounded in whether so blaming enhances the wrongdoer’s ability to be suitably responsive to moral reasons (see also McGeer and Pettit 2015). Our capacity to recognize and respond to moral considerations is an elastic, socially scaffolded capacity that relies on our blaming in order to build out that capacity (see also Fricker 2016; Jefferson 2019). A second and related development has been greater attention to the way blaming practices build out the moral powers of agents through a process of “responsibilizing” (McGeer 2013), “agency cultivation” (Vargas 2013), or “prolepsis” (Fricker 2016; McGeer

12   Manuel Vargas 2019). For many instrumentalists, a central function and justification for responsibility practices is that these practices enhance the capacity of those agents to recognize and respond to relevant moral considerations, broadly proleptic effects are unavoidably central to these accounts. Blame is proleptic inasmuch as the effect of blame leads suitable agents to have reasons or rational capacities that were either not had or not actively recognized by the agent in the moment of wrongdoing. This is the sense in which these accounts focus on cultivating a particular kind of agency though a process of “responsibilizing” practices of moral praise and blame. However, instrumentalists differ on where to locate proleptic effects. Many instrumentalist accounts emphasize the proleptic aspect in individual instances of blaming (as in McGeer and Pettit 2015; Fricker 2016; McGeer 2019; Jefferson 2019); some emphasize its systemic effects (as in Vargas 2013, 2020). Although most contemporary instrumentalist accounts avail themselves of the idea of blame that aims to enhance moral considerations-​ sensitive agency, proleptic effects are plausibly more pronounced on accounts that emphasize individualized prolepsis, and less pronounced on accounts that appeal to the systemic effects of blame on groups of agents, or that otherwise limit the scope of prolepsis to a class of independently specified responsible (or non-​exempt) agents. This latter thought, about a potentially independent basis of candidates for responsibility practices, has been the subject of a third notable development in the literature. Instrumentalist accounts have taken more seriously the need to provide an account of responsible agency, that is, an account that identifies the kinds of agents that are proper candidates for our responsibility practices, attitudes, and judgments. (Alternately, one may think of this as a theory of exemptions, if we use Watson’s (1987) terminology.) An articulated theory of responsible agency is important for instrumentalist approaches, in part because one traditional challenge to these accounts was that the extension of the influenceable and the extension of responsible agents was not the same. One solution is to simply invoke an independent account of responsible agency (as in Vargas 2013, which appeals to a form of reasons-​responsiveness). However, as McGeer (2015) has noted, this raises concerns about the basis of that independence. A different strategy is to accept that responsible agency involves a tight connection to agency susceptible to moral influence, but to then shore up this account by specifying the nature of that susceptibility to influence—​for example, as susceptibility to the specifically reactive attitudes, or to the moral import of the expression of reactive attitudes. Concerns about extensional adequacy and the costs of revisionism on this issue remains a matter of ongoing discussion (Jefferson 2019). Given the foregoing, it is perhaps unsurprising that a fourth development that has shaped contemporary instrumentalist accounts is the gradual emergence of an explicitly revisionist approach to moral responsibility. On one standard definition of revisionism about moral responsibility, a theory of moral responsibility is revisionist if the truth of the theory’s account of moral responsibility is in conflict with common-​sense views about that thing.11 The appeal of revisionism for instrumentalists should be apparent: If 11 

Compare McCormick’s formulation of revisionism: “Revisionism is the view that we can and should distinguish between what we think about moral responsibility and what we ought to think about it, that

Instrumentalist Theories of Moral Responsibility    13 one has a principled basis for revisionism, it may offer a principled solution to lingering gaps between the extension of ordinary judgments about moral responsibility and the instrumentalist’s account of responsibility. The core of this idea was clearly present in Smart’s (1961) account, and it has been picked up and developed in the current crop of instrumentalist theories. For example, Arneson claims that instrumentalism “has to be regarded as a substitute for the ordinary idea of responsibility . . . A precondition of finding influenceability acceptable is having good grounds for finding the ordinary notion of responsibility unsustainable. So the fact that influenceability does not mesh perfectly with the ordinary notion of responsibility is not per se an objection to it” (2003, 249). Versions of revisionism about free will and/​or moral responsibility can also be found in the instrumentalist accounts endorsed by Vargas (2013), McGeer (2015), and Jefferson (2019). Of course, one need not be a revisionist to be an instrumentalist, and there are revisionists who are not obviously instrumentalists (Nichols 2015; Doris 2015a). Smart, Arneson, and Vargas are all explicit that their revisionism is motivated by the thought that folk thinking about responsibility is at least partly committed to an implausible libertarian metaphysics of agency. One need not be a revisionist to be an instrumentalist, though. One might think that folk understandings of responsibility have always been compatibilist (this is perhaps the traditional approach to compatibilism—​ see Vargas 2011). On that account, instrumentalism just is an account of the normative foundations of the practices as we have them. Partly connected to the question of whether various instrumentalist accounts are revisionist, there is the matter of what is to be revised. Vargas’s indirect, two-​tiered account functions so as to limit or avoid the need for notable revisions in our ordinary practices. The revisions in his account are primarily at the level of the kinds of things we think and believe about the basis of our responsibility practices. In contrast, more direct forms of instrumentalism tend to more readily accept the need for important revisions in our practices, that is, in what we do (as in Arneson 2003; McGeer 2013; Jefferson 2019, and if we count his positive proposal as revisionist, especially so in the case of Pereboom’s 2014 proposal). A fifth and final innovation worth noting is that some instrumentalists have employed a distinction between the justification of responsibility norms and their content. Something like this distinction is suggested in “Freedom and Resentment,” in Strawson’s distinction between questions internal and external to the responsibility practice. Building on later remarks by Strawson (1985), Dale Miller (2014a) has argued that the idea that instrumental concerns are the wrong kind of answer to questions about responsibility was intended by Strawson as an invocation of an idea in Carnap, namely, the idea that there are some questions whose answers are internal to a framework, and that it is a separate question to ask about the appeal of having that framework. the former is in some important sense implausible and conflicts with the latter, and so we should revise our concept accordingly” (2016, 109). For overviews of revisionism, see McCormick (2016); McKenna and Pereboom (2016 286–​293), and Vargas (forthcoming).

14   Manuel Vargas The internal/​external to the framework idea seemed to be in the air at the time, with a related idea figuring in the work of Rawls (1955) and HLA Hart (1959). On their accounts, we ought to distinguish between the justification of rules and institutions (which may be given in instrumental terms) and the content of the rules or the norms of the institutions (which may be backward-​looking, desert-​based, or otherwise non-​instrumental in character). Vargas (2013; 2015), in particular, has made use of this distinction in distinguishing between the instrumental character of the responsibility system, and the potentially desert-​based, backward-​looking character of first-​order (or substantive) responsibility norms. This section canvassed five ideas that have figured prominently in contemporary instrumentalist approaches—​the independence of instrumentalism from consequentialism; the agency cultivating or proleptic aspect of responsibility practices; more attention to responsible agency; the evolution of revisionist theorizing; and arguably greater clarity about questions internal and external to frameworks. As we will see in the next section, these ideas tend to figure in how contemporary instrumentalist accounts respond to various standard challenges.

5.  Objections and Debates In this section, I discuss several standard objections and ongoing debates for instrumentalist approaches. These include the question of scapegoating, objections that instrumentalists are relying on the wrong kind of reason, concerns about self-​ effacement, and the objection that instrumentalists are unable to capture the notion of desert that figures in debates about free will and moral responsibility.

5.1  Scapegoating One traditional complaint about consequentialist theories of ethics is that they seem to support scapegoating, or the punishing of people who are innocent. One might wonder whether instrumentalist approaches to responsibility are vulnerable to a parallel complaint. Whatever the prospects are for consequentialist theories of normative ethics, instrumentalist accounts have good resources for deflecting concerns about the equivalent of scapegoating in a theory of moral responsibility—​in this case, blaming the non-​responsible when it is expedient to do so. Two replies, at least, are available: the definitional response and the modularity response. They can be employed jointly or individually, depending on the particulars of the account. The definitional response to scapegoating is well-​expressed by Arneson: “one is responsible for an act if one did it and doing of this sort are influenceable by blaming or

Instrumentalist Theories of Moral Responsibility    15 punishing. One cannot squeeze hard on this admittedly thin notion of responsibility to somehow induce it to imply that one can be responsible for a crime one did not commit, because one’s doing it is by definition required for responsibility” (2003, 245). One might put pressure on Arneson’s formulation in various ways. For example, depending on how one thinks of an agent’s doings, omissions could be a problem for the idea that one’s doing is required for responsibility. Still, the underlying point is plausible: Whatever one thinks responsibility is, it requires that some outcome be suitably related to the agent’s behavior. Adding to that thought the further thought that the agent must be able to be influenced by blame, or that it must be the case that rules licensing blame must generally produce better responsiveness to moral considerations, does not eliminate that conceptual or definitional point. Of course, one can ask why that conceptual constraint? To this, the instrumentalist might reply that this new question is no longer a question about when someone is responsible. Rather, it is a question about why we should have a notion of responsibility at all. Perhaps we do better to give up responsibility and go in for some other thing, something that does an even better job at achieving the instrumentalist’s end. However, if this is the question, it is no longer the instrumentalist who can be accused of changing the subject. Thus, the instrumentalist may demur, insisting that her account is merely an account of responsibility, and that it is beyond her ambition to offer an account of what collection of practices, all things considered, best serves the identified instrumental end. I leave it to others to decide whether this is a satisfactory response. Whatever the answer might be, there are further things that the instrumentalist can say about these issues. In response to the “why responsibility and why not something else?” question, the instrumentalist can appeal to some of the same considerations suggested by Strawson in “Freedom and Resentment.” It might be that, for example, given our psychologies, no nearby arrangement of practices would do. Alternately, it might simply be a matter of feasibility—​perhaps there are normatively more appealing arrangements, but they are not readily available to us. If so, the most we can effectively do is to militate for modest adjustments of the practices we find. Or, one might be moved by epistemic concerns, citing concerns about unintended consequences in transforming practices in some more radical way. This might seem especially plausible if one was inclined to think that the underlying affective psychology was ineradicably entangled with other things—​love, human affection, social coordination, and so on. If so, the question of the comparative appeal of this set of practices [including responsibility and its entanglements] and that set [devoid of responsibility along with transformations in the things with which it is entangled] cannot be readily answered. One might even hold, as Strawson seems to have, that the only answers that can be given are always internal to some or another set of value-​fixing affective orientations. The point of the foregoing remarks is not to take a stand on how that conversation should go. It is simply to note that it is not obvious that the instrumentalist faces a unique or distinctive challenge in employing the definitional response to the scapegoating objection. If the instrumentalist elects to take up subsequent “metanormative”

16   Manuel Vargas questions about those conceptual constraints, there are widely utilized ideas to which she might appeal to explain why we should embrace responsibility, rather than joining the eliminativist in calling for its ejection from our moral lives.12 The modularity response was already highlighted in section 2, without the benefit of a label. The idea is as follows. A theory of moral responsibility is a theory of moral responsibility, and not some other thing. To be an instrumentalist about moral responsibility need not entail that one is a consequentialist about normative ethics. Some instrumentalist accounts forge a relatively tight link between susceptibility to the influence of moral blame and being blameworthy (e.g., Miller 2014a; McGeer 2019; Jefferson 2019). Others do not (Vargas 2013). However strong that link, the normative appeal of scapegoating—​blaming those who are not blameworthy; or even, expanding the scope of the putatively blameworthy to those whom it is simply expedient to blame—​depends on one’s wider philosophical commitments. If one accepts consequentialism, then it was already normatively appealing to think that when the payoffs are high enough, it may do to blame those who are not blameworthy (or in general, to expand the class of the blameworthy to the expediently blamed). Similarly, if one’s normative ethics funds a prohibition against blaming those for whom it is merely expedient to do so—​suppose that agents should always be treated as ends in themselves, and not merely as means—​then one already has the resources for explaining the impermissibility of scapegoating the non-​responsible. These prohibitions do not disappear because one’s theory of moral responsibility is justified in instrumentalist terms. Instead, the normative ethical theory can be understood as providing side constraints on the scope of the instrumental reasons.

5.2  The Wrong Kind of Reasons In a famous passage in “Freedom and Resentment,” Strawson imagines the instrumentalist’s interlocutor saying “the only reason you have given for the practices of moral condemnation and punishment in cases where this freedom is present is the efficacy of these practices in regulating behavior in socially desirable ways. But this is not a sufficient basis, it is not even the right sort of basis, for these practices as we understand them” (1962, 4). This observation is sometimes taken as the locus classicus of the “wrong kind of reasons” objection against instrumentalist theories. It prominently figures in Darwall’s 12 

Here, the issue of revisionism lurches into the debate again. One may hold that there are some benefits for moralized praise and blame, so long as they are denuded of any commitments to, say, libertarian conceptions of free will. Such views have typically been taken up under the banner of offering explicitly revisionist theories of moral responsibility which aim to (sometimes selectively) purge commitments that involve incompatibilism, implausible psychologies, or overly demanding commitments on desert within our conceptual or practical lives (e.g., Smart 1961; Hurley 2003; Vargas 2013a; McCormick 2015; Nichols 2015; Doris 2015a). As I understand it, the positive account in Pereboom (2014), but perhaps not Pereboom (2001), is best construed as an instance of this approach—​albeit one distinguished by its eschewal of desert. More about this below.

Instrumentalist Theories of Moral Responsibility    17 (2006) claim that “Desirability is a reason of the wrong kind to warrant the attitudes and actions in which holding someone responsible consists in their own terms” (15). “Strawson’s Point,” as Darwall calls it, is supposed to identify a fatal problem for instrumentalist theories.13 It bears noting that immediately after introducing the pessimist’s accusation that the optimist is relying on the wrong kind of reason, Strawson makes it clear that his project is to give the optimist more to say. On the face of it, Strawson did not seem to think the objection was fatal to instrumentalism (Cf. McGeer 2014, Miller 2014a). Looking at the details of Strawson’s account suggests one reason for thinking that the wrong kind of reasons objection can be met by the instrumentalist. Recall the idea, already noted, that Strawson had a roughly Carnapian picture of responsibility practices. There are questions internal to the practice, and there are questions external to the practice. On this reading, what makes instrumental considerations the wrong kind of reason is that they raise questions external to the practice. For Strawson, the idea that social efficacy is not the right basis for the practice is a point internal to the practice. If instrumental reasons have a place, it is external to the practice. This is, of course, exactly what at least some contemporary instrumentalists maintain (Vargas 2008, 2013, 2015; Miller 2014a). Here is what the “wrong kind of reasons” objection plausibly gets right: Consequences are mostly irrelevant to the propriety conditions of blame, gratitude, and resentment. Each of these things has its own aptness or propriety conditions, conditions that don’t appeal to consequences. Consequences may speak to whether it is advantageous to express our blame, gratitude, or resentment, but not to whether those attitudes are merited, deserved, or proper. These things are, for Strawson, in large part structured by a relatively fixed set of reactive dispositions that are keyed to perceptions of whether others have exercised due concern or adequate quality of will. One can accept all of the preceding, while allowing that instrumental questions might arise at a different order of inquiry. Many norm-​structured practices have exactly this structure to them. Foul calls in a sport are typically justified by the thought that the safety of the players must be preserved, but this must be balanced with the enjoyment of spectators and players in the flow of the game. However, whether a particular play is a foul or not is clearly not settled by appeal to those framework questions. They are settled internal to the framework, by appeal to the rules of the game. In the context of the game—​within the framework of participants, one might say—​questions of safety and enjoyment are irrelevant to whether something is a foul (unless the rule specifies that those are considerations). We can, of course, step outside those rules and ask if we want to have those rules, whether there is a better set of rules available to us, and whether we

13  For a helpful discussion of why one might doubt that Darwall’s account has adequate resources for motivating a wrong kind of reasons objection against instrumentalists, independent of Strawsonian concerns, see Miller (2014a). Even if we accept Darwall’s account of the closed “second-​personal” circle of accountability, the considerations in the present section suggest reasons that this picture is compatible with some forms of instrumentalism.

18   Manuel Vargas have reason to allow ourselves to be bound by those rules. But the rules are the rules, at least until we change them. Can the instrumentalist exploit this idea to make sense of the role of instrumental considerations? Strawson entertains the idea that we might step outside the “reactive” or participant framework, and take up a standpoint of objectivity, but he was cautious about the possibility of undertaking normative inquiries external to the participant stance.14 The suggestion in “Freedom and Resentment” is that it is unclear where we might stand if we take up framework questions, and on what basis we might adopt a different set of attitudes and practices. Here, though, instrumentalists may depart from Strawson in his holding that, for example, normative questions bottom out in our attitudes; or that the attitudes that ground normative matters rise and fall together; or that one cannot privilege some of these attitudes, and not others; or that one might identify some axiological notions as the basis on which to evaluate various attitudes and practices. An instrumentalist who appeals to non-​Strawsonian views on any of these matters might find grounds to insist that there is some place we can stand outside the practice of responsibility to assess whether we might retain it, reject it, or revise it. What the proponent of the wrong kind of reasons objection needs is for it to be impossible to intelligibly take up questions about normative frameworks that are external to that framework, or at least, some special reason to think one cannot do so in the context of moral responsibility. That many of our normative practices permit these sorts of questions, and that at least some instrumentalist accounts mimic that structure, suggests that the wrong kind of reasons objection is, so far, the wrong kind of objection to make against indirect or two-​tiered instrumentalist theories of moral responsibility. The existing wrong kinds of reasons objection is most promising against direct and classical consequentialist accounts of the sort discussed in section 3, above. Instrumentalisms that recognize a difference between questions internal and external to a framework, or that can distinguish between the contents and the justifications of practices and institutions (Cf. Rawls 1955; Hart 1959) have resources for deflecting the concern.

5.3  Self-​Effacement The foregoing remarks can give rise to concerns about self-​effacement (Doris 2015b; Miller 2014b). There are different ways to put the concern, but in the abstract the idea is that instrumentalism, particularly those that allow for indirectness at the level of first-​ order practices, can produce an unacceptable bifurcation in the thinking of blamers. The worry is that blamers must employ a set of first-​order principles about blame, while 14 

In general, the question of normativity in Strawson is an elusive one (cf. Watson 1987).

Instrumentalist Theories of Moral Responsibility    19 simultaneously recognizing that there is a second-​order set of concerns about whether these principles ought to be amended or otherwise adjusted. This concern has taken various forms, including skepticism about the psychological tenability of accepting two orders of concern and the risk that this sort of “divided” thinking might de-​motivate the efficacy of the first-​order commitments (Williams 1988). The details vary by instrumentalist theory, but one view to which instrumentalists can avail themselves is the idea that first-​order principles, beliefs, and judgments may be most effective when they are intuitive, routinized, and habitual (Arneson 2003). Routinized or habituated principles will tend to shape one’s dispositions for perceiving what is morally significant (Vargas 2013; Miller 2014b; McGeer 2019). If the locution can be forgiven, we can begin by noting that the Williams-​style worry about self-​effacement has two faces. There is the point about the psychological tenability of two-​level thinking, and a worry about its motivational efficacy. Experience speaks to its tenability. As John Doris has observed, it does not seem impossible for one to think that one shoots a three pointer and to also hold “the justification of participating in a practice where things are done for the sake of winning games is that this participation ethically improves the participant” (2015b, 2632). On the matter of motivational efficacy, the issue is that awareness of the possibility of second-​order concerns will impair the ordinary efficacy of otherwise entrenched habits of mind. This worry is especially acute for act-​consequentialist style versions of instrumentalism where the propriety of blaming in any instance depends on its effects. First, it is not obvious that instrumentalists must require that agents be aware of the distinction between the first-​order norms of responsibility holding and the normative foundations of that practice. Further, it is simply unclear that much follows from awareness of this distinction. Whether this awareness would in any interesting way alter the character of the first-​order norms (or the typical person’s relationship to them, for that matter) seems to be an entirely empirical question. Second, Dale Miller (2014b) has argued that Strawson’s account of the reactive attitudes provides resources for instrumentalists in this context. To the extent to which instrumentalist accounts are relying on first-​order judgments and principles that correspond with the alignment of our affects—​a matter about which many contemporary instrumentalists are explicitly committed—​then the frame of mind of the instrumentalist is one where the reactive attitudes are governing in their usual way. If there is any special demand on the mind of the instrumentalist, it may be as minimal as resolving to inhabit those attitudes, and to let them fully become habits of mind, subject to disavowal only under isolated conditions. As Strawson himself emphasizes, we can step back from our attitudes and ask questions about them, and whether they might admit of alteration. That is, however, a different frame of mind than the frame of mind where one is operating within the space of the attitudes as they present themselves to us. Of course, radical departures from the dispositions of our reactive attitudes may be difficult to sustain, but the pressure toward conservatism about revision tends to be recognized by many contemporary instrumentalists.

20   Manuel Vargas

5.4  Desert What about desert? Some philosophers have allowed that there may be multiple senses of responsibility, but that the one that figures in debates about free will and moral responsibility concerns some or another version of the idea of desert (Pereboom 2001, 2014, 2017; Caruso and Morris 2017). However, desert is a notion with no central role in consequentialist theories of ethics. So, one might think, it is difficult to see how desert might operate in instrumentalist theories that emphasize the consequences of the practice in establishing their justification. Pereboom has argued that instrumentalists of various stripes do not offer basic desert theories of responsibility (2014, 2; 2017, 260). Perhaps this is true of some instrumentalists (e.g., Dennett 1984; Arneson 2003), but at least some instrumentalist theories have resources for capturing the operative sense(s) of desert. According to Pereboom’s justly influential account, the notion of desert that is at stake in debates about free will is something he labels basic desert, where the operative notion of desert is one according to which an agent “would deserve to be blamed or praised just because she has performed the action, given an understanding of its moral status, and not, for example, merely by virtue of consequentialist or contractualist considerations” (2014, 2). The issues here are delicate, but the general contours of the main options for the instrumentalist are relatively clear. First, it is open to instrumentalists (and others) to reject the idea that desert is essential to moral responsibility.15 Second, instrumentalists might accept the basic desert conception, and argue that it can be met. Third, instrumentalists might accept that some notion of desert is involved in attributions of responsibility, but not the one that figures in Pereboom’s characterization of basic desert. In what follows, I’ll focus on the second and third approaches, as they have figured in most contemporary instrumentalist accounts. Few instrumentalist accounts have explicitly repudiated any notion of desert.16 Some accounts have maintained that the basic desert condition can be met by at least some instrumentalist accounts (Vargas 2015). To see why, we must first get clear on the dialectical burdens in this context. On the standard construal of basic desert, the idea is that desert is generated by the nature of the agent and the action, and blaming isn’t licensed by consequentialist or contractualist considerations. This characterization cannot, by itself, settle the question

15 

McKenna (2012), who is not an instrumentalist in the sense under discussion, has explored a version of this view. He ultimately argues that a desert-​invoking formulation is compatible with his account of responsibility. David Shoemaker (2015) has suggested that questions of desert are restricted to harsh treatment and not central to responsibility. 16  In recent work, Nelkin (2016) has argued that although “desert” has been used in a variety of ways, desert and accountability are mutually entailing. If this is right, then quite apart from whether we accept Pereboom’s specific account of basic desert, it may not be open to proponents of a theory of accountability to insist that desert has no role to play.

Instrumentalist Theories of Moral Responsibility    21 of whether responsibility is compatible with determinism or not. If basic desert is, from the outset, construed as an incompatibilist notion, it is a nonstarter in its primary dialectical function, namely, as a characterization of the notion of responsibility in the dispute. Of course, Pereboom and other putative incompatibilists think that the only way to secure basic desert, and thus moral responsibility in the relevant sense, is via some form of agency incompatible with determinism. Similarly, compatibilists who accept the basic desert characterization must think that their accounts can produce basic desert. The point here is that there is a further debate to be had, and that basic desert’s supposed entailment of incompatibilism cannot be definitional, on pain of basic desert failing as a neutral characterization of the stakes of the debate. So, can instrumentalists offer an account of responsibility that, on the face of it, meets the conditions of basic desert? Compatibilists who are not instrumentalists in the sense at stake here have offered such accounts. For example, David Brink maintains that desert is to be understood as the product of two independent variables, wrongdoing and culpability (2012, 498; see also Brink and Nelkin 2013).17 The issue is whether instrumentalists can help themselves to this sort of account, or something like it. Here, one’s variety of instrumentalism seems to matter. For instrumentalist accounts that entirely ground desert talk in consequences, Pereboom is right that such accounts fail to be basic desert accounts, because to deserve praise or blame would be determined “merely by virtue of consequentialist or contractualist considerations.” However, two-​ tier versions of instrumentalism seem not to run afoul of that prohibition. Recall that on a two-​tiered instrumentalist account, the norms, judgments, and attitudes that make up the responsibility system can be backward-​looking. On this sort of account, the propriety conditions for deserving blame can be the moral qualities of the agent and the act. Thus, the fact that the agent deserves blame is not merely by virtue of consequentialist considerations. Indeed, the two-​tiered theorist can allow that no judgments of blame are settled “merely by virtue of consequentialist considerations.” Instead, they are settled by, for example, whether the agent has acted wrongfully, and whether the agent was culpable in so acting. The incompatibilist might contest the instrumentalist’s construal of culpability and acting. Even so, the constraint that the deserving of blame can be established independent of its consequences can readily be satisfied by a two-​tiered instrumentalist theory. Here’s what’s the basic desert-​ers “no consequentialism” constraint rightly captures: There is a conception of desert—​one important to many of our social and, more specifically, moral practices—​according to which consequences are irrelevant. Trying to

17 

One further virtue of Brink’s account is that it also provides a compatibilist-​friendly way to capture the core retributivist idea, that punishment is proportional to that desert. So, incompatibilists convinced that the stakes of the responsibility debate must also involve a notion of desert or responsibility that supports retributivism (cf. Caruso and Morris 2017) cannot readily dismiss this version of compatibilism as failing to even join the putative debate. It is notable, however, that not all compatibilists have thought it a desideratum that any notion of desert captured by their account can support retributive practices, and practices of punishment (e.g., Scanlon 1998 and Wallace 1994).

22   Manuel Vargas capture that notion of desert with a conception of desert that is sensitive to consequences is bound to do considerable violence to this ordinary and central notion of desert (Doris 2015b). Inasmuch as a consequence-​insensitive notion of desert is implicated in attributions of moral responsibility, then at least one variety of instrumentalism about moral responsibility can capture this notion of desert. In doing so, the two-​tiered instrumentalist captures something of the flavor of Strawson’s observation that there are questions internal to a practice, and questions external to a practice. Questions about basic desert are questions internal to the practices of responsibility. On the two-​tiered version of instrumentalism, instrumentalism is constrained to the grounds we have for that practice as a whole, and so its internal-​to-​the-​practice account of desert can be compatible with basic desert. Are there other ways a (potentially non-​ two-​ tiered) instrumentalist might accept the demand for a basic desert notion of responsibility? Following the definitional strategy suggested by Arneson’s (2003) remarks about scapegoating, an instrumentalist could accept that basic desert is a conceptual requirement on responsibility, and that it operates on non-​consequentialist terms. So, the constraint of basic desert would be accepted on conceptual grounds, settled by the definitional features of responsibility, and not because of consequentialist considerations. It would then be open to the instrumentalist to add that the reason for keeping this package of conceptual commitments is that it produces instrumentally valuable results. This wouldn’t require endorsement of a two-​tiered instrumentalism, but it would raise some of the same questions that occur in the context of scapegoating, for example, raising questions about why we should accept basic desert practices vs. some other practice that does without them.18 It is unclear how far apart these strategies are from one another. What matters for present purposes is this: On the face of it, instrumentalist accounts of moral responsibility can satisfy basic desert construals of the philosophical stakes. Whether the resultant conception of basic desert has other unappealing features, or whether one accepts the implied account of culpability is a further matter. As noted at the outset of this subsection, there are other routes available to the instrumentalist seeking to address concerns about blame. Instrumentalists can, for example, accept that desert figures in responsibility attributions, but reject the claim that the stakes are basic desert. Would such a position mean that one is willfully ignoring the central philosophical dispute about moral responsibility? It need not. Recall the difference between conceptualist and phenomenalist construals of the philosophical stakes of a theory of responsibility. Many proponents of the importance of basic desert are most naturally interpreted as conceptualists, that is, endeavoring to capture our concepts, thoughts, or beliefs about the term “responsibility.” However, the instrumentalist can avail herself of the phenomenalist idea that this is a methodological mistake. Instead, on the phenomenalist reading, the proper philosophical stakes are about the

18 

Pereboom’s desert-​free account of responsibility (2014) might be understood as a contribution to this debate.

Instrumentalist Theories of Moral Responsibility    23 nature and normative integrity of our practices.19 These practices, she might say while gesturing at various phenomena in our social world—​and not some armchair stipulation of a metaphysics of desert—​are the subject of my account, and plausibly the subject of most accounts in the long history of philosophical debates about responsibility. For the phenomenalist, if it turns out that what we need to explain the nature, function, and normative integrity of our practices as we find them is some non-​basic notion of desert, then so much the worse for basic desert.20 A different but complementary strategy is to lean on the revisionist impulses that oftentimes accompany instrumentalist accounts. On this approach, one could grant that basic desert captures the ordinary understanding of desert. However, one can also maintain that some attenuated notion of desert is the proper successor notion (as in Smart 1961 and Arneson 2003) or that we can make do with forms of responsibility that do away with desert (as in Pereboom 2014). Although there are a number of paths open to instrumentalists in addressing concerns about basic desert, it is also true that one might place a special value in securing even more demanding forms of desert, including pre-​institutional or practice-​independent forms of desert, or desert of the “making sense of heaven and hell” variety (Strawson 1994). Whatever the attractions of these other forms of desert may be, they are less plausible as neutral characterizations of the stakes of the responsibility debates. They are more plausible as substantive accounts of more particular conceptions of desert, for which even the evaluative criteria for these notions—​for example, normative appeal, folk recognizability, conceptual importance, etc.—​remains a matter of robust theoretical dispute.

6.  Conclusion Instrumentalist theories of moral responsibility are enjoying something of a renaissance. Given that these accounts have resources for addressing the traditional worries raised against them, and given their recent proliferation in the literature, there is some

19  It is an error, I think, to reply that accounts of this sort make compatibilism too easy, or that they eliminate any substantive difference between compatibilism and hard incompatibilism (this thought is in the spirit of remarks in Pereboom 2014, 2–​3; 2017, 260). Nothing on phenomenalist construals of the debate rule out the possibility that our ordinary practices may presume that we have impossible or unlikely forms of agency, or that our practices are normatively indefensible. The former is claimed by some incompatibilists and many revisionists and disputed by conventional compatibilists. Disagreements about the latter is what separates typical revisionists and conventional compatibilists from eliminativists or hard incompatibilists. 20  I take it that something like this is the spirit of how McGeer and Funk (2017) think we should understand the significance of findings that show that our putatively retributive attitudes are sensitive to consequences.

24   Manuel Vargas reason to think that these accounts are (rightly or wrongly) once again an important approach for understanding the nature of moral responsibility.

Bibliography Arneson, R. J. (2003). The Smart Theory of Moral Responsibility and Desert. In S. Olsaretti (Ed.), Desert and Justice (pp. 233–​258). Oxford: Oxford University Press. Bennett, J. (1980). Accountability. In Z. Van Straaten (Ed.), Philosophical Subjects. New York: Clarendon. Brink, D. O. (2012). Retributivism and Legal Moralism. Ratio Juris, 25(4), 496–​512. Brink, D. O., & Nelkin, D. (2013). Fairness and the Architecture of Responsibility. Oxford Studies in Agency and Responsibility, 1, 284–​314. Caruso, G. D. (2012). Free Will and Consciousness: A Determinist Account of the Illusion of Free Will. Lanham, MD: Lexington Books. Caruso, G. (2015). Free Will Eliminativism: Reference, Error, and Phenomenology. Philosophical Studies, 172(10), 2823–​2833. Caruso, G. D., & Morris, S. G. (2017). Compatibilism and Retributive Desert Moral Responsibility: On What Is of Central Philosophical and Practical Importance. Erkenntnis, 82(4), 837–​855. Daniels, N. (1979). Wide Reflective Equilibrium and Theory Acceptance in Ethics. Journal of Philosophy, 76, 256–​282. Darwall, S. L. (2006). The Second-​Person Standpoint: Morality, Respect, and Accountability. Cambridge, MA: Harvard University Press. Dennett, D. (1984). Elbow Room. Cambridge: MIT Press. Doris, J. (2015a). Talking to Our Selves: Reflection, Ignorance, and Agency. Oxford: Oxford University Press. Doris, J. (2015b). Doing Without (Arguing About) Desert. Philosophical Studies, 172(10), 2625–​2634. Double, R. (1996). Metaphilosophy and Free Will. New York: Oxford University Press. Dworkin, G. (1986). Review of Elbow Room. Ethics, 96(2), 423–​425. Dworkin, G. (Ed.). (1970). Determinism, Free Will, and Moral Responsibility. Englewood Cliffs, NJ: Prentice-​Hall. Fricker, M. (2016). What’s the Point of Blame? A Paradigm Based Explanation. Nous, 50(1), 165–​183. Hart, H. L. A. (1959). Prolegomena to the Principles of Punishment. Proceedings of the Aristotelian Society New Series, 60, 1–​26. Hook, S. (Ed.). (1958). Determinism and Freedom in the Age of Modern Science. New York: New York University Press. Hurley, S. L. (2003). Justice, Luck, and Knowledge. Cambridge, MA: Harvard University Press. Jackson, F. (1998). From Metaphysics to Ethics: A Defense of Conceptual Analysis. New York: Oxford University Press. Jefferson, A. (2019). Instrumentalism About Responsibility Revisited. The Philosophical Quarterly, 69(276), 555–​573. McCormick, K. (2015). Companions in Innocence: Defending a New Methodological Assumption About Moral Responsibility. Philosophical Studies, 172(2), 515–​533.

Instrumentalist Theories of Moral Responsibility    25 McCormick, K. (2016). Revisionism. In K. Timpe, M. Griffith, & N. Levy (Eds.), Routledge Companion to Free Will (pp. 109–​120). New York: Routledge. McGeer, V. (2013). Civilizing Blame. In J. D. Coates & N. A. Tognazzini (Eds.), Blame: Its Nature and Norms (pp. 162–​188). Oxford: Oxford University Press. McGeer, V. (2014). P.F. Strawson’s Consequentialism. In D. Shoemaker & N. A. Tognazzini (Eds.), Oxford Studies in Agency and Responsibility Volume 2 (pp. 64–​92). New York: Oxford University Press. McGeer, V. (2015). Building a Better Theory of Responsibility. Philosophical Studies, 172(10), 2635–​2649. McGeer, V. (2019). Scaffolding Agency: A Proleptic Account of the Reactive Attitudes. European Journal of Philosophy, 27(2), 301–​323. McGeer, V., & Funk, F. (2017). Are “Optimistic” Theories of Criminal Justice Psychologically Feasible? The Probative Case of Civic Republicanism. Criminal Law and Philosophy, 11(3), 523–​544. McGeer, V., & Pettit, P. (2015). The Hard Problem of Responsibility. In D. Shoemaker (Ed.), Oxford Studies in Agency and Responsibility, Vol. 3 (pp. 160–​188). Oxford: Oxford University Press. McKenna, M. (2012). Conversation and Responsibility. New York: Oxford University Press. McKenna, M., & Pereboom, D. (2016). Free Will: A Contemporary Introduction. New York: Routledge. Miller, D. E. (2014a). “Freedom and Resentment” and Consequentialism: Why “Strawson’s Point” Is Not Strawson’s Point. Journal of Ethics and Social Philosophy, 8(2), 1–​22. Miller, D. E. (2014b). Reactive Attitudes and the Hare-​Williams Debate: Towards a New Consequentialist Moral Psychology. The Philosophical Quarterly, 64(254), 39–​59. Moore, G. E. (1903). Principia Ethica. Cambridge: Cambridge University Press. Nelkin, D. (2016). Accountability and Desert. Journal of Ethics, 20(1–​3), 173–​189. Nichols, S. (2015). Bound: Essays on Free Will and Responsibility. New York: Oxford University Press. Nowell-​Smith, P. (1948). Free Will and Moral Responsibility. Mind, 57(225), 45–​61. Nowell-​Smith, P. (1954). Determinists and Libertarians. Mind, 63(251), 317–​337. Nozick, R. (1981). Philosophical Explanations. Oxford: Clarendon Press. Pears, D. F. (Ed.). (1963). Freedom and the Will. London: Macmillan & Co. Pereboom, D. (2001). Living Without Free Will. Cambridge: Cambridge University Press. Pereboom, D. (2014). Free Will, Agency, and Meaning in Life. New York: Oxford University Press. Pereboom, D. (2017). Response to Daniel Dennett on Free Will Skepticism. Rivista Internazionale di Filosofia e Psicologia, 8(3), 259–​265. Rawls, J. (1955). Two Concepts of Rules. Philosophical Review, 64, 3–​32. Rawls, J. (1971). A Theory of Justice. Cambridge, MA: Harvard University Press. Scanlon, T. M. (1988). The Significance of Choice. In S. M. McMurrin (Ed.), The Tanner Lectures on Human Values (pp. 150–​216). Cambridge: Cambridge University Press. Schlick, M. (1939). When Is A Man Responsible? (D. Rynin, Trans.). In The Problems of Ethics (pp. 143–​158). New York: Prentice Hall. Shoemaker, D. (2015). Responsibility From the Margins. New York: Oxford University Press. Shoemaker, D., & Vargas, M. (2021). Moral Torch Fishing: A Signaling Theory of Blame. Nous, 55, 581–​602. Retrieved from https://​doi.org/​10.1111/​nous.12316 Smith, A. (2013). Moral Blame and Moral Protest. In D. J. Coates & N. A. Tognazzini (Eds.), Blame: Its Nature and Norms (pp. 27–​48). New York: Oxford University Press.

26   Manuel Vargas Smart, J. J. C. (1961). Free Will, Praise, and Blame. Mind, 70, 291–​306. Strawson, G. (1994). The Impossibility of Moral Responsibility. Philosophical Studies, 75, 5–​24. Strawson, P. F. (1962). Freedom and Resentment. Proceedings of the British Academy, XLVIII, 1–​25. Oxford University Press: United Kingdom. Strawson, P. F. (1985). Skepticism and Naturalism: Some Varieties. New York: Columbia University Press. Talbert, M. (2016). Moral Responsibility: An Introduction. Malden, MA: Polity Press. Vargas, M. (2004). Responsibility and the Aims of Theory: Strawson and Revisionism. Pacific Philosophical Quarterly, 85(2), 218–​241. Vargas, M. (2008). Moral Influence, Moral Responsibility. In N. Trakakis & D. Cohen (Eds.), Essays on Free Will and Moral Responsibility (pp. 90–​122). Newcastle, UK: Cambridge Scholars Press. Vargas, M. (2011). The Revisionist Turn: Reflection on the Recent History of Work on Free Will. In J. Aguilar, A. Buckareff, & K. Frankish (Eds.), New Waves in the Philosophy of Action (pp. 143–​172). New York: Palgrave Macmillan. Vargas, M. (2013a). Building Better Beings: A Theory of Moral Responsibility. Oxford: Oxford University Press. Vargas, M. (2013b). If Free Will Does Not Exist, Then Neither Does Water. In G. Caruso (Ed.), Exploring the Illusion of Free Will and Moral Responsibility (pp. 177–​202). Lanham, MD: Lexington Books. Vargas, M. (2015). Desert, Responsibility, and Justification: Reply to Doris, McGeer, and Robinson. Philosophical Studies, 172(10), 2659–​2678. Vargas, M. (2017). Contested Terms and Philosophical Debates. Philosophical Studies, 174(10), 2499–​2510. Vargas, M. (2020). Negligence and Social Self-​Governance. In A. R. Mele (Ed.), Surrounding Self-​Control (pp. 400–​420). New York: Oxford University Press. Vargas, M. (forthcoming). Revisionism. In J. Campbell, K. M. Mickelson, & V. A. White (Eds.), The Wiley Companion to Free Will. Oxford: Wiley-​Blackwell. Wallace, R. J. (1994). Responsibility and the Moral Sentiments. Cambridge, MA: Harvard University Press. Watson, G. (1987). Responsibility and the Limits of Evil. In F. D. Schoeman (Ed.), Responsibility, Character, and the Emotions (pp. 256–​286). New York: Cambridge University Press. Watson, G. (2003). Introduction. In G. Watson (Ed.), Free Will (2nd ed., pp. 1–​25). Oxford: Oxford University Press. Williams, B. (1988). The Structure of Hare’s Theory. In D. Seanor & N. Fotion (Eds.), Hare and Critics: Essays on Moral Thinking (pp. 185–​196). Oxford: Clarendon Press.

CHAPTER 2

Re asons-​R esp onsi v e ne s s , F r an kfu rt Exa mpl e s , a nd the Free Will A bi l i t y Michael McKenna

Reasons-​responsive theories of freedom explain exercises of free agency in terms of responsiveness or sensitivity to reasons. To a first approximation, an agent acts freely just in case, in acting, she is responsive to a sufficiently rich range of reasons. This formulation requires numerous qualifications, but it will serve as a simple starting point.1 A familiar confusion about such theories is that they require of free action that an agent actually respond to reasons, or actually respond to the good reasons there are. But of course, barring a skeptical conclusion about free will, agents sometimes freely act contrary to the reasons they have for how they ought to behave, either from routine, whim, recklessness, weakness of will, or perversity for that matter. Thus, the responsiveness or sensitivity component in a reasons-​responsive theory is modal. It involves dispositions or abilities that need not be manifested or exercised, or even if manifested or exercised, need not be manifested or exercised optimally. One way, then, to put the thesis is that a reasons-​responsive theory concerns acting from the ability to be responsive to reasons, even in contexts in which one is not acting in ways that are responding to the best reasons, or for that matter responding to any reasons.2 Although reasons-​responsive theories are amenable to differing ways of approaching the free will debate, for the purposes of this chapter, I will understand free will in terms

1 

For helpful comments, I would like to thank Josh Cangelosi, Randolph Clarke, Justin Coates, Taylor Cyr, Andrew Eshleman, Joseph Metz, Dana Nelkin, Derk Pereboom, Carolina Sartorio, Matthew Talbert, Mark Timmons, and Robert Wallace. I would also like to thank Dana and Derk for inviting me to contribute to this volume. For further assessment of reasons-​responsive theories, see my 2011, and 2017; and also McKenna and Pereboom 2016. 2  A complication I will not work out here is that on some views of action, actions are necessarily done for reasons (e.g., Setiya 2003).

28   Michael McKenna of the ability of an agent to control her behavior in the strongest sense necessary for moral responsibility (e.g., Haji 2009; McKenna 2013; McKenna and Pereboom 2016; Mele 2006; Pereboom 2014; and Timpe 2008). A free act just is an act that issues from an exercise of this ability. As for moral responsibility, I will restrict attention to moral responsibility for blameworthy conduct, and to the accountability sense of moral responsibility (see Shoemaker 2015; and Watson 1996). For an agent to be morally responsible and blameworthy for an act in this sense is for her to be accountable for it, thereby licensing pertinent others in the moral community to hold her to account by confronting her, making moral demands, and so forth (McKenna 2012: Shoemaker 2015: 87). As such, she would deserve their overt and directed blame. Free will on this view just is whatever ability is required of a morally responsible agent to have sufficient control over her culpable conduct that she would deserve blame for it.3

1.  Initial Appeal In contemporary work on free will, reasons-​responsive theories have been advanced mostly in the service of compatibilism, but their initial appeal is independent of the compatibilism/​incompatibilism debate.4 How so? I’ll offer three considerations. First, a reasonable constraint on a theory of free will, regardless of whether it is a libertarian or compatibilist proposal, is that the freedom identified is unique to persons—​or even if it is not strictly speaking unique to persons, it still marks something of importance for persons. Consider the elegant Aristotelian thesis that man is a rational animal.5 Admittedly, being sensitive to reasons does not fully capture what it is for a human person to be a rational animal. A creature might be sensitive to reasons and yet not be sufficiently sensitive to count as one whose nature or essence is best captured in terms of being rational. Sensitivity to some reasons, after all, might count as a mark of irrationality, depending on the force of other countervailing reasons. Nevertheless, if it is true that a human person is essentially or most fundamentally a rational animal, then for her

3  Some philosophers do not understand free will by reference to moral responsibility and considerations of deserved blame (e.g., Clarke 2003; Ginet 1990; Kane 1996; van Inwagen 1983; Vihvelin 2013). Despite this difference between us, what follows should be of use to them as well. 4 Treat compatibilism as the thesis that it is metaphysically possible for an agent to possess the free will ability and to act freely at a world in which determinism obtains. Determinism can be understood as the thesis that only one future is physically possible (van Inwagen 1983: 3). Although not strictly required of the metaphysical thesis, compatibilism typically involves the further empirical thesis that most actual human persons do have free will and often act freely. Incompatibilism is the denial of compatibilism. And libertarianism is the view that incompatibilism is true and that persons do possess free will and (sometimes) act freely. 5  Terrence Irwin (1980) argues that Aristotle can be interpreted through the lens of reasons-​responsiveness.

Reasons-Responsiveness,FrankfurtExamples,andtheFreeWillAbility    29 freedom to consist in her sensitivity to reasons would be for her freedom to involve an expression of something that is distinctive of her nature. Second, reasons-​responsive theories are especially effective at capturing the range of excuses and exemptions. One test of a satisfying theory of freedom is that it does a good job of explaining a rich range of our judgments about familiar cases. Especially telling are our judgments about when agents are characteristically excused or exempted from responsibility and blameworthiness due to lack of freedom. Agents suffering from phobias, addictions, compulsive disorders, delusions, hallucinations, and so forth, will have their responsiveness to pertinent reasons somehow compromised. So will agents who are deceived, misinformed, tied up, coerced, and so on. The difference, for instance, between the weak-​willed drinker who freely takes a drink and the alcoholic suffering from a genuine addictive disease can be marked in terms of whether the former would respond to pertinent reasons not to take the next drink were such reasons presented to her, as in contrast to the addict who, we can suppose, would not. Third, reasons-​responsive theories also present an appealing picture of an agent’s relation to her environment—​to, so to speak, the space of reasons. As practical agents, freedom construed as responsiveness to reasons affords an agent a way to engage with the various conditions in her environment that provide considerations for or against pertinent courses of action. To put it simply, on such a view, an agent freely relates to the reasons provided by her environment. Even when an agent acts suboptimally, failing to respond to the actually present good reasons to not act as she does, when she is free, it remains true that various reason-​providing factors either did or would have influenced how she exercised her own agency. In this way, to coin a phrase, she is free in the space of reasons.

2.  Reasons-​R esponsiveness, a First Pass It is easiest to get a purchase on reasons-​responsive theories by comparing cases, as suggested earlier when considering the weak-​willed and addicted drinkers.6 So to begin, consider a familiar case of unfree action: an agent, Handy, washes his hands from an extreme compulsive handwashing disorder. Suppose Handy gets his hands dirty one day and washes them. It might be tempting to think that in washing his hands, he does so freely and that this consists in his responding appropriately to a good reason to wash his hands. But as it happens, Handy would have washed his hands at the time whether they were dirty or not in response to any number of whacky reasons. He’d have washed them if they were clean and he just saw a garbage truck down the street, if someone within earshot had whispered the word “germ,” or even if, when his hands were truly dirty, doing so would result in his being seriously injured (suppose for some reason that in

6 

In this and the next four paragraphs, I draw upon McKenna (2017).

30   Michael McKenna the circumstances the only way for him to wash his hands requires breathing poisonous gas). What this suggests is that, in this situation, when Handy washes his dirty hands, it is fortuitous that he does so in response to a good reason. The role of good reasons is not properly integrated with what leads him to action. This in turn suggests that Handy is not sufficiently in control of his handwashing in a stable way. So he does not wash his hands freely. Contrast Handy with Dandy. Dandy is a bit of a dandy. He enjoys fancy dress and a very clean, well-​groomed appearance. As such, Dandy might have an idiosyncratic commitment to keep his hands (and the rest of him) super clean—​far more so than most normal people with good hygiene. Thus, he finds lots of good reasons for washing his hands and few good reasons not to. So his particular sensitivity to reasons for handwashing is admittedly odd. But his handwashing is still well-​integrated with the full range of reasons animating how he acts. Dandy, for instance, would not risk serious harm to himself just to wash his hands. While he does find lots of good reasons to wash them, as in keeping with his dandy ways, these many reasons are still related to his overall goals and projects as a sane person. He does not wash them, for instance, just because the very mention of germs causes him to think horrible dirty thoughts. Suppose, like Handy, Dandy dirties his hands in a situation much like the one Handy was in, and so Dandy washes his hands. It seems that while Handy does not wash his hands freely, Dandy does. The cases of Handy and Dandy suggest two related features of a credible reasons-​ responsive theory. First, to understand freedom in terms of reasons-​responsiveness involves an agent’s control in relation to a spectrum of potential but non-​actual reasons, and this spectrum comes in degrees. Second, the causal processes issuing in free action must involve the agent’s reasons-​responsive resources. To explain, consider first the matter of degree. Note that Handy would be regarded as unfree in his hand washing even if there were some sliver of reasons that would get him not to wash his hands compulsively in some situations. If, for instance, a murderer were threatening to kill Handy’s child if Handy washed his hands, or if Handy’s house were on fire and he were about to be incinerated, he’d forgo washing his hands. It is just that the range of reasons to which Handy would be responsive is too limited, and so is inconsistent with his being free in any credible sense that would bear on free will and moral responsibility. Now consider Dandy. Suppose that there is some range of reasons to which Dandy would not be responsive but should be, even by his own lights. While he often would wash his hands in response to good reasons, there are a few occasions where even he is a bit reckless and does not wash them. For instance, perhaps he should wash them just before tending to his toddler’s breakfast, but when in a hurry he fails to do so. In this way, he is not perfectly reasons-​responsive regarding his handwashing. Still, this is enough to model his freedom on the spectrum of reasons to which he is responsive. Turning to the matter of the actual causal process, when Dandy washes his hands because they are dirty, we have reason to think that the reasons for handwashing are related to his rational abilities as an agent, and that these abilities function well in playing a proper role in his act of handwashing. This is in part because his adequate degree of

Reasons-Responsiveness,FrankfurtExamples,andtheFreeWillAbility    31 reasons-​sensitivity is good evidence that Dandy is “wired” in the right way. Handy, by contrast, appears to be driven largely by his compulsion, at least in this sphere of his life. When he washes his dirty hands, there is reason to think that his dirty hands do not play the right sort of role as a reason for him that is part of the etiology leading to his act of handwashing. To the extent that his dirty hands are part of the actual cause of his handwashing, it is by way of a deviant causal process that is not an appropriately rational one, and so is in this way not freedom-​conferring. It is likely to be a fluke or an unreliable occasion for him to give expression to his compulsion.7 To help illustrate the preceding point, consider the following analogy (McKenna 2013). Imagine two rooms in two different buildings, each with a temperatures of 76 degrees. In each room is a thermostat, and each thermostat is set to 76 degrees. But as it happens, in one room the heating and cooling unit is not working properly. Although the room is 76 degrees and the thermostat is set to 76 degrees, the room would remain 76 degrees if it were set to values as high as 82 or as low as 70 degrees. In the other room, the room’s temperature is reliably linked to the thermostat’s setting so that setting the thermostat at 76 plays a proper role in the room’s temperature of 76 degrees. Were that thermostat set to 77 or 75, the room would regulate and adjust to be 77 or 75. It would thus be aptly responsive or sensitive. Key here is the following insight: These dispositional facts do not just tell us about other non-​actual but possible ways for the thermostat to behave or fail to behave in these two rooms. They also tell us about the actual causal processes, the way that heating and cooling machinery is actually causally functioning. In the former case, we can see that the operation of the thermostat is too fluky to think that it is causing properly (or causing at all) the room’s being the temperature it is. In the latter case, it is apparently finely so tuned, and is thus causing in the right way the room’s being 76 degrees. The same causal insights apply to the differences between Handy and Dandy’s handwashing. This causal thesis bears on an important distinction between two ways of theorizing about freedom. Some philosophers explain freedom in terms of alternatives to what an agent does do. These philosophers, leeway theorists, focus on the leeway an agent has to do something other than what she does, and her freedom consists, at least in part, in her ability to access this leeway and do something other than what she does.8 Others, source theorists, focus on the actual source of an agent’s action and on the etiology pertaining to what she does do. On such an approach, freedom is not explained in terms of what other things an agent might have done or was able to do but on what 7  The issue of deviant as opposed to non-​deviant causal processes in the production of action is familiar to those who work on the theory of action. To illustrate, consider a famous case due to Donald Davidson (1963) of a climber who realizes that to survive he must let go of a rope. The thought so unnerves him that he does so, but in a way that was not intentional. Here, he does respond to reasons, but not through a causal process that involves his controlling his action. As readers might suspect, it is extremely difficult in theses contexts to nail down just what non-​deviant as opposed to deviant causal processes consist in. 8  For example, see Brink and Nelkin 2013; Campbell 1997; Clarke 1993; Ginet 1990; Horgan 1979; Kane 1996; Lehrer 1976; Lewis 1981; Nelkin 2011; van Inwagen 1983; Vihvelin 2013; and Wolf 1990.

32   Michael McKenna she did do and what agential ability was manifested in her doing that thing.9 Because reasons-​responsive theories are typically characterized in terms of sensitivity to a range of reasons, including reasons that were not actually present when an agent acted, it is tempting to think that reasons-​responsive theories are leeway theories. An agent’s freedom, on this view, consists in her freedom to have done something other than what she would have done, were different reasons present. While it is true that reasons-​ responsive theories do seem to invite a leeway interpretation (a point we’ll consider in detail later), this is not required. Just as illustrated in the preceding paragraph, one way to understand the relevance of an agent’s being sensitive to a range of non-​actual but possible reasons to do otherwise is that this reveals an important aspect of the actual causal process whereby an agent brings about her own actions when she acts as she does. In this respect, it is at least an open possibility for a reasons-​responsive theorist to opt for a source theory.

3.  Reasons-​R esponsiveness and Compatibilism Now consider the special appeal reasons-​responsive theories have for compatibilists. Here I’ll highlight just two. First, recall the classical compatibilist thesis that it is not causation itself (deterministic causation) this is incompatible with freedom, but only certain kinds of causes. So, for instance, Moritz Schlick (1939) argued that freedom was not to be contrasted with causation, but instead with compulsion. Compulsion is of course a kind of causing, and it defeats freedom, but it is a simple fallacy of reasoning to infer from this that any sort of (deterministic) causation defeats freedom. This particular way of advancing the argument for compatibilism is transparently weak. There are other causal defeaters to freedom other than compulsion. As the list grows, one wonders if what they have in common is just (deterministic) causation itself. Assuming that this is not so just begs the question in favor of compatibilism. What the compatibilist needs is a principled way of showing what the various defeaters have in common other than merely being (deterministic) causes. But now consider a well-​developed reasons-​responsive theory, and note the point made in the preceding section, that a reasons-​responsive theory ought to be thought of as a theory about the causes leading to free action. Here it seems that we have a viable compatibilist candidate for distinguishing between freedom-​defeating and freedom-​conferring causes, one that lives up to the aspirations of the classical compatibilists. Agents whose acts are caused by a reasons-​responsive process are free; 9  For example, see Frankfurt 1971; Fischer 1994; Haji 1998; Hunt 2000; Fischer and Ravizza 1998; McKenna 2013; Pereboom 2001, 2014; Sartori 2016; Stump 1990; and Wallace 1994.

Reasons-Responsiveness,FrankfurtExamples,andtheFreeWillAbility    33 agents whose acts are not so caused are not free.10 The kinds of causes, then, that are freedom-​defeating are those that thwart reasons-​responsiveness. Of course, the plausibility of this strategy is to be found in the details. A reasons-​responsive theorist owes us an account of what constitutes an adequate freedom-​conferring degree of responsiveness or sensitivity to reasons, and whether that account offered really can avoid importing any incompatibilist presuppositions. Second, reasons-​responsive theories concern an agent’s sensitivity to a spectrum of reasons, and so to a range of non-​actual but possible reasons.11 To know whether an agent is responsive to a pertinent class of reasons, we have to ask how she would have responded to certain reasons were they present. This is so despite the fact that these reasons were not actually present and she did not actually respond to them. Key, then, is a collection of counterfactuals. When true of an agent, it seems she is reasons-​ responsive; when not true, it seems she is not. Relying upon such truths is appealing to compatibilists, since it is widely agreed that determinism per se is compatible with the truth of counterfactuals. How so? Compatibilism is a thesis about what is physically possible given the actual past (and laws of nature). It permits different physical possibilities if, say, the past were slightly different—​especially if an agent were acting in the context of different reasons from the reasons actually bearing on her situation. Recall Dandy. He acted freely in washing his hands. Suppose, as the compatibilist would argue, he did so freely in a world at which determinism is true. Hence, when he washed his hands, his doing so was the only future physically possible at any point antecedent to his washing. If he was reasons-​responsive, then in response to some range of reasons, he’d have refrained from washing. For instance, suppose that if he were to wash his hands, he would have to breathe poisonous gas. In this scenario, he’d have a strong reason not to wash his hands and as an upshot would not do so. Thus, the following two propositions are compatible in the sense that the truth of one does not entail the falsity of the other: OF: When Dandy elected to wash his hands, only one future was physically possible. CF: If Dandy were to have to breathe poisonous gas, he would not wash his hands.

Since according to a reasons-​responsive theory, Dandy’s washing his hands freely is supported by a collection of true counterfactuals like CF, the truth of OF is no threat to Dandy’s freedom.

10 

The philosopher who has done the most to develop this thesis explicitly in these terms is Carolina Sartorio (2016). But many reasons-​responsive theorists should also be interpreted similarly. See, for instance, Arpaly 2006; Dennett 1984; Fischer 1994; Fischer and Ravizza 1998; Haji 1998; McKenna 2013; Nelkin 2011; Vargas 2013; Vihvelin 2013; and Wolf 1990. 11  An exception is Carolina Sartorio (2016), who relies upon actual absences of reasons rather than upon non-​actual but possible reasons.

34   Michael McKenna

4.  Reasons-​R esponsiveness and Frankfurt Examples: The Problem Reasons-​responsive theories appear to be in tension with Harry Frankfurt’s influential strategy for showing that the freedom moral responsibility requires does not involve the ability to do otherwise. This puts pressure on reasons-​responsive theorists to take some stand on whether Frankfurt’s argument is sound, and if it is, how to resolve the apparent tension. Frankfurt (1969, see article 5 in this volume) argued by way of counterexample that the ability to do otherwise is not required for moral responsibility. For instance, suppose Jones is planning to shoot Smith and he does so on his own, and for his own sane reasons. Suppose, further, that at the time of his action, he was able to do otherwise. Barring skeptical worries about free will, there is no reason to think he did not act freely and was not morally responsible for shooting Smith. But now imagine a duplicate scenario with the further detail that there is a demon, Black, who wants Jones to shoot Smith, and would prefer that Jones shoot Smith on his own. Still, Black arranges the situation so that, if Jones shows any sign that he will not shoot Smith, then Black will manipulate him and cause him to shoot Smith. However, as it happens, Black never has to intervene, because Jones shoots Smith on his own, just as it was in the previous scenario—​the one with no Black on the scene. In the case with Black, it seems that Jones is unable to do otherwise, since Black is prepared to ensure the outcome that Jones shoots Smith. Yet Jones acts in this case just as the Jones counterpart did in the Black-​ free scenario. Intuitively, because Black was passive and did not interfere at all in Jones’s action, Jones acted freely and is just as responsible as Jones who acted in the Black-​free scenario. Hence, the Frankfurtian will reason, the freedom required for moral responsibility does not involve the ability to do otherwise. There are various objections to the soundness of Frankfurt’s argument, but many endorse it. If sound, it provides a powerful case for a source theory of free will and moral responsibility. What matters in a successful Frankfurt example is not whether the agent has any robust leeway to what she does; she doesn’t. It is, rather, that her action emanates from her; she is its source. Note, moreover, Frankfurt’s argument provides important dialectical advantages for compatibilists because it allows them to avoid a powerful argument for the incompatibility of determinism and the ability to do otherwise.12 If free will (as the control condition for moral responsibility) does not require the ability to do otherwise, it does not matter if that ability is incompatible with determinism. So the source compatibilist can simply sidestep that dispute. 12  Most notably, it helps avoid taking on the well-​known Consequence Argument for Incompatibilism (Ginet 1966; and van Inwagen 1975, 1983), an argument to the effect that in deterministic contexts, the ability to do otherwise requires the ability to alter either the past or the laws of nature, something no human person can do.

Reasons-Responsiveness,FrankfurtExamples,andtheFreeWillAbility    35 One might reasonably think that Frankfurt’s argument should fit well with a reasons-​ responsive theory, since as noted before a reasons-​responsive theory can be used to support a causal thesis about the actual etiology of an agent’s action. Frankfurt’s argument shows that the source of an agent’s action is what matters regarding its freedom. A reasons-​responsive theory provides a positive account of what kind of source is freedom-​conferring. So why aren’t these two allies? Why are they in tension? To put it plainly, it appears that in the context of a Frankfurt example, a free agent is not reasons-​ responsive. Why? Because she would not do otherwise were other reasons present. The counterfactual intervener’s presence, Black in our case mentioned, would ensure that the agent not respond differently to a pertinent reason to do otherwise. In the nearest possible world in which such reasons are present, the intervener remains with his nefarious intention to see to it that the agent pursue one course of action come what may. So, isn’t reasons-​responsiveness incompatible with the freedom an agent retains in a Frankfurt example?

5.  Reasons-​R esponsiveness and Frankfurt Examples: A Solution? It is open to a reasons-​responsive theorist, even a source reasons-​responsive theorist, to deny that an agent in a Frankfurt example acts freely. She might argue that the agent is not free just because the agent is not reasons-​responsive.13 But either of the two other options seems more appealing. One is to favor the soundness of Frankfurt’s argument and so deny that reasons-​responsiveness offers the proper way to account for source freedom. Some other feature of agency instead would explain why an agent in a Frankfurt example acts freely.14 Another is to deny that Frankfurt examples really do expunge an agent’s ability to do otherwise, and so it can remain true that an agent is in some way able to respond differently to reasons when she acts freely in a Frankfurt example (more on this momentarily).15 Despite the appeal of at least two of the three options suggested in the preceding paragraph, some have instead argued that Frankfurt’s argument is, after all, compatible with a reasons-​responsive account of the freedom an agent exercises in a Frankfurt example (see Fischer and Ravizza 1994; Haji, 1998; McKenna 2013; and Sartorio 2016). The strategies for doing so can be parsed into mechanism-​based and agent-​based theories. I’ll consider each in turn. 13  The source reasons-​responsive theorist can, after all, reject Frankfurt’s argument for a source theoretic conclusion and instead discharge a different argument for a source theory. However, I do not know of a single philosopher who has pursued this strategy. 14  For instance, Frankfurt (1971) would argue that the agent acts on the desires with which she identifies. 15  See Campbell 1997; Fara 2008; Horgan 2015; Nelkin 2011; Smith 2003; and Vihvelin 2004, 2013.

36   Michael McKenna John Martin Fischer and Mark Ravizza (1998) offer an ingenious solution to the problem of rendering Frankfurt’s argument compatible with a reasons-​responsive account of freedom. Fischer and Ravizza concede that an agent in a Frankfurt example is not reasons-​responsive since an agent is not able to respond otherwise to reasons to do otherwise given the presence of the counterfactual intervener (Black in the example above). Nevertheless, Fischer and Ravizza argued, we should focus on the actually unfolding causal path of the action in a Frankfurt example, wherein Black remains dormant. That process itself can be accounted for as the sort which, when it is operating unfettered, would be responsive to reasons to do otherwise. Fischer and Ravizza simply call “the process that leads to action” (38) the mechanism of action. And they argue that when the agent owns the mechanism, then in a Frankfurt example, the agent’s own mechanism of action is responsive to reasons—​even if that agent is not. This can account for the agent’s freedom in a Frankfurt example. To test the responsiveness of the agent’s mechanism, Fischer and Ravizza argue, one must consider possible worlds in which the agent’s mechanism operates without interference, and in the context of different reasons. If the agent’s mechanism is such that, were it to respond differently to a sufficiently rich range of reasons, then it would be reasons-​responsive. Otherwise not. But in worlds where the mechanism operates unfettered, the counterfactual intervener is no longer on the scene, or instead remains dormant. While Fischer and Ravizza’s mechanism-​based proposal has much to recommend it, it also brings with it further challenges. One involves identifying what, intuitively, the mechanism of action is, such that we can then hold it fixed and consider how it would respond to different reasons (e.g., see McKenna 2001, 2013; and Ginet 2006). What might be required is that an agent respond differently to different reasons when we hold fixed all the intrinsic internal properties of the agent involved in the causal process leading to her action. But this risks being too restrictive. Perhaps doing so would require the falsity of determinism to allow that a mechanism, so restricted, be able to respond differentially to a variety of different reasons.16 A different strategy, one that is agent-​based, focuses on an agent’s responsiveness to reasons rather than on mechanisms. An agent-​based theorist might instead deny the claim that an agent who acts freely in a Frankfurt example is not reasons-​responsive, despite it remaining true that the agent is not able to do otherwise. So, for instance, I have argued (2013) that an agent might be responsive to reasons to do otherwise even if it remains true that in a Frankfurt example that agent would not act differently in response to good reasons for doing otherwise. In particular, in response to a range of good reasons to do otherwise, she would not act as she does for her own reasons. Hence, reactivity to reasons to do otherwise would not be manifested in her actually doing otherwise but instead in the reasons on the basis of which she acted as she did. This is so even if, due to the role played by the intervener, that difference in reactivity would not be manifested

16 

Fischer (2004) has offered thoughtful replies to this and other objections to his mechanism-​ based view.

Reasons-Responsiveness,FrankfurtExamples,andtheFreeWillAbility    37 in acting any differently. To illustrate, consider Jones’s act of freely shooting Smith in the context of a Frankfurt example. Imagine that Jones’s actual reason for shooting is revenge for Smith’s past poor treatment of Jones. Outside the context of a Frankfurt example, grant that Jones would be reactive to the following reason for not shooting Smith: R: Smith’s child is with him, and Jones would not want to traumatize the child, despite his reasons for revenge.

But now place Jones in a Frankfurt example. The following remains true: if Smith’s child were with him, Jones would not act on his own reasons of revenge for shooting Smith. There would thus be a variation in his own reactivity. Admittedly, he would still shoot Smith since Black would make him do it. But this just shows that acting otherwise is not required for an agent to be reactive to reasons to do otherwise, although of course in standard cases—​outside of Frankfurt examples—​the two go hand in hand. Yet a different agent-​based solution has been developed by Carolina Sartorio (2016). Sartorio’s elegant proposal turns to the metaphysics of causation and relies upon the thesis that absences can be causes.17 If so, then the absence of certain reasons can be part of the actual cause for an agent’s acting as she does. Key to Sartorio’s argument is the carefully formulated Difference-​Making Thesis (94–​101) wherein causes make a difference to their effects in that the effects wouldn’t have been caused by the absence of their causes. To explain, suppose Jones is not in a Frankfurt example when he decides to shoot Smith and that he is suitably reasons-​responsive. As earlier, among the reasons to which he would react differently and not shoot Smith is (R), Smith’s child happens to be with him at the time. Thus, it is true in this case—​because there is no counterfactual intervener—​that if Smith’s child were with him, Jones would not shoot Smith. Now add the presence of the counterfactual intervener, Black. This renders the preceding counterfactual false. But granting Sartorio’s thesis about absence causation, this remains true: In the actual scenario in which Jones acts—​when Black remains dormant—​the absence of a reason not to shoot Smith (the absence of R) is one of the actual causes of Jones’s shooting Smith. Were that absence not part of the causal mix, Jones would not have shot Smith on his own. The causes of Jones’s shooting Smith would have been different. In this way, in acting as he does, Smith is responsive to a range of reasons that include not shooting Smith if his child is with him. One source of resistance to both the mechanism-​based and agent-​based proposals enlisted earlier turns on denying that an agent who acts freely in a Frankfurt example is unable to do otherwise. It might be true that if the agent in a Frankfurt example had reason to do otherwise, she would not do otherwise. The intervener would see to that. But when the agent does not have reason to do otherwise, and when she acts unencumbered, she retains the ability to respond to reasons to do otherwise. It is just that this

17 

Sartorio also allows for the possibility that absences be quasi-​causes, which involve some metaphysical relation similar to the role played by causation (2016: 45).

38   Michael McKenna ability, what Nelkin (2011: 67) calls an interference-​free ability, is not manifested when she acts as she does. Admittedly, in cases in which she does have reason to do otherwise, the intervener then interferes with the agent’s exercise of that ability. But that does not bear on the nature of the abilities the agent has when she acts as she does on her own while the intervener remains passive. Possession of abilities, after all, even specific abilities, does not require that they are exercised. Various compatibilists have defended a view along these lines (e.g., see Campbell 1997; Fara 2008; Nelkin 2011; Smith 2003; and Vihvelin 2004, 2013). One way to develop the preceding leeway version of a reasons-​responsive strategy is to draw on work done on the metaphysics of dispositions (Lewis 1997; and Martin 1994). Kadri Vihvelin (2013) has ably developed this approach in impressive detail.18 I offer here a highly simplified picture just to capture the central line of reasoning. To begin, consider salt’s solubility. Solubility is a dispositional property that need not be manifested. Salt has this property even when it is not in fact dissolved—​when a piece of it is sitting dry on a shelf. Its solubility consists in its being disposed to dissolve when placed in water. But now consider a piece of salt encased in wax and then placed in water. It will not dissolve since the wax masks the manifestation of its actually dissolving. So the following counterfactual is false: if this piece of salt were placed in water, it would dissolve. Regardless, it remains soluble. Note that the wax’s presence does not alter in any way the intrinsic properties of the salt, which intuitively would seem to be the basis for the salt’s being soluble (Lewis 1997). Likewise, consider an agent who acts from a reasons-​responsive ability in a Frankfurt example. Take Jones from our previous example. When he acts freely, he is unperturbed by Black, who is passive. Jones, it might be argued, retains the ability to respond to different reasons when he acts as he does. Naturally, if it should happen that some reasons were present and he would (absent Black) do otherwise, Black would then intervene. In that case, Jones would not, just then, be reasons-​responsive. So the following counterfactual is false: if Jones had good reason to do otherwise, he would do otherwise. Regardless, when Jones acts on his own, he retains the ability to do otherwise—​just like the salt retains its property of solubility when encased in wax. So Black’s presence when he remains passive essentially functions as a mask of Jones’s ability to respond differently to reasons. Jones retains that ability when he freely acts as he does. Note that insofar as Black does not alter any of Jones’s intrinsic properties, it is natural to think that the basis for the pertinent ability to respond differently to reasons remains intact.19

18 

See also Fara 2008; and Smith 2003. There is a technical detail I elide here. Perhaps it is better to think of Black as a fink rather than a mask. Where masks conceal the manifestation of a disposition, finks alter the object that is said to have the dispositional property in just those conditions that would otherwise elicit manifestation of it, so that in those cases, the object does not just then actually have the pertinent dispositional property at all. For instance, a magician is prepared to turn a fragile glass vase to stone should it be dropped. It never is, so it remains fragile, but if it were dropped, the magician would change it to stone before it landed and so render it not fragile. Still, while on the shelf undisturbed, it remains fragile. 19 

Reasons-Responsiveness,FrankfurtExamples,andtheFreeWillAbility    39 Randolph Clarke (2009) has labeled the basic strategy just delineated the New Dispositionalism. While the gloss just offered is far too simplified to do justice to the care with which the view is advanced, it provides us enough to consider the question of whether an agent does or does not retain the dialectically pertinent ability to do otherwise in a Frankfurt example. Here, as Clarke notes, it might well be that there is a kind of interference-​free ability an agent retains in a Frankfurt example, just as the new dispositionalists contend, but it also seems that such an agent lacks an important ability, namely the ability to do otherwise under the circumstances (2009: 341).20 It then becomes a key point of dispute between source reasons-​responsive theorists on the one hand, and leeway reasons-​responsive theorists on the other, what sort of ability is at issue in tending both to Frankfurt cases and to the sort of freedom taken to be in dispute in the free will debate. My own position, which I will not argue for here, is that Frankfurt’s argument is most usefully employed as a way of theorizing about an ability to do otherwise in the precise context one is in. Absent some sort of thought experiment like Frankfurt’s, “precise” is open to different treatments. The appeal of Frankfurt’s argument, then, even for those who wish to deny its soundness, is that it suggests a way for adversaries to test intuitions about whether the freedom that matters for responsibility requires that, just when one acts, as one is, one is able, just then, to do otherwise given one’s present conditions. If this is what is at issue, then it seems source reasons-​responsive theorists are right to seek an account of the freedom an agent retains in a Frankfurt example that does not concern an ability to do otherwise.

6.  Capturing the Proper Scope of Responsiveness Recall Handy and Dandy. As explained, Handy’s lack of freedom, according to a reasons-​responsive theory, consists in his being unresponsive to too many good reasons to refrain from handwashing, as well as being overly responsive to too many bad reasons to do so. Dandy’s freedom consists in his being adequately responsive to 20  In making this point, Clarke is simply invoking the distinction between general and specific abilities (Clarke confirms this in correspondence.) General abilities can be retained while sleeping, paralyzed, and so forth, whereas specific ones involve having all it takes to do something along with the opportunity to do it. If this is all Clarke’s point comes to, while still forceful, it has less bite than I think can be made of it. As I see it, one might instead employ Clarke’s point to a more fine-​grained distinction between different sorts of specific abilities. One requires the ability to do otherwise in the circumstances one is in—​what Austin (1956) called the “all-​in” sense of ability. One does not, and yet is specific. Note that Nelkin insightfully introduces this distinction when she favors an interference-​free notion of ability that is not the all-​in notion, but does goes “beyond the notion of a general capacity” (Nelkin 2011: 67). Here I would employ Clarke’s challenge to the new dispositionalists by reference to Nelkin’s notion of an interference-​free ability. I am grateful to Derk Pereboom for encouraging me to clarify this point, and to Randy Clarke for helping develop it.

40   Michael McKenna many good reasons to refrain, even if he also finds lots of idiosyncratic reasons to do so. Nevertheless, as noted, Handy was responsive to at least some good reasons not to wash, like being incinerated in a house fire, and Dandy was not responsive to some good reasons to wash, like doing so before preparing his toddler’s meal. So, despite being unfree, Handy was still a wee bit reasons-​responsive, while Dandy was, despite his being free, still not as responsive as an ideally reasons-​responsive agent would be. The key, then, to capturing the relevant sense of freedom is a matter of identifying the apt range of sensitivity to reasons for or against performing a free act. That will then account for free will, understood as the strongest sense of freedom required for moral responsibility. Fischer and Ravizza (1998) offer a carefully worked-​out proposal, which they call moderate reasons-​responsiveness. As they rightly explain, reasons-​responsiveness has different components, both a receptivity component and a reactivity component (1998: 69–​76). Receptivity concerns recognizing the good reasons there are for pertinent courses of action. They focus on potential reasons to act other than as an agent does act when she acts freely. Reactivity concerns acting on those reasons. An important insight Fischer and Ravizza bring out is that the scope of an agent’s receptivity might be greater than her reactivity, and this is consistent with acting freely.21 After all, it is one thing to recognize the good reasons there are. It is another to act accordingly. People who act recklessly, or from weakness of will, or just with knowledge of the immorality of their acts do not react to good reasons for doing otherwise even though they recognize the force of those reasons. Still, they might very well act freely. Thus, Fischer and Ravizza allow for a certain asymmetry. For an agent to be adequately free, she must be fairly robustly reasons-​receptive. Yet she need not be as reasons-​reactive. That is to say, the reasons to which she would react differently were she to see them as good reasons for acting otherwise might encompass a narrower set than that set of reasons to which she would be receptive. Call this the receptive-​reactive asymmetry thesis. Consider first the receptivity component. According to Fischer and Ravizza, for an agent to act freely, she must be receptive to a fairly robust pattern of possible reasons for acting other than as she does. For instance, the reasons must hang together as a class in such a way as to show that the agent satisfies some simple conditions of practical rationality. For instance, if offering Dandy a bribe of $100 to wash his hands would be sufficient reason for him to judge it best not to wash his hands, then so would bribes of greater amounts. Likewise, if bribes of $80 would not, so would bribes of lesser amounts. Also, the pattern of reasons must show that the agent is sane and that what she takes to be reasons are grounded in reality. She would not, for instance, take someone’s whispering the word “germ” to be a reason to wash her hands (unless further context made clear a

21  Two minor points regarding the current discussion. First, Fischer and Ravizza write in terms of control rather than free will. I will ignore that difference for present purposes. Second, I will also set aside their qualification that it is mechanisms of an agent that are reasons-​responsive, and not necessarily agents themselves.

Reasons-Responsiveness,FrankfurtExamples,andtheFreeWillAbility    41 genuine need to wash hands). Moreover, she must be able to appreciate moral reasons. Agents incapable of appreciating non-​instrumental reasons for regarding the interests of others are not adequately attuned to reasons to possess a kind of freedom in the space of reasons (to use my phrase). So they are not morally responsible for what they do. As for the reactivity component, Fischer and Ravizza (1998: 73) were initially prepared to commit to a fairly radical thesis insofar as they allowed an agent to be merely weakly reasons-​reactive. This came to the minimal requirement that a free agent simply react to just one reason to do otherwise to which she was receptive. That, they contended, was sufficient for her to possess the reactivity needed to act freely. But as should be clear from cases like Handy’s, merely being reactive to some reasons to do otherwise is consistent with acting from a freedom-​defeating compulsion.22 Hence, what is needed is a more robust degree of reactivity than mere weak reactivity, one that still might be weaker by comparison with the agent’s degree of receptivity. Fischer (2005) has subsequently modified the view to accommodate this point. It is easy to misconstrue Fischer and Ravizza’s initial proposal of weak reactivity. On their earlier view, the reason or reasons to which an agent would be reactive do not set the limits of an agent’s ability to respond to reasons. That would be perverse. Rather, on their earlier view, the fact that an agent would respond to a reason to do otherwise if she were to recognize it as sufficient for doing otherwise establishes an agent’s general ability to respond to any reason she deems sufficient for doing otherwise (1998: 73). That’s closer to right. The problem with this proposal, however, is just that cases like Handy’s seem to be clearly incompatible with this thesis. Handy seems not to be able to react to a range of good reasons to refrain from washing his hands even when he is receptive to them. His compulsion sees to that. But if so, it is now unclear what the relation is between the limits of an agent’s ability to react to reasons to do otherwise and the collection of counterfactuals establishing an agent’s reactivity. The range of reasons to which an agent would react if she received them just does not settle the scope of her ability to respond to reasons. Insofar as we are trying to understand the free will ability in terms of an ability to respond to reasons, it is now indeterminate what the relationship is between the counterfactuals establishing the scope of an agent’s reactivity and an agent’s ability to react to a range of reasons. Yet a very different challenge to the condition of reactivity comes not from the criticism that weak reactivity is too weak, but that it is too strong. As Derk Pereboom (2006) has pointed out, some agents who apparently do act freely and are morally responsible act in contexts in which there simply are not reasons to which they would react otherwise. Some, for instance, are unconditionally committed to a prohibition of blaspheming God. They would never do it. Surely a person could freely refrain from such action. Does this undermine a reactivity requirement on a reasons-​responsive theory? Possibly. One

22 

For criticisms based on this concern, see McKenna 2005; Mele 2000; and Watson 2001.

42   Michael McKenna way a reasons-​responsive theorist might resist is by casting about for action descriptions that could permit some (more remote) contexts wherein an agent would react differently. Suppose, for instance, that there are no conditions in which Joseph would blaspheme God. But suppose there are conditions whereby Joseph might come to believe that the being he thought was God was not God, and in such a bizarre circumstance, he would blaspheme the being he (formerly) thought was God. Yet a different way to save the reasons-​responsive approach is to widen an understanding of reactivity to include not just reacting differently, but to stably persisting in reacting similarly as conditions vary (e.g., Pereboom 2006: 202). This is an amendment to a reasons-​responsive theory that is worth developing, but to my knowledge, no one has developed it in any detail. Set aside the preceding challenges with accounting for reasons-​reactivity. Return to the matter of reasons-​receptivity. Consider Fischer and Ravizza’s requirement than an agent who is moderately reasons-​responsive be receptive to moral reasons. Is this right? Some argue that psychopaths are unable to recognize moral reasons, and yet despite this, they are morally responsible and blameworthy for their wrongdoing. Why? Because failing to recognize the moral reasons for why one should not act as one does is itself a reason to hold that agent responsible, to resent, blame, and in pertinent cases even to punish her. Hence, psychopathy is not a reason to exempt them from blame (Talbert 2008, 2014). Of course, insofar as blame is fitting only for the blameworthy, it follows that these agents are blameworthy. If in turn blameworthiness for one’s actions requires acting freely, and so requires possession of the free will ability, then fitting blame directed at psychopaths entails that psychopaths do act freely. If all of this is right, receptivity to moral reasons is not a condition for free will. Thus, Fischer and Ravizza’s moderate reasons-​responsiveness thesis would need amending accordingly. As I see it, the controversy over the moral blameworthiness of psychopaths turns primarily, even if not exclusively, on what moral responsibility is.23 If one takes blame and its fittingness to involve little in the way of deserved harsh treatment, then it is not as puzzling to think it fitting to blame a person for failing to respond to reasons that she not only did not appreciate, but was literally incapable of appreciating because of her very own constitution. But consider those who think blame is fitting because a culpable person deserves to suffer a setback to her interests by being blamed, and maybe even deserve to be punished and so intentionally caused to suffer. They will find it more plausible to require that a blameworthy agent who freely does morally wrong be able to appreciate the moral reasons that she failed to act upon (Nelkin 2015). As I am committed to the more demanding account of moral responsibility, I advise that reasons-​ responsive theorists like Fischer and Ravizza retain the requirement of receptivity to moral reasons.

23 

Although, I am assuming something that might be no more than a philosophers’ fiction: that psychopaths are literally incapable of recognizing moral reasons.

Reasons-Responsiveness,FrankfurtExamples,andtheFreeWillAbility    43

7.  Abilities, Dispositions, and Counterfactuals Reflect upon the role of the counterfactuals enlisted in the preceding section. An agent’s responsiveness to reasons is accounted for in terms of the counterfactual conditions in which, were other reasons present, an agent would respond differently to those reasons. This general strategy is not unique to Fischer and Ravizza. Numerous reasons-​ responsive theorists have appealed to counterfactuals to do similar work, even if, for instance, they disagree with Fischer and Ravizza about a mechanism-​based approach or about the need for a leeway condition on free agency.24 There is, however, a constellation of delicate questions about just how these counterfactuals relate to the notion of ability at the heart of the free will debate. Consider the following strategy, especially tempting for compatibilists. Begin with the basic reasons-​responsive proposal. Free will should be explained in terms of the ability to respond to reasons—​to be free in the space of reasons. This ability is then identified with a set of dispositions to respond to a pertinent collection of reasons. The identification is offered as a reductive proposal. The free will ability just is a disposition or a collection of dispositions to respond to a range of reasons. These dispositions are in turn reductively analyzed as a set of counterfactuals. So there is a second reduction. To be disposed to respond to various reasons just is for it to be true that, if those reasons pertained to an agent’s context of action, she would be receptive to them by recognizing them, and for some subset, she would do otherwise in response to them.25 This reductive strategy is strongly suggested by Fischer and Ravizza (1998),26 and it is openly advanced by Kadri Vihvelin (2013). The reason it is so appealing to compatibilists is because, as explained before, determinism per se just does not pose any threat to the truth of counterfactuals. So if the compatibilist can offer an extensionally adequate account of our ability claims about free agency, and when free agency fails, by appeal to this robust collection of counterfactuals, she has a very powerful compatibilist theory of free will. The preceding picture, however, is worrisome, and for at least two different reasons. One concerns a question of metaphysical priority. Another concern doubts about extensional adequacy between ability claims and claims about dispositions. First consider the issue of metaphysical priority—​what some call grounding. Suppose one gets right an agent’s “profile” of responsiveness to reasons in such a way that this captures her free will ability. Call this agent Franny. When Franny freely performs act, A, 24 

For example, see Haji 1998; McKenna 2013; Smith 2003; Vargas 2013; and Vihvelin 2004, 2013. Naturally, here I simplify. The sort of counterfactual I offer is a stand-​in for more carefully crafted proposals meant to accommodate problems with accounting for dispositions in terms of simple counterfactual conditionals (e.g., see Vihvelin 2013). Space does not permit a more cautious discussion of this issue. 26  Nevertheless, I know of no passage that commits Fischer and Ravizza to this reductive thesis, which in my estimation is all to the good for them. 25 

44   Michael McKenna she is disposed to recognize reasons R1 through same vast number of reasons to Rn and to react to some subset of them. Hence, when Franny As, some set of counterfactuals is true that are roughly of the following form: • CF1: If R1, then Franny would be receptive and reactive to R1 and so do otherwise; • CF2: If R2, then Franny would be receptive and reactive to R2 and so do otherwise; • . . . and so on through CFn. But now consider the simple Euthyphro question: When Franny acts, is she able to respond to reasons because she is disposed to respond to these reasons R1-​Rn, given that CF1-​CFn are true, or is she so disposed, given that CF1-​CFn are true, because she is able to respond to reasons? Here is a reason to favor an abilities-​first view, and so to reject a reductive analysis of the free will ability in terms of dispositions and counterfactuals. R1-​Rn is simply a snapshot of the particular reasons to which Franny would be responsive when she performs A. But like any normal person, the reasons that might or might not move Franny are fluid. She can change her mind. R14, for instance, might be a reason that she’d swap out for some other, say R14*, should she happen to read a newspaper just prior to performing A. Still, she’d be just as able to respond to reasons. It is just that she would do so in slightly different ways. There is, in short, a flexibility or variability to what particular set of reasons would move a person at any time. What reasons a person would respond to if they were to bear on her context of action are what they are because she is able to respond to reasons, not vice versa. Her ability fixes, so to speak, the scope of her responsiveness when she performs A. Suppose the brief argument set out before is sound and so an abilities-​first thesis is correct. That is, suppose the reductive project fails. Does this pose much of a threat to the overall approach? Maybe not. Consider Franny’s act, A. Assuming we can get an extensional equivalence between our ability claim about Franny’s freely A-​ing and the disposition to respond to a range of reasons, we are still entitled to something like the following biconditional: • FA: Franny freely A-​s if and only if (CF1 & CF2 & . . . CFn) It is just that, on the current nonreductive proposal, (CF1 & CF2 &. . . CFn) is true because Franny’s free will ability is what it is, an ability which she exercised in A-​ing. Given this, a compatibilist might argue that we are on safe ground so long as we are still entitled to the claim that the conjunctive proposition (CF1 & CF2 & . . . CFn) is compatible with determinism. This should be enough to establish that Franny freely A-​s is compatible with determinism. That seems right. But note that this turns on the prospects for analyzing dispositions in terms of counterfactuals. And as it happens, that turns out to be a tall order.27 It simply is not clear that any account of abilities in terms of dispositions 27 

Perhaps the most carefully worked out attempt to specify the counterfactuals that would be the reduction base for a free will ability is Vihvelin’s 2004, 2013. But as Randolph Clarke has pointed out

Reasons-Responsiveness,FrankfurtExamples,andtheFreeWillAbility    45 can succeed (e.g., see Vetter 2014). If not, then, at best, the compatibilist would be entitled to is the following less ambitious biconditional: • FA: Franny freely A-​s if and only if Franny is disposed to respond to R1 through Rn. The trouble with this weaker thesis is simply that without the appeal to counterfactuals, we have no more reason to think that dispositions to respond to reasons themselves are compatible with determinism than we do to think an ability like the free will ability is.28 The compatibilists’ aspiration has been to find a tight link between the free will ability and something uncontroversially thought to be compatible with determinism. Without an appeal to propositions like (CF1 & CF2 & . . . CFn) there is not much incentive for compatibilists to explain the free will ability in terms of dispositions to respond to sets of reasons. She might as well just theorize directly in terms of abilities to respond to reasons and drop the focus on dispositions altogether. Even setting aside the preceding worries, there is also the further question of whether it is possible to achieve the relevant extensional equivalence.29 There is reason to be skeptical. To begin, consider Peter van Inwagen’s observation about the difference between dispositions and abilities (or as van Inwagen put it, capacities and abilities). In response to pertinent stimulus conditions, dispositions yield manifestations “willy-​nilly” (1983: 11). But agential abilities seem to involve the “power to originate” (11). The capacity to understand French results in one’s simply understanding it willy-​nilly upon hearing it. The ability to speak it does not. An agent originates speech when, in acting, she speaks French. But now notice that the characterization of the free will ability is cast in terms of how an agent would respond if certain other reasons were present. This suggests a model in which the agent as the being who would initiate the varied response is passive and her responding otherwise is simply something that happens to her willy-​nilly rather than something she settles upon doing and originates.30 My point here is only that these conceptual differences between (mere) dispositions and agential abilities pose a challenge for compatibilists who wish to rely upon the strategy of tending to a collection of dispositions to respond to reasons.31

(2009), Vihvelin’s proposal is plagued by some of the same problems facing prior efforts. For a helpful assessment of work in abilities, see Choi and Fara 2018. For an equally helpful assessment of work on abilities, see Maier 2014. 28 

To clarify, the point is not about general abilities and similarly general dispositions. It is, rather, about specific abilities to freely perform some free act and equally specific dispositions to respond or not to reasons in the conditions one is in. 29  For a careful examination of the issues discussed here, see R. Wallace (unpublished manuscript). 30  I am indebted to Randolph Clarke, who in conversation first brought this problem to my attention several years ago. I know of no explicit attempt to take up this challenge for reasons-​responsive theorists. 31  But maybe this is not a problem. As Derk Pereboom has remarked in correspondence, it is not clear why some dispositions could not have active rather than passive manifestations. I cannot fully answer this here, although I am not sure that put this way, it amounts to more than a verbal dispute. Perhaps abilities just are dispositions involving active manifestations rather than passive ones.

46   Michael McKenna Yet a further difficulty has to do with a simple conceptual point about what an agent would do and what she is able to do. Recall the receptivity-​reactivity asymmetry thesis mentioned before. It allows for the prospect that an agent might be disposed to recognize certain reasons and yet not be disposed to react to those reasons. As noted, this is needed to explain free action that involves weakness of will, clear-​eyed moral wrongdoing, and so forth. To illustrate the problem, consider the following simple example. David the professional cyclist freely rides alongside Michael the inept, aging cycling enthusiast. David, as it happens, is a sweetheart of a guy, whereas Michael is a bit of a braggart. Were Michael to start trash-​talking about his cycling powers, he’d give David good reason to crush Michael on the bike, just leave him suffering on the road in the middle of nowhere. But as it happens, David is such a sweetheart that he would not do that to Michael even though Michael would deserve it. Regardless, David is no dummy. He can see that Michael probably deserves to eat a bit of humble pie. So David would be disposed to be receptive but not reactive to stomping Michael’s arse. However, regardless of what he would do, it is clear what he is able to do, which is teach Michael a good lesson.32 So it seems we cannot account for an ability like the free will ability exclusively in terms of counterfactuals regarding what an agent would do were she to have reasons to do otherwise. An agent’s abilities to act simply outstrip what she would do given some specification of the reasons she might have to act one way rather than another.33 The preceding discussion suggests that a reasons-​responsive theorist might just as well plump for an abilities-​first view and forgo any attempt to exhaustively account for an ability to respond to reasons in terms of dispositions. Perhaps a better strategy is simply to account for the free will ability in terms of abilities. This, for instance, was how Susan Wolf (1990) advanced her reasons-​responsive proposal, and Dana Nelkin (2011) has also adopted the same strategy. This is consistent with tending to pertinent dispositions and counterfactuals as a kind of forensic evidence for an agent’s having or not having the pertinent ability. When she has it, some range of counterfactuals such as (CF1 & CF2 & . . . CFn) will be true. When there is no such true proposition, this will be decisive evidence that she lacks the ability. Call this the direct-​abilities approach. Note that on a direct-​abilities view, one could enlist the fact that propositions like (CF1 & CF2 & . . . CFn) are compatible with determinism as an abductive basis for inferring that the abilities to which they pertain are also compatible with determinism.34 32  As I am imagining the case, it is not that David is perhaps disposed to give characters the likes of Michael a dose of what they deserve, where this is just masked by another disposition, like the disposition to get along with others. It is just that he really does not have these inclinations. Thanks to Dana Nelkin for encouraging me to clarify this. 33  It would be one thing if somehow we could infer a general ability to react to all reasons just by registering that an agent would react to some reason were it to apply to her. Recall that this was one reason for Fischer and Ravizza to commit to their reactivity-​is-​all-​of-​a-​piece thesis. But as noted above, this is not a live option, since some who suffer from impairments of agential abilities are nevertheless disposed to react to some reasons despite the impairment. Mele’s (2000) convincing example is of an agoraphobic who would leave her house if it were on fire. 34  I’ve recently suggested this strategy for a theory of free will built from Gary Watson’s compatibilist proposals (McKenna 2019).

Reasons-Responsiveness,FrankfurtExamples,andtheFreeWillAbility    47 There may be an even more austere way to avoid the troubles for the reductive strategy. Rather than seek an account of abilities and so treat the most fundamental explanatory question to be about an essentially modal notion, explain free action in non-​modal terms altogether. Perhaps Carolina Sartorio’s (2016) approach could be advanced in this way. She argues that free action can be grounded in the actual causes issuing in an action. This will involve reasons-​reactivity. But the reactivity will be accounted for in terms of the actual causal history, which will include both the presence and the absence of various reasons. This will even license inferences about agent’s abilities. However, if the pertinent causal facts are themselves not modal, the truths about these abilities will not be directly grounded in counterfactuals, dispositions, or other modal notions. They will instead be grounded in the actual causal facts about the agent, including, crucially, the intrinsic features of the agent that played the relevant casual role in generating free action.35

8.  Further Issues There are several other exciting issues worth exploring regarding reasons-​responsive theories of freedom and moral responsibility. Space does not permit exploring these topics in any detail. Here I will simply mention a few of them. To begin, there is a common objection to Fischer and Ravizza’s (1998) source compatibilist proposal, and to similar proposals (e.g., Haji 1998; McKenna 2013), that they are repackaged versions of classical compatibilism—​a leeway view. Why? Because they rely upon “alternative possibilities,” and so on a condition of the ability to do otherwise. So, when Fischer and Ravizza promise to give a theory that does not rely upon a condition of alternative possibilities for moral responsibility, some might protest that counterfactual reasoning is used, and this just is to show that in some sense, agents are able to act otherwise, precisely because of their reliance on counterfactual reasoning. Note that this criticism is friendly to the new dispositionalists’ thesis that agents in Frankfurt examples retain the ability to do otherwise. Chris Franklin (2015), for instance, has recently developed this criticism. However, Taylor Cyr (2017) responds by arguing that mere alternative possibilities as ways of understanding the conditions of free agency do not establish that an agent is able to access such alternatives, and so a source compatibilist conclusion is not refuted.36

35 

Sartorio’s considered view is designed to be neutral on whether causal facts that ground an agent’s freedom are themselves modally grounded (Sartorio, unpublished manuscript). Still, she at least allows for a non-​modalized grounding of a causal theory of free action. 36  My impression of the dispute between Franklin and Cyr, as in keeping with Clarke’s (2009) diagnosis of the new dispositionalism, is that Franklin is right to argue that even source theorists should grant that some sort of ability to do otherwise is required for moral responsibility—​at least a general ability. (Although see Cyr and Swensen, forthcoming, for an argument against even this thesis.) But if we tend to a more carefully clarified sense of the ability at issue, Cyr is right that the alternatives that help

48   Michael McKenna Here is another problem that also bears on some of the issues discussed in the preceding section. Consider an agent at a determined world who is blameworthy for some act of moral wrongdoing but who failed to recognize the reasons for acting morally. The reasons-​responsive theorist contends that her freedom consists in her being reasons-​responsive. But note that in the nearest possible world in which the actual good reasons to do otherwise were present—​the actual world—​that agent does not recognize those reasons. But then why shouldn’t we worry that the agent is not able to be receptive to those reasons? If she is not, it does not seem like much of a consolation to a skeptic to say that the agent’s freedom consists in her ability to respond to other reasons that might obtain in other possible worlds. At this juncture, the incompatibilist is likely to pounce and just point out that a reasons-​responsive proposal only works if it also permits an agent the ability to respond to the actual reasons for doing otherwise that a culpable agent failed to react to. This, however, requires both the direct ability to do otherwise, and the falsity of determinism. Patrick Todd and Neil Tognazzini (2008) have thus proposed an amendment to Fischer and Ravizza’s proposal that includes a requirement of the ability to be receptive to the actual moral reasons bearing on the context of a blameworthy agent’s actions. Note that, for related reasons, there is a similar worry about reactivity to the moral reasons to which a blameworthy agent knowingly failed to react. Here what is needed is, seemingly, the ability to react to the very reasons at play in an agent’s actual situation, and to do so in the actual situation one is in. But now, again, don’t we need an incompatibilist requirement of the ability to react otherwise?37 Yet another interesting issue concerns whether a reasons-​responsive theory stands best on its own or would instead be better advanced in the context of a broader theory. One question, for instance, concerns the justification for featuring reasons-​ responsiveness in a theory of moral responsibility. One elegant answer to this question is that reasons-​responsiveness is one element in a set of conditions needed to provide an agent with the fair opportunity to avoid the costs of being penalized for wrongdoing when held morally to account (Brink and Nelkin 2013). On this proposal, reasons-​ responsiveness is but one element in a wider overall account of moral responsibility. A still different motivation for treating a reasons-​responsive theory as an ingredient in a more general account of freedom is that perhaps reasons-​responsiveness fails to capture important elements in a comprehensive account of what is distinctive about free human agency—​the sort thought to be central to the free will debate. Philosophers like Harry Frankfurt (1971) and Gary Watson (1975), for instance, have focused on the internal architecture of a free agent’s mental life and argued that freedom is best captured in terms of a kind of harmony or mesh between the motivations leading to action and the elements crucial to an agent’s own self-​identifications. Are the motivations impediments to or expressions of one’s own self? Recently, Chad Van Schoelandt and

capture the freedom an agent possesses in these contexts do not support the contention that an agent retains the relevant sense of ability. 37 

For an attempt to resolve this problem for the source compatibilist, see my (2005).

Reasons-Responsiveness,FrankfurtExamples,andtheFreeWillAbility    49 I (2015) have proposed a hybrid theory that merges elements of a reasons-​responsive theory with a mesh theory. There is yet the question of whether a well-​developed compatibilist version of a reasons-​responsive theory does an adequate job of answering the concerns of incompatibilists, or for that matter simply open-​minded undecided neutral inquires who might remain unconvinced. Carl Ginet (2006), for instance, offered a rich defense of crucial ingredients of Fischer and Ravizza’s (1998) compatibilist proposal, declaring himself to be a reasons-​responsive theorist. However, he also argued that for the proposal to offer an adequate account of the freedom required for moral responsibility, we need a further incompatibilist requirement that determinism is false.38 In general, it should be born in mind that reasons-​responsive theories should not be thought as exclusively the domain of compatibilists.

9.  Concluding Consideration In closing, it is worth noting that there is a conception of free agency that is not rooted most fundamentally in the ability to respond to reasons at all, but simply to act unfettered, perhaps merely as an expression of oneself, or as a determiner of one’s yet-​ to-​be-​formed self (e.g., Sartre 1956; G. Strawson 1986). Moreover, some might argue that in the contexts that are especially important to theorizing about human freedom, an agent’s predicament is precisely that all of her reasons for and against a course of action are in equipoise, and so she remains torn about what to do (Balaguer 2010; Kane 1996). Her freedom to act between such options, it seems, cannot be explained in terms of responding to reasons. Her reasons have run out. That’s her problem. Her freedom, then, it might be argued, is a freedom to act without any further reasons. As a reasons-​ responsive theorist, I remain uncertain how to respond to these concerns. But short of dogmatic dismissal, they merit attention.

References Aristotle. 1985. Nicomachean Ethics. Trans., Terrence Irwin. Indianapolis, IN: Hackett. Arpaly, Nomy. 2006. Merit, Meaning, and Human Bondage. Princeton, NJ: Princeton University Press. Austin, J.L. 1956. “Ifs and Cans.” Proceedings of the British Academy 42: 109–​132. Oxford University Press. United Kingdom. Balaguer, Mark. 2010. Free Will as an Open Scientific Question. Cambridge, MA: MIT Press.

38 

See also Christopher Franklin’s “A Theory of the Normative Source of Pleas” (2013).

50   Michael McKenna Brink, David, and Dana Nelkin. 2013. “Fairness and the Architecture of Responsibility.” In David Shoemaker, ed., Oxford Studies in Agency and Responsibility (vol. 1, pp. 284–​313). Oxford: Oxford University Press. Campbell, Joseph Keim. 1997. “A Compatibilist Theory of Alternative Possibilities.” Philosophical Studies 88: 319–​330. Choi, Sungho, and Michael Fara. 2018. “Dispositions.” Stanford Encyclopedia of Philosophy. Clarke, Randolph (1993). Toward a credible agent-​causal account of free will. Noûs 27 (2):191–​203. Clarke, Randolph. 2009. “Dispositions, Abilities to Act, and Free Will: The New Dispositionalism.” Mind 118: 323–​351. Clarke, Randolph. 2003. Libertarian Accounts of Free Will. New York: Oxford University Press. Cyr, Taylor W. 2017. “Semicompatibilism: No Ability to Do Otherwise Required.” Philosophical Explorations 20: 208–​221. Cyr, Taylor, and Philip Swensen. Forthcoming. “Moral Responsibility without General Abilities.” Philosophical Quarterly 69 (274):22–​40. Davidson, Donald. 1963. “Actions, Reasons, and Causes.” Journal of Philosophy 60: 685–​700. Dennett, Daniel. 1984. Elbow Room: Varieties of Free Will Worth Wanting. Cambridge, MA: MIT Press. Dennett, Daniel. 2003. Freedom Evolves. London: Penguin Books. Fara, Michael. 2008. “Masked Abilities and Compatibilism.” Mind 117 (468): 843–​865. Fischer, John Martin. 2005. “Reply: The Free Will Revolution.” Philosophical Explorations 8 (2): 145–​156. Fischer, John Martin. 2004. “Responsibility and Manipulation.” Journal of Ethics 8 (2): 145–​177. Fischer, John Martin. 1994. The Metaphysics of Free Will. Oxford: Blackwell Publishers. Fischer, John Martin, and Mark Ravizza. 1998. Responsibility and Control: An Essay on Moral Responsibility. Cambridge: Cambridge University Press. Frankfurt, Harry. 1969. “Alternate Possibilities and Moral Responsibility.” Journal of Philosophy 66: 829–​839. Frankfurt, Harry. 1971. “Freedom of the Will and the Concept of a Person.” Journal of Philosophy 68: 5–​20. Franklin, Chrisopher. 2013. “A Theory of the Normative Force of Pleas.” Philosophical Studies 163 (2): 479–​502. Franklin, Christopher. 2015. “Everyone Thinks an Ability to Do Otherwise Is Necessary for Moral Responsibility.” Philosophical Studies 172: 2091–​2107. Gert, Bernard, and Tim Duggan. 1979. “Free Will as the Ability to Will.” Nous 13: 197–​217. Ginet, Carl. 1966. “Might We Have No Choice?” In Keith Lehrer, ed., Freedom and Determinism (87–​104). New York: Random House. Ginet, Carl. 1990. On Action. Cambridge: Cambridge University Press. Ginet, Carl. 2006. “Working with Fischer and Ravizza’s Account of Moral Responsibility.” Journal of Ethics 10: 229–​253. Haji, Ishtiyaque. 2009. Incompatibilism’s Allure: Principle Arguments for Incompatibilism. Peterborough, Ontario: Broadview Press. Haji, Ishtiyaque. 1998. Moral Appraisability. New York: Oxford University Press. Horgan, Terrance. 1979. “‘Could,’ Possible Worlds, and Moral Responsibility.” Southern Journal of Philosophy 17: 345–​358 Horgan, Terrance. 2015. “Injecting the Phenomenology of Agency into the Free Will Debate.” In David Shoemaker, ed., Oxford Studies in Agency and Responsibility (vol. 3, pp. 34–​61). Oxford: Oxford University Press

Reasons-Responsiveness,FrankfurtExamples,andtheFreeWillAbility    51 Hunt, David P. 2000. “Moral Responsibility and Unavoidable Action.” Philosophical Studies 97: 195–​227. Irwin, T.H. 1980. “Reason and Responsibility in Aristotle.” In A.O. Rorty, ed., Essays on Aristotle’s Ethics (117–​155). Berkeley: University of California Press. Kane, Robert. 1996. The Significance of Free Will. Oxford: Oxford University Press. Lehrer, Keith. 1976. “‘Can’ in Theory and Practice: A Possible Worlds Analysis.” In Myles Brand and D. Walton, eds., Action Theory (241–​270). Dordrecht, Netherlands: D. Reidel. Lewis, David. 1981. “Are We Free to Break the Laws?” Theoria 47: 113–​121. Lewis, David. 1997. “Finkish Dispositions.” Philosophical Quarterly 47: 143–​158. Maier, John. 2014. “Abilities.” Stanford Encyclopedia of Philosophy. Martin, C.B. 1994. “Dispositions and Conditionals.” Philosophical Quarterly 44: 1–​8. McKenna, Michael. 2005. “Reasons Reactivity & Incompatibilist Intuitions.” Philosophical Explorations 8 (2): 131–​143. McKenna, Michael. 2011. “Contemporary Compatibilism: Mesh Theories and Reasons-​ Responsive Theories.” In R. Kane, ed., Oxford Handbook of Free Will, 2nd ed. (175–​198) New York: Oxford University Press. McKenna, Michael. 2013. “Reasons-​Responsiveness, Agents, and Mechanisms.” In David Shoemaker, ed., Oxford Studies in Agency and Responsibility (vol. 1, pp. 151–​184). Oxford: Oxford University Press. McKenna, M. (2012). Conversation and Responsibility. Oxford University Press. McKenna, Michael. 2017. “Reasons-​ Responsive Theories of Freedom.” In Kevin Timpe, Meghan Griffith, and Neil Levy, eds., Routledge Handbook on Free Will (27–​40). New York: Routledge. McKenna, Michael. 2001. “Review of John Martin Fischer and Mark Ravizza’s Responsibility & Control.” Journal of Philosophy 98 (2): 93–​100. McKenna, Michael. 2019. “Watsonian Compatibilism.” In D. Justin Coates and Neal A. Tognazzini, Oxford Studies in Agency and Responsibility (vol. 5, pp. 5–​37). Oxford: Oxford University Press. McKenna, Michael, and Derk Pereboom. 2016. Free Will: A Contemporary Introduction. New York: Routledge. McKenna, Michael, and Chad Van Schoelandt. 2015. “Crossing a Mesh Theory with a Reasons-​ Responsive Theory.” In A. Buckareff, C. Moya, and S. Rosell, eds., Agency, Freedom, and Responsibility (44–​64). New York: Palgrave Macmillan. Mele, Alfred. 2006. Free Will and Luck. New York: Oxford University Press. Mele, Alfred. 2000. “Reactive Attitudes, Reactivity, and Omissions.” Philosophy and Phenomenological Research 61: 447–​452. Nelkin, Dana. 2005. “Freedom, Responsibility, and the Challenge of Situationism.” Midwest Studies in Philosophy 29: 181–​206. Nelkin, Dana. 2011. Making Sense of Freedom and Responsibility. Oxford: Oxford University Press. Nelkin, Dana. 2015. “Psychopaths, Incorrigible Racists, and the Faces of Responsibility.” Ethics 125: 357–​390. Pereboom, Derk. 2014. Free Will, Agency, and Meaning in Life. New York: Oxford University Press. Pereboom, Derk. 2001. Living without Free Will. Cambridge: Cambridge University Press. Pereboom, Derk. 2006. “Reasons Responsiveness, Alternative Possibilities, and Manipulation Arguments Against Compatibilism; Reflections on John Martin Fischer’s My Way.” Philosophical Books 47: 198–​212.

52   Michael McKenna Sartorio, Carolina. 2016. Causation and Free Will. Oxford: Oxford University Press. Sartorio, Carolina. Unpublished. “The Grounds of Our Freedom.” Sartre, Jean-​Paul. 1956. Being and Nothingness. New York, NY: Washington Square Press. Schlick, Moritz. 1939. “When Is a Man Responsible?” In Schlick, Problems of Ethics (143–​156). Prentice-​Hall. Setiya, Kieran. 2003. “Explaining Action.” Philosophical Review 112: 339–​393. Shoemaker, David. 2015. Responsibility from the Margins. Oxford: Oxford University Press. Smith, Michael. 2003. “Rational Capacities, or: How to Distinguish Recklessness, Weakness, and Compulsion.” In Sarah Stroud and Christine Tappolet, eds., Weakness of Will and Practical Irrationality (17–​38). New York: Oxford University Press. Strawson, Galen. 1986. Freedom and Belief. Oxford: Clarendon Press. Stump, Eleonore. 1990. “Intellect, Will, and the Principle of Alternate Possibilities.” In Michael D. Beaty, ed., Christian Themes and the Problems of Philosophy (254–​285). Notre Dame, IN: University of Notre Dame Press. Talbert, Matthew. 2008. “Blame and Responsiveness to Moral Reasons: Are Psychopaths Blameworthy?” Pacific Philosophical Quarterly 89: 516–​535. Talbert, Matthew. 2014. “The Significance of Psychopathic Wrongdoing.” In T. Schramme, ed., Being Amoral: Psychopathy and Moral Incapacity (275–​300). Cambridge, MA: MIT Press. Timpe, Kevin. 2008. Free Will: Sourcehood and Its Alternatives. New York: Continuum Press. Todd, Patrick, and Neal Tognazzini. 2008. “A Problem for Guidance Control.” Philosophical Quarterly 58 (233): 685–​692. van Inwagen, Peter. 1983. An Essay on Free Will. Oxford: Clarendon Press. van Inwagen, Peter. 1975. “The Incompatibility of Free Will and Determinism.” Philosophical Studies 27:185–​199. Vargas, Manuel. 2013. Building Better Beings. Oxford: Oxford University Press. Vetter, Barbara (2014). Dispositions without Conditionals. Mind 123 (489):129–​156. Vihvelin, Kadri. 2013. Causes, Laws, & Free Will: Why Determinism Doesn’t Matter. New York: Oxford University Press. Vihvelin, Kadri. 2004. “Free Will Demystified: A Dispositional Account.” Philosophical Topics 32: 427–​450. Wallace, R. Jay. 1994. Responsibility and the Moral Sentiments. Cambridge, MA.: Harvard University Press. Wallace, Robert. Unpublished. “A Dilemma for Abilities-​Based Compatibilism.” Watson, Gary (1975). Free agency. Journal of Philosophy 72 (April):205–​220. Watson, Gary. 2001. “Reason and Responsibility.” Ethics 111: 374–​394. Watson, Gary. 1996. “Two Faces of Responsibility.” Philosophical Topics 24 (2): 227–​248. Wolf, Susan. 1990. Freedom within Reason. Oxford: Oxford University Press.

CHAPTER 3

At t ribu tionist T h e ori e s of Moral Resp onsi bi l i t y Matthew Talbert

1.  An Austere Approach to Responsibility Several related factors make it difficult to give a confident account of attributionism and the attributionist approach to moral responsibility. First of all, the view is a fairly recent innovation (though its central components are not without precedent). It wasn’t until 20051 that several related perspectives on moral responsibility were collected under the title “attributionism,” and this was largely in reference to work that had appeared not much earlier. One might worry, then, that the boundaries of the view are not yet clearly defined, and perhaps that authoritative presentations have yet to appear. Secondly, “attributionism” is used to refer to the views of several authors, but depending on who is using the term, the members of this group are not always the same; moreover, none of these authors defend identical theories, so, again, there’s room for disagreement about what the approach is supposed to rule in and rule out. Finally, the so-​called “attributionists” have generally been reluctant to accept this label. There’s some justification for this tendency since, as I explain in this chapter, the name is potentially misleading and has perhaps led to some confusion in the literature. Despite these concerns, progress can certainly be made in defining the topic of this chapter. I will take “attributionism” to refer centrally to an approach to moral responsibility that comprises the common elements in the perspectives defended by Pamela Hieronymi, T. M. Scanlon, Angela Smith, and myself.2 The approach can be glossed 1 

Levy (2005). Other views described as versions of attributionism include Adams (1985), Arpaly (2003), and Sher (2009). For discussion of the relation between Sher’s view and attributionism, see Talbert (2016, 53–​56, 148–​149). 2 

54   Matthew Talbert this way: assessments of moral responsibility are, and ought to be, centrally concerned with the morally significant features of an agent’s orientation toward others that are attributable to her, and an agent is eligible for moral praise or blame solely on the basis of these attributions.3 In this context, attributionists often focus on an agent’s evaluative judgments (or perhaps her “quality of will”), as revealed through her actions, omissions, beliefs, and attitudes. We are interested in these judgments because we are interested in how other agents stand with respect to us and to one another: Do their judgments, and the actions they inform, indicate an appropriate degree of respect and concern for us and for others, or not? Representative summaries of “attributionism” include the following: On attributionist accounts, an agent is responsible for an action just in case that action is appropriately reflective of who she most deeply is . . . that is sufficient for us to hold her responsible. (Levy 2007, 132; third emphasis added) [The view] often labeled “attributionism,” . . . asserts that agents are responsible for all of the actions, beliefs, and attitudes, conscious or not, that reflect their judgments about what they have reason to do, believe, or feel. (Sher 2009, 120; emphasis added) [Attributionists] argue that responsibility for an action or omission . . . requires only that the action be appropriately attributable to the agent. . . . An action is . . . attributable to the agent if it is expressive of her attitudes and values. (Levy and McKenna 2009, 115; emphasis added)

Angela Smith notes that what is distinctive about attributionism is not its interest in specifying grounds for attributing attitudes, actions, and omissions to agents—​many other approaches share this feature (Smith, unpublished).4 Instead, and as I meant the added emphases above to bring out, the distinctive feature of the view is that the relevant attributions are taken to be all that is required for responsibility. The above quotations point, then, to the relative austerity of attributionism, to borrow a term that Gary Watson once used to describe Scanlon’s approach (Watson 2002, 240). Attributionism is “austere” in the sense of lacking adornment: It gets by without positing much in the way of necessary conditions on moral responsibility.5 For example, attributionists tend to reject both the claim that you are morally responsible only for what is under your voluntary control, as well as the claim that you are responsible for wrongdoing only if you could have responded to the moral considerations that spoke

3 

This is compatible with holding agents morally responsible for morally neutral acts that are attributable to them in such a way that they would have been open to praise or blame if the acts had not been neutral (for a similar point, see Hieronymi 2008, 363 note 13). Throughout most of this chapter I will focus on blameworthiness. 4  I thank Smith for permission to cite and quote her unpublished paper. Some of the themes in this paper find their way into Smith (2015). 5  In calling Scanlon’s view “austere,” Watson also meant to suggest that it can seem coolly detached, concerned merely with assessing agents’ qualities. I take up this criticism later in the chapter.

Attributionist Theories of Moral Responsibility    55 against your behavior.6 These conditions are rejected for the same reason: Agents may reveal features of their selves, in a way that is sufficient for moral responsibility, even in cases in which they are not met. This suggests another sense in which attributionism can seem “austere.” Perhaps it is too stern and severe a view, one that, in rejecting the above conditions, holds too many people responsible. At any rate, this has been one of the central criticisms of the view.7

2.  The Recent History of “Attributionism” In her 1990 book, Freedom Within Reason, Susan Wolf criticizes what she calls the “Real Self View” of moral responsibility. According to this view, “an agent is responsible only for those actions which are attributable to her real self,” and whether one’s actions are attributable to one’s real self depends on whether “one is at liberty to govern one’s actions on the basis of one’s valuational system” (Wolf 1990, 34, emphasis added). These quotations suggest a connection between the Real Self View and attributionism, but the suggestion is a bit misleading. Certainly, both approaches elucidate the conditions under which behavior is attributable to agents (though, as already noted, this does little to distinguish these views from many competitors), but the Real Self accounts with which Wolf is concerned (e.g., Watson 2004 [1975]) focus on a fairly narrow set of issues having to do with an agent’s relationship to her own desires: Does she endorse these desires, do they align with her values, or is she alienated from them in a way that calls her responsibility into question? Contemporary attributionists give little detailed attention to this particular issue, focusing instead on developing broader accounts of responsibility and blameworthiness. However, one of Wolf ’s criticisms of the Real Self View brings out an important connection with attributionism. For Wolf, a traumatic upbringing can give us reason to “question an agent’s responsibility for her real self,” which may raise doubts about the agent’s responsibility for her present behavior (1990, 37). But for the Real Self theorist, as long as an agent has a properly integrated “real self ” (as long as she is moved by desires

6 

Associated with the rejection of these conditions is a tendency among attributionists to also reject historical conditions on responsibility (i.e., conditions that exempt from blame agents with certain sorts of personal histories) as well as certain requirements on moral knowledge (such as the requirement that blameworthy wrongdoers recognize, or could have recognized, the moral status of their behavior). 7  Attributionism is thus criticized for setting the bar for blameworthiness too low, but some versions of the view may also seem to set the bar too high. Suppose that the attribution of a morally faulty attitude is not only sufficient for blameworthiness, but necessary as well. The problem is that certain kinds of unwitting wrongdoers (different from the morally ignorant wrongdoers alluded to in the previous note) do not harbor the attitudes that this version of attributionism requires for blameworthiness. For development of the view along these lines, see Talbert (2017).

56   Matthew Talbert that accord with her values), she will not be exempted from responsibility on the basis of historical considerations: It simply “does not matter where her real self comes from” (Wolf 1990, 35). As we shall see, attributionism reaches a similar conclusion: The fact that one’s bad behavior is the predictable result of one’s upbringing does not necessarily get one off the hook. In order to make sense of the discrepancy between her own view and the Real Self theorist’s, Wolf suggests that the latter may be working with a relatively superficial conception of responsibility (1990, 39–​40). The picture that Wolf paints is of an arid and detached—​one might say austere—​take on moral responsibility: On this view, to blame someone is simply to see her as a producer of bad outcomes, as merely “a bad act-​maker” (1990, 39). But, Wolf argues, simply identifying an action as bad, and an agent as its author, misses the depth and significance of our moral responsibility practices (1990, 40–​41). In an important 1996 paper, Gary Watson responds to Wolf. He agrees that some perspectives on moral responsibility focus on the nature of agents’ actions and on the attributability of these actions to their respective agents. Watson calls these “self-​ disclosure” views: When an agent acts in a way that he reflectively endorses, this discloses something important—​and relevant to responsibility—​about his “fundamental evaluative orientation” (2004 [1996], 271). Watson contends, against Wolf, that this is not a superficial form of moral responsibility: It is, rather, “a core notion of responsibility that is central to ethical life and ethical appraisal” (2004 [1996], 263). But Watson concedes to Wolf that identifying instances of self-​disclosure, and the moral quality of these disclosures, is not all that there is to moral responsibility: Holding people responsible is not just a matter of the relation of an individual to her behavior; it also involves a social setting in which we demand (require) certain conduct from one another and respond adversely to one another’s failures to comply with these demands. (2004 [1996], 262)

For Watson, the self-​disclosure theorist’s focus on attributability is associated with a lack of attention to the conditions an agent must fulfill in order to be properly held responsible for his behavior. For example, self-​disclosure views “are silent . . . about the capacity of self-​governing agents to comprehend the grounds on which moral requirements rest, and about our authority to hold one another to these” (2004 [1996], 262–​263). Thus, following Wolf, Watson suggests that self-​disclosure views lose sight of the fact that agents formed by unfortunate circumstances might not be properly held responsible for bad behavior that is admittedly attributable to them. These agents may act badly, but without being blameworthy in a sense that goes beyond attribution “to involve the idea that agents deserve adverse treatment or ‘negative attitudes’ in response to their faulty conduct” (2004 [1996], 266). Watson is distinguishing here, as he puts it in the title of his paper, between two faces of responsibility. A person is morally responsible in the attributability sense if she satisfies conditions that allow us to attribute behavior to her. But this person may fail to satisfy

Attributionist Theories of Moral Responsibility    57 the more demanding conditions—​historical conditions, moral competence conditions, etc.—​that apply to responsibility in the accountability sense. Such an agent may be open to aretaic appraisal (virtues and vices may be attributed to her), but not properly held accountable for her behavior and targeted with the negative responses that seem to be at the heart of blame. What does this have to do with contemporary attributionism? One worry is that readers may conflate the contemporary view with Watson’s “responsibility as attributability” (2004 [1996], 271). This conflation is understandable: Both views take moral responsibility to depend, roughly, on morally significant disclosures of an agent’s fundamental evaluative orientation (as we saw Watson put it) or of her “moral personality” (Hieronymi 2008). In addition, T. M. Scanlon, for one, does refer to the sort of responsibility that is relevant to moral blame as “responsibility as attributability” (1998, 248), and he tends to speak interchangeably of “moral criticism” and “blame,” which suggests a deflated account of blame to many readers (1998, 268). However, Scanlon also says, “when moral criticism applies this makes various reactive attitudes such as guilt, resentment, and indignation appropriate” (1998, 276). It is not correct, then, to suppose that Scanlon is concerned just with specifying the conditions for attributing moral faults to agents, and not with drawing conclusions about whether agents are open to the responses involved in holding agents accountable for their behavior. The same goes for the other attributionists considered in this chapter: They take themselves to be specifying grounds for holding agents accountable for their behavior; it’s just that these grounds are found in the fulfillment of attributability conditions. In other words, attributability is enough for accountability.8 To repeat, the worry about conflation is, as Angela Smith puts it, that “when Scanlon says he is presenting an account of responsibility in the ‘attributability’ sense, one will mistakenly think that he is offering an account of the conditions of [just] aretaic appraisal” (Smith, unpublished). And Smith suggests that we have an example of this mistake in the paper that first described Scanlon’s view, and Smith’s own, as instances of “attributionism”: Neil Levy’s “The Good, The Bad and the Blameworthy.” There, Levy says that Watson’s distinction between attributability and accountability “seems to map neatly onto the distinction between responsibility as understood by attributionism and responsibility as understood by volitionism” (2005, 3).9 Smith worries that this “falsely implies that . . . [views like hers] are concerned only with the conditions for . . . aretaic appraisal,” whereas Smith aims to provide conditions for more substantial blaming responses (Smith, unpublished).

8  Michael McKenna says that Angela Smith “collapses Watson’s distinction by, so to speak, downgrading all cases of accountability-​responsibility to attributability-​responsibility” (2012, 193). It would be better to say that attributability-​responsibility is upgraded. I take Pamela Hieronymi to be making a related point when, in her discussion of Watson’s distinction, she says, “I will suggest that the reactive attitudes cannot be separated from ‘aretaic’ appraisal” (2004, 140 note 12). 9  “Volitionism” is Levy’s term for views that place a voluntary-​control condition on responsibility.

58   Matthew Talbert Smith argues that this worry—​together with her point, mentioned earlier, that attributionists are not alone in their interest in the attribution of behavior—​should lead us to dispense with the label “attributionism.” Smith is right that it is an error to run together contemporary attributionism and the view that Watson described. However, for several reasons, I’ve come to think that we should be content with “attributionism” as the name for the view discussed in this chapter. For one thing, I don’t think that the error in question has significantly derailed debate. Debate would be derailed if writers failed to engage attributionism because they mistakenly believe that the view is not concerned with giving an account of real (i.e., resentment-​involving) moral blame. But this isn’t what has happened. Even in the paper to which Smith refers, Levy argues that attributionism generates the wrong result—​one that affirms responsibility—​in cases of impaired agents. As I read him, Levy means that attributionism regards these agents as open to something like accountability-​blame when they should not be so regarded. He argues, for instance, that, contrary to the attributionist account, a suitably impaired agent may plausibly be “taken to be a disagreeable character, [but] he is not therefore a blameworthy person” (2005, 8). I take Levy to be saying that negative aretaic appraisals of such an agent may be appropriate, but he will not be blameworthy in some other, more robust way. But this makes sense as a reply to attributionism only if Levy takes that view to hold that impaired agents may be accountable in a way that goes beyond aretaic appraisal. Generally, this is what we find in the literature. As I noted in the last section, critics regard attributionism as inclined to hold too many people responsible. But this isn’t supposed to be a problem because the view engages too readily in mere aretaic appraisal; it’s supposed to be a problem because the view blames too many people in a more profound way. As long as those who argue against attributionism are arguing about the appropriate grounds for serious moral blameworthiness, then the mistake described above hasn’t seriously distorted the debate. A couple other considerations speak in favor of preserving “attributionism.” First, this ship seems to have sailed: The term is fairly well ensconced in the literature. Second, alternative proposals have their own problems. Smith sometimes substitutes “non-​volitionalism” for the view that others call “attributionism” (2008), but this obscures the degree to which the view rejects conditions on responsibility besides those having to do with volitional control. And while Levy has substituted talk of “quality of will” views for “attributionism” (2011, 158), too many other accounts—​ ones that are clearly not varieties of attributionism—​can lay equal claim to this title: for example, McKenna (2012), Shoemaker (2015), and Wallace (1996). Finally, Smith has recently turned to speaking of “answerability” because, on her account, people are morally responsible for that which they can reasonably be asked to answer (2015; see also Hieronymi 2008 and 2014). However, “answerability” suffers from a version of the problem that “attributionism” has: Some authors (Shoemaker 2011 and 2015 and Pereboom 2014) use the term to describe a type of responsibility associated with non-​ resentment-​involving blame. And, as Smith herself notes, it’s sometimes simply easier “to make certain points using the language of ‘attributability’ ” (2017, 37 note 2). I propose, then, that we so-​called “attributionists” accept this title.

Attributionist Theories of Moral Responsibility    59

3.  The Voluntary Control Condition It’s often assumed that we are blameworthy only for what is in our control, either in an immediate or a mediated fashion. After all, many of the things for which we are potentially blameworthy—​our conscious choices, for example—​have this feature. However, attributionism holds that we are open to blame on account of those things that reflect our objectionable evaluative judgments, and that not all such things are under our immediate control or are associated with prior exercises of control. In What We Owe to Each Other, Scanlon argues that blameworthiness tracks instances of morally faulty self-​governance: If an action is blameworthy . . . the agent’s mode of self-​governance has ignored or flouted requirements flowing from another person’s standing as someone to whom justification is owed. This is what . . . makes it appropriate for the person who was wronged to feel resentment rather than merely anger and dismay. (1998, 271)

Understood in this way, blame—​though, again, Scanlon often speaks simply of “moral criticism”—​will be appropriate only with respect to agents who are capable of governing themselves in the face of reasons. And in the case of these agents, such a response will be apt “only in regard to their judgment-​sensitive attitudes: That is, those attitudes that . . . should be ‘under the control of reason’ ” (Scanlon 1998, 272).10 Of course, not all criticism is moral criticism since not all reproaches call into question a person’s judgment-​sensitive attitudes. There might be a criticism implicit in a question like “ ‘Why are you so tall?,’ ” but it’s not a moral criticism—​at least not an apt one—​since a person’s height does not implicate morally significant judgments on her part (Scanlon 1998, 272). It’s also true, of course, that a person does not choose her height. But we should not conclude from this that we reasonably blame people only for what they choose. In fact, some features of ourselves that we do not choose—​and that we might not be able to change—​may reflect our judgments, and so we may be open to blame on their account. Thus, on Scanlon’s view, it is not the case “that moral criticism applies only to actions or attitudes that arise directly from an agent’s conscious judgments” (1998, 272; also see Scanlon 2008, 170). Such criticism can also apply to involuntary attitudes that are judgment-​sensitive. Angela Smith has developed this aspect of attributionism with particular force (see also Hieronymi 2008 and 2014), emphasizing the ways in which unchosen features of our selves—​attitudes, desires, what we notice or fail to notice—​can reflect our commitments. We may not be causally responsible for these facts about ourselves, but

10  Don’t be misled by Scanlon’s reference to “control.” Attitudes are under the control of reason if they are sensitive to judgments about reasons, but these judgments themselves need not be the result of a volitional process (see Scanlon 1998, 281). See note 12 for a related point.

60   Matthew Talbert we are responsible for them in the sense that they reflect how we are oriented toward others, and we may properly be asked to alter or defend this orientation. What Smith is noticing is that there is a rational relation11 between various involuntary states of agents and what they take to be important or valuable. There is, a rational connection between many of the thoughts and desires that occur to us and the evaluative judgments and commitments we accept. If we value something and judge it to be worth promoting, protecting, or honoring in some way, this should (rationally) have an influence on our unreflective patterns of thought and feeling. (2005, 247)12

For the attributionist, the presence of this relation explains why we hold one another responsible for what we do voluntarily, and since the same relationship can hold in the case of things that are not voluntary, we can be blameworthy for some things that are not voluntary. If a person’s (involuntary) indignation indicates an objectionably contemptuous judgment, or her (involuntary) inadvertence indicates an objectionable lack of concern, then she is open to moral criticism (blame), which responds just to the “content of that attitude and not to facts about its origin in a person’s prior voluntary choices, or to facts about its susceptibility to influence through a person’s future voluntary choices” (Smith 2005, 251). Scanlon makes related points in the context of discussing manipulation scenarios of the sort that incline some to put historical conditions on moral responsibility. Of course, some varieties of manipulation do undermine responsibility. If a person’s behavior is brought about by hypnosis or direct brain stimulation, then it is unlikely that he is morally responsible for that behavior (Scanlon 1998, 277). However, Scanlon argues that this isn’t because the manipulated agent was subject to causal pressures over which he lacked control. Rather, responsibility is absent in these cases because the causal pressures at issue “are of a kind that sever the connection between the action or attitude and the [manipulated] agent’s judgments and character” (Scanlon 1998, 278). If an agent’s action is brought about by direct brain stimulation, there are no grounds for “attributing to him the discernment or lack of discernment that would be revealed by [the agent] thinking that he had good reason” to perform the action and doing so on that basis (Scanlon 1998, 278). But what if an action is brought about “by ‘implanting’ in the agent the thought that it is warranted” (Scanlon 1998, 278)? Such an agent may still be relieved of responsibility even though the action in question expresses an objectionable judgment, but, once again, this is not mainly because of lack of control on the agent’s part. Instead,

11 

Smith has referred to her own account, therefore, as the “rational relations view” (2005, 240). Smith uses “judgment” loosely: Judgments “are not necessarily consciously held propositional beliefs, but rather tendencies to regard certain things as having evaluative significance. . . . ‘Judgments’ in this sense do not always arise from conscious choices or decisions . . .” (2005, 251–​252). 12 

Attributionist Theories of Moral Responsibility    61 responsibility is called into question to the degree that the judgment is aberrant and transitory, and thus not fully attributable to the agent. However, in cases of longer-​term involuntary changes, our conception of what belongs to the agent for purposes of moral assessment may shift. A previously kind person might undergo a change of character over which she lacks control “after being hit on the head or given drugs for some medical condition” (Scanlon 1998, 279). Here, “at least at first,” we might not see the agent’s out-​of-​character behavior “as grounds for modifying our opinion of her” (Scanlon 1998, 279). But if the pattern of behavior persists, we may come to see it as a part of the agent as she is now, and as forming an appropriate basis for moral criticism going forward (Scanlon 1998, 279; also see Smith 2005, 261 note 46). Some readers may think that if a person owes her bad character to a bump on the head, or to a bad upbringing, then she cannot be open to serious moral blame. Perhaps it is not enough for blameworthiness that an agent’s actions reflect her bad attitudes and judgments; perhaps blameworthy agents must also be (in a causal sense) responsible for these facts about themselves. Neil Levy, for example, holds that “I am responsible for my attitudes if I have genuinely been . . . active with regard to them; if I have chosen them,” and he objects to the attributionist’s contention that “[i]‌t need not be the case that I have controlled [my attitudes], or even that I could control them, for me to be responsible for them” (2005, 10). Levy concludes, “[i]f control matters, then its absence cancels responsibility (unless of course the agent is responsible for her absence of control) (2005, 10).13 According to Levy, what leads the attributionist astray here is a failure to acknowledge the distinction between agents that are bad and agents that are blameworthy. For the attributionist, if objectionable attitudes are attributable to an agent, then she is blameworthy, but, Levy argues, “even when attitudes are rightly attributed to agents, it is a further question whether they are responsible for them” (2005, 15). And with respect to the lapses of attention and concern that Angela Smith describes, Levy says, unless we exercise relevant control over them, we are not responsible for our lapses. Since there seems to be conceptual room for a distinction between a faulty attitude . . . and one for which the agent is at fault, and attributionists have given us no good reason for thinking that this distinction should not be made, we should reject attributionism as an account of moral responsibility.” (2005, 15; see Smith 2008 for her reply)

Despite Levy’s suggestion, attributionists are willing to grant that, beyond establishing the attributability of an attitude to an agent, there is also the question of whether the agent is responsible for her attitudes in the sense of having played a certain sort of role in their acquisition. But attributionists take the attributability of the attitude to settle—​if it is suitably objectionable—​the question of the agent’s blameworthiness, even if she is not responsible for bringing it about that she has that attitude. Likewise, the attributionist 13 

Levy, I take it, affirms the antecedent of this conditional.

62   Matthew Talbert is happy to distinguish between a faulty attitude and an attitude for which an agent is at fault. But, again, the attributionist insists that an agent’s being at fault for an attitude is not required for it to be faulty, and that the faultiness of the attitude is enough for an agent to be open to blame on account of it (even if it is not her fault—​causally speaking—​that she has that attitude). Levy and others (e.g., Rosen 2004) appear to believe that being at fault for one’s faulty attitudes is required for blameworthiness. From the attributionist’s perspective, this runs together questions about causal responsibility and moral responsibility in an unhelpful way.

4.  The Moral Competence Condition Many accounts of responsibility hold that wrongdoers are open to serious moral blame only if they are morally competent in such a way that they could have recognized, and responded appropriately to, the moral considerations that counted against their behavior.14 Again, attributionism rejects this condition on responsibility. In What We Owe to Each Other, Scanlon says that a “plausible test” for deciding whether a certain type of impairment of moral competence rules out moral criticism is to think about “whether the behavior of a creature which has that condition would, for that reason, lack the distinctive significance that moral failings generally have for relations with others” (1998, 287–​288). So, for example, if a being is non-​rational and does not make judgments about reasons at all, then, though it may harm us, its behavior would not have the significance that wrongdoing typically has for us. This is because such a being’s behavior would not express a judgment about the sort of treatment to which we are open: Its behavior would not be of the sort that can “challenge our moral standing and make resentment an appropriate reaction” (Scanlon 1998, 288). However, Scanlon argues, “a rational creature who fails to see the force of moral reasons,” but who is otherwise sensitive to rational considerations, is capable of expressing objectionable evaluative judgments, and is, therefore, open to moral blame (1998, 288). Some agents have localized impairments of moral competence. In this context, the debate often focuses on individuals who form inaccurate moral judgments because those judgments are widely accepted and nurtured in their cultural contexts. One side in this debate holds, roughly, that widespread acceptance, and cultural sanctioning, of an objectionable judgment can impair moral competence insofar as agents in these contexts often have difficulty assessing the moral status of the judgment in question. This can make it unreasonable to expect an agent to avoid wrongdoing associated with this objectionable judgment.

14 

For different versions of this condition, see McKenna (2012), Nelkin (2011), Wallace (1996), and Wolf (1987, 1990), among others.

Attributionist Theories of Moral Responsibility    63 In a well-​known discussion along these lines, Susan Wolf considers “persons who, though acting badly, act in ways that are strongly encouraged by their societies—​the slave owners of the 1850s, the Nazis of the 1930s, and many male chauvinists of our fathers’ generation, for example” (1987, 56–​57). These individuals had the misfortune of being raised in environments that cultivated bad values and bad judgmental tendencies in them, which made it difficult for them to recognize that some of their behavior was objectionable. Wolf concludes that if we believe that these “agents could not help but be mistaken about [the moral status of] their values”—​if we believe that they were unavoidably impaired in this way—​then “we do not blame them for the actions those values inspired” (1987, 57). The attributionist has a number of replies to this line of thought (some of which I will save for the discussion of resentment and protest in the next section). To start with, it should be emphasized that when we consider the racist behavior of a morally benighted slave owner, we are considering voluntary behavior that is often conducted in full awareness of the disagreeable consequences of that behavior for others: The slave owner simply dismisses—​or at least he does not assign appropriate weight to—​these considerations. Thus, it should be admitted by all sides that the actions of such a slave owner can display objectionable attitudes that are at least very similar to the kinds of attitudes that are often relevant to judgments of blameworthiness. Those who reject attributionism will maintain, however, that while the expression of such attitudes is often associated with blameworthiness, it is not sufficient for blameworthiness: In addition, blameworthiness requires (for example) that it is reasonable to expect a wrongdoer to have avoided acquiring the attitudes for which we blame her. But now the attributionist might argue that it is not clear how adding the satisfaction of this condition (or related ones) to the story of a slave owner’s wrongful behavior would change that behavior from something that does not warrant blame to something that does. Such an addition would be relevant if the fact that a slave owner might have come to a different conclusion about, say, the moral status of slavery means that his behavior is morally objectionable in a way that it could not otherwise be. But it is not clear that this is so: It is not clear that the behavior of slave owner A is objectionable in some special way, as compared to the behavior of similarly motivated slave owner B, simply because A could have (in some sense) judged that slavery is wrong. Or, more modestly, even if A’s behavior would be made especially objectionable by access to this alternative judgment, it doesn’t follow that B’s behavior is not sufficiently objectionable to make blame appropriate in his case (Talbert 2012a, 98–​101). But this doesn’t get to the heart of the position of someone like Wolf. The central idea isn’t that a slave owner’s behavior would be more objectionable if he could have been receptive to a different moral outlook. The thought is, rather, something like this: If a slave owner’s moral competence were unimpaired, then he would be a fair target for blame because it would be reasonable to expect him to have avoided acquiring the bad values that he possesses and that inform his wrongful behavior. Put the other way around, the thought is that it is unfair to blame an agent if it was unreasonably difficult for her to avoid the behavior for which she is blamed.

64   Matthew Talbert Attributionists will agree that it is sometimes unfair to blame those who have significant difficulty avoiding wrongdoing. But, the attributionist will note, this is most clearly the case when the difficulty in question is associated with a failure of self-​governance. A coerced agent, for example, may be blameless and may have difficulty avoiding wrongdoing, but the blamelessness here stems from the fact that her behavior doesn’t express the evaluative commitments that it would have expressed had she been acting as she pleased.15 Quite often, an agent who has difficulty avoiding an action fails (like the coerced agent) to govern her behavior as she sees fit, but these two things can come apart: An agent may be unable to avoid wrongdoing, but not because her control over her behavior is impaired. For example, a slave owner’s difficulty avoiding wrongdoing may stem from an entrenched commitment to seeing members of a certain race as open to being treated like property. Such an agent controls her behavior as she sees fit, and her difficulty avoiding wrongdoing is a function of an abundance of the sort of thing (a morally reprehensible point of view) that often grounds blame. The attributionist concludes that it is much less clear in this type of case (than it is in the case of a coerced agent) that it is unfair to blame one who has significant difficulty avoiding wrongdoing.16 So far, I have focused on limited impairments of moral competence, but what about wrongdoers with broader impairments? Discussions of this topic have focused on psychopathy—​or at least a philosophical approximation of this disorder. For the purposes of this chapter, it is enough to say that a psychopath is unable to make use of moral concepts and judgments (in the normal way) and unable to find motivation in moral considerations (in the normal way). As such, the psychopath is quite different from the slave owner considered before. The slave owner may recognize that there is such a thing as “moral standing,” so it is straightforward to attribute to him, and to see his behavior as expressing, judgments about moral standing. But the psychopath, as conceived here, is a figure who is not in the business of using moral concepts at all, which makes it difficult to interpret his behavior as expressing judgments about standing. Neil Levy argues on this basis that the psychopath’s behavior should not be thought of as “expressing contempt, ill-​will, or moral indifference” of the sort that would, on the attributionist approach, ground blame (2007, 135). “Contempt,” Levy maintains, “is a thoroughly moralized attitude; only a moral agent [which the psychopath is not] is capable of it” (2007, 135; see also Shoemaker 2011, 629). The psychopath certainly fails to express appropriate regard for others, but since success in this domain was never an option, his failure of regard is not an expression of disregard. Attributionists, however, have argued that psychopaths’ behavior can have blame-​ grounding moral significance even if they cannot form judgments with explicit moral content. This argument (Talbert 2008, 2012b, 2014) takes its cue from Scanlon’s 15  A coerced agent may “act as she pleases” (or least as she chooses) in one sense, but in another very clear sense, she is not doing what she wants. 16  See Scanlon (1998, 288) and Talbert (2013, 237–​238). This reasoning is related to Frankfurt’s (1998 [1969]) argument that access to behavioral alternatives is not required for moral responsibility.

Attributionist Theories of Moral Responsibility    65 observation that, “[a]‌person who is unable to see why the fact that his action would injure me should count against it still holds that this doesn’t count against it” (1998, 288). The psychopath may not make moral judgments, but insofar as he makes judgments about reasons at all, he dismisses factors that ought to inform one’s practical judgments. Thus, the psychopath inhabits a perspective according to which the suffering of others may be overlooked. We can, for example, attribute to him the judgment, “The fact that this will hurt you is no reason to refrain,” and the attributionist argues that this judgment is contemptuous in a way that is appropriately met with moral blame.17 But what is the point of blaming a psychopath? A number of philosophers emphasize the conversational aspects of our blaming practices. Blame seems not only to communicate moral condemnation and moral demands, but to also invite replies from wrongdoers: They should come to share our views, to apologize, reform themselves, and so forth. But if blame has an essential conversational point, then perhaps it is appropriate only in contexts in which such communication is possible, and the psychopath is not a potential partner in moral conversation. Gary Watson has emphasized the way in which the attitudes involved in blame often play this kind of conversational role (see also McKenna 2012 and Shoemaker 2015). Watson has said that the “negative reactive attitudes express a moral demand, a demand for reasonable regard,” and that “[t]‌o be intelligible, demanding presumes understanding on the part of the object of the demand” (2004 [1987], 230). In a recent discussion of psychopathy, Watson adds that resentment, in particular, commits us “to the appropriateness of [taking up] an inherently communicative stance” toward the target of these responses (Watson 2011, 328 note 35). The conclusion to draw, Watson suggests, is that since we cannot hope to communicate our moral perspective to the psychopath, we cannot regard him as an appropriate target of resentment. Attributionism has a response here, one which will become apparent in the next section when I discuss the roles that protest and resentment play in the attributionist perspective.

5.  Protest, Resentment, and Relationships Attributionists are united in the view that blame is not, in itself, a sanction or a punishment. As Scanlon puts it, “moral blame is fundamentally a judgment of condemnation, not a penalty” (1998, 267). Though it is often unpleasant to be the object of moral criticism and blame, that is not their purpose, nor are they tactics deployed “in order to 17  Angela Smith (2015, 118) appears to endorse this conclusion about psychopaths, and Pamela Hieronymi suggests, more generally, that blame is not rendered unfair by the fact that an agent is not “capable of controlling her behavior by the light of moral reasons” (2004, 126).

66   Matthew Talbert enforce norms of behavior” (Scanlon 1998, 285). Thus, judgments of blameworthiness are measured by different standards, and answer to different constraints, than decisions to impose hardships on others. A similar contrast can be drawn between judgments of blameworthiness and expressions of these judgments. And once we make this distinction, concerns about the fairness of blame may seem less pressing since, as Angela Smith argues, they may have more to do with the propriety of expressing our judgments of blameworthiness than with the accuracy of these judgments. The problem, Smith suggests, is that responsibility theorists often mistakenly suppose that if it is wrong “to express moral criticism or to sanction people for their desires, emotions, and other attitudes,” then “they cannot be morally responsible for them” (2008, 379; also see Smith 2007). Making a related point, Pamela Hieronymi notes that the negative attitudes involved in blame “are not action-​like. That is, they are not voluntary responses to a judgment in the way an intentional action might be” (2004, 120). So, again, whether one ought to hold these attitudes will not be measured by the same standards as a decision to harm a wrongdoer. Indeed, as Hieronymi argues, “adopting the reactive attitudes could be rendered unfair only by considerations that bear on the content of the judgments they reveal” (2004, 133). And the judgment that is revealed by an attitude like resentment is something along the lines of “I was wronged” or “I have been disrespected.” Various considerations might show that such judgments are erroneous, but the considerations at issue in debates about the fairness of blame are not typically of this sort. On Hieronymi’s account, resentment’s function is to mark the fact that one was wronged. In contrast with Watson’s conversational approach, nothing about this picture of resentment commits one to viewing the object of resentment as a potential moral interlocutor. Our private resentment need not commit us to the possibility of moral understanding on the part of the one we blame, and even expressions of resentment need not take such an end to be achievable. Of course, the attributionist does not deny that we often hope that our expressions of resentment will inspire contrition, apology, and reform in a wrongdoer. The attributionist’s point is only that resentment—​or at least a closely related, morally serious blaming emotion—​can be appropriately experienced in the absence of these possibilities (for discussion of the “closely related” emotion, see Hieronymi 2014, 31–​32 and Talbert 2014, 289–​291). But still, what is the point of resentment when we cannot inspire moral understanding in a wrongdoer? It’s tempting to say that, for the attributionist, resentment has no point beyond its sensitivity to the fact that one’s standing has been called into question. But a bit more can be said. For example, Hieronymi argues that “resentment is best understood as a protest” against a wrong done to you (2001, 546). The fact of your having been wronged makes the claim “that you can be treated in this way, and that such treatment is acceptable,” and “[i]‌n resenting . . . you challenge” this claim (Hieronymi 2001, 546). I have also argued that we should think of blame and resentment as a form of protest, and that doing so will help the attributionist respond to those who insist that expressions of blame lose their sense if they cannot be interpreted as communicative efforts (Talbert 2012a). Going further, Angela Smith has identified the “desire to protest and repudiate

Attributionist Theories of Moral Responsibility    67 conduct” as “[t]he fundamental motivational element underlying all instances of moral blame” (2013, 36–​37; emphasis added). One of Smith’s targets in making this last point is Scanlon’s account of blame, which brings me to the final issue I want to discuss: Scanlon’s perhaps problematic stance with respect to the harsher side of blame, particularly resentment. If Scanlon’s account of blame is one that does not go in for resentment (and similar responses), then this again raises the concern (described in section 2) that attributionism—​or at least Scanlon’s version of it—​is not defending a serious form of blame. In Permissibility, Meaning, and Blame, Scanlon says that a claim of blameworthiness asserts that an agent’s “action shows something . . . that impairs the relations that others can have with him or her” (2008, 128). Further, “[t]‌o blame a person is,” in addition to forming a judgment of blameworthiness, “to take your relationship with him or her to be modified” (Scanlon 2008, 128), and “to hold the attitude toward him or her that this impairment makes appropriate” (Scanlon 2008, 131). Attitudes like resentment are not a necessary component of blame on this account. Scanlon imagines a case in which his friend Joe betrays him. In response, Scanlon might “decide not to rely on or confide in Joe . . . and not to seek his company . . . or to have . . . special concern for his feelings and well-​being” (2008, 136). Scanlon takes all these ways in which he might revise his relationship with Joe to count as forms of blame. In addition, Scanlon might “also resent [Joe’s] behavior,” yet “this is not required for blame . . . I might just feel sad” (2008, 136).18 Perhaps Scanlon is right that there is more to blame than resentment, but should we agree that his sadness at being mistreated by Joe is, as he suggests, an instance of blame? Surely this is in tension with much of our ordinary talk about “blame” (Wallace 2011; Wolf 2011). And even if we count Scanlon’s decision to not seek out Joe’s company as a form of blame, it is not the robust sort that is typically at issue in debates about moral responsibility. After all, even someone who is largely skeptical about blame, as conventionally understood, can accept that it is reasonable for Scanlon to choose to spend less time with Joe (Pereboom 2014, 131–​132). Something seems to be missing from Scanlon’s account. But the missing element may not be resentment. Instead, I’m inclined toward Smith’s proposal that what is missing is protest. Perhaps, Smith says, “[i]‌t is only those modifications of attitudes that are undertaken as a way of protesting the relationship-​impairing attitudes of others that qualify as instances of moral blame” (2013, 39). Of course, resentment is a common way of registering moral protest, but it is not the only way. Playing off Scanlon’s example, Smith says,

18 

It’s worth emphasizing that Scanlon makes room for resentment even in his recent work on blame:

It would be foolish to deny that moral emotions such as resentment . . . are appropriate responses . . . [to blameworthy actions]. But these emotions are not all that blame normally involves. Other attitudes, such as modified intentions, are also important. (2013, 99)

68   Matthew Talbert I may have lost my ability to feel anger toward an unreliable friend, yet I may still protest his treatment of me by cutting off relations with him. In doing so, and doing this in protest of his latest let-​down, I make clear that I blame him, even if my predominant feeling is one of sadness. (2013, 41)

6.  Conclusion I have characterized attributionism as austere because of its narrow focus on the quality of a specific range of evaluative judgments that are attributable to agents. Successful attribution of these judgments is enough to make an agent blameworthy, regardless of whether he fulfills a number of other proposed conditions on moral responsibility. However, attributionism may also be thought austere in the sense that it is interested only in detached appraisal of agents, or in the sense that it is excessively severe and heedless of constraints that rightly apply to our blaming practices. In response to these criticisms, attributionism clarifies its focus on evaluative judgments and the attitudes that are sensitive to these judgments. Since the judgments that attributionism appraises are of paramount interpersonal significance, blaming attitudes are rightly in play as responses to these appraisals. Moreover, the blaming responses the attributionist envisions are not interpreted as sanctions, but rather as natural responses to the quality of the evaluative judgments attributed to agents. The fairness of a reactive response depends, then, on the accuracy of the attribution of an objectionable judgment, and the degree to which the response is attuned to, or calibrated by, the attributed judgment.19

References Adams, Robert. 1985. “Involuntary Sins.” Philosophical Review 94: 3–​31. Arpaly, Nomy. 2003. Unprincipled Virtue: An Inquiry Into Moral Agency. New York: Oxford University Press. Frankfurt, Harry. 1998 [1969]. “Alternate Possibilities and Moral Responsibility.” In H. Frankfurt, The Importance of What We Care About. Cambridge: Cambridge University Press, pp. 1–​10. Hieronymi, Pamela. 2001. “Articulating an Uncompromising Forgiveness.” Philosophy and Phenomenological Research 62: 529–​555. Hieronymi, Pamela. 2004. “The Force and Fairness of Blame.” Philosophical Topics 18: 115–​148. Hieronymi, Pamela. 2008. “Responsibility for Believing.” Synthese 161: 357–​373. Hieronymi, Pamela. 2014. “Reflection and Responsibility.” Philosophy and Public Affairs 42: 3–​41.

19 

I thank Randolph Clarke and Angela Smith for their helpful comments on a draft of this chapter.

Attributionist Theories of Moral Responsibility    69 Levy, Neil. 2005. “The Good, the Bad and the Blameworthy.” Journal of Ethics and Social Philosophy 1: 1–​16. Levy, Neil. 2007. “The Responsibility of the Psychopath Revisited.” Philosophy, Psychiatry, and Psychology 14: 129–​138. Levy, Neil. 2011. Hard Luck. New York: Oxford University Press. Levy, Neil and Michael McKenna. 2009. “Recent Work on Free Will and Moral Responsibility.” Philosophy Compass 4: 96–​133. McKenna, Michael. 2012. Conversation and Responsibility. New York: Oxford University Press. Nelkin, Dana. 2011. Making Sense of Free Will and Responsibility. New York: Oxford University Press. Pereboom, Derk. 2014. Free Will, Agency, and Meaning in Life. New York: Oxford University Press. Rosen, Gideon. 2004. “Skepticism About Moral Responsibility.” Philosophical Perspectives 18: 295–​313. Scanlon, T. M. 1998. What We Owe to Each Other. Cambridge, MA: Harvard University Press. Scanlon, T. M. 2008. Moral Dimensions: Permissibility, Meaning, and Blame. Cambridge, MA: Harvard University Press. Scanlon, T. M. 2013. “Interpreting Blame.” In D. Coates and N. Tognazzini (eds.), Blame: Its Nature and Norms. New York: Oxford University Press, pp. 84–​99. Sher, George. 2009. Who Knew?: Responsibility Without Awareness. New York: Oxford University Press. Shoemaker, David. 2011. “Attributability, Answerability, and Accountability: Toward a Wider Theory of Moral Responsibility” Ethics 121: 602–​632. Shoemaker, David. 2015. Responsibility from the Margins. New York: Oxford University Press. Smith, Angela. Unpublished. “The Myth of Attributionism.” Smith, Angela. 2005. “Responsibility for Attitudes: Activity and Passivity in Mental Life.” Ethics 115: 236–​271. Smith, Angela. 2007. “On Being Responsible and Holding Responsible.” The Journal of Ethics 11: 465–​484. Smith, Angela. 2008. “Control, Responsibility, and Moral Assessment.” Philosophical Studies 138: 367–​392. Smith, Angela. 2013. “Moral Blame and Moral Protest.” In D. Coates and N. Tognazzini (eds.), Blame: Its Nature and Norms. New York: Oxford University Press, pp. 27–​48. Smith, Angela. 2015. “Responsibility as Answerability.” Inquiry 58: 99–​126. Smith, Angela. 2017. “Unconscious Omissions, Reasonable Expectations, and Responsibility.” In D. K. Nelkin and S. C. Rickless (eds.), The Ethics and Law of Omissions. Oxford University Press, pp. 36–​60. Talbert, Matthew. 2008. “Blame and Responsiveness to Moral Reasons: Are Psychopaths Blameworthy?” Pacific Philosophical Quarterly 89: 516–​535. Talbert, Matthew. 2012a. “Moral Competence, Moral Blame, and Protest.” The Journal of Ethics 16: 89–​109. Talbert, Matthew. 2012b. “Accountability, Aliens, and Psychopaths: A Reply to Shoemaker.” Ethics 122: 562–​574. Talbert, Matthew. 2013. “Unwitting Wrongdoers and the Role of Moral Disagreement in Blame.” In D. Shoemaker (ed.), Oxford Studies in Agency and Responsibility, Vol. 1. Oxford: Oxford University Press, pp. 225–​245. Talbert, Matthew. 2014. “The Significance of Psychopathic Wrongdoing.” In T. Schramme (ed.), Being Amoral: Psychopathy and Moral Incapacity. Cambridge MA: MIT Press, pp. 275–​300.

70   Matthew Talbert Talbert, Matthew. 2016. Moral Responsibility: An Introduction. Cambridge: Polity Press. Talbert, Matthew. 2017. “Omission and Attribution Error.” In D. K. Nelkin and S. C. Rickless (eds.), The Ethics and Law of Omissions. Oxford: Oxford University Press, pp. 17–​35. Wallace, R. Jay. 1996. Responsibility and the Moral Sentiments. Cambridge, MA: Harvard University Press. Wallace, R. Jay. 2011. “Dispassionate Opprobrium: On Blame and the Reactive Sentiments.” In R. J. Wallace, R. Kumar, S. Freeman (eds.), Reasons and Recognition: Essays on the Philosophy of T. M. Scanlon. New York: Oxford University Press, pp. 348–​372. Watson, Gary. 2002. “Contractualism and the Boundaries of Morality: Remarks on Scanlon’s What We Owe to Each Other.” Social Theory and Practice 28: 221–​241. Watson, Gary. 2004 [1975]. “Free Agency” In G. Watson, Agency and Answerability: Selected Essays. New York: Oxford University Press, pp. 13–​32. Watson, Gary. 2004 [1987]. “Responsibility and the Limits of Evil” In G. Watson, Agency and Answerability: Selected Essays. New York: Oxford University Press, pp. 219–​259. Watson, Gary. 2004 [1996]. “Two Faces of Responsibility.” In G. Watson, Agency and Answerability: Selected Essays. New York: Oxford University Press, pp. 260–​288. Watson, Gary. 2011. “The Trouble with Psychopaths,” In R. J. Wallace, R. Kumar, and S. Freeman (eds.), Reasons and Recognition: Essays on the Philosophy of T. M. Scanlon. New York: Oxford University Press, pp. 307–​331. Wolf, Susan. 1987. “Sanity and the Metaphysics of Responsibility.” In F. Schoeman (ed.), Responsibility, Character, and the Emotions: New Essays on Moral Psychology. Cambridge: Cambridge University Press. Wolf, Susan. 1990. Freedom Within Reason. New York: Oxford University Press. Wolf, Susan. 2011. “Blame, Italian Style” In R. J. Wallace, R. Kumar, and S. Freeman (eds.), Reasons and Recognition: Essays on the Philosophy of T. M. Scanlon. New York: Oxford University Press, pp. 332–​347.

PA RT I I

K I N D S OF R E SP ON SI B I L I T Y

CHAPTER 4

Ac c ounta bi l i t y, Answerabili t y, a nd At tribu ta bi l i t y On Different Kinds of Moral Responsibility Sofia Jeppsson

1.  Different Kinds of Moral Responsibility: An Introduction Traditionally, moral responsibility was assumed to be one thing. Lately, however, a substantial number of philosophers distinguish between different kinds of moral responsibility. They might argue, for instance, that it makes no sense to be angry with a psychopath who cannot understand morality anyway, but it does make sense to label her cruel; there is a sense in which she is not responsible for hurting others and another sense in which she is. They might also argue that skepticism about moral responsibility does not hit every kind of responsibility there is—​even if no one deserves to be punished, for instance, moral criticism and the placing of moral demands on others might still have their place in the world. We might call philosophers who distinguish between different kinds of moral responsibility moral responsibility pluralists; and their opponents, conversely, moral responsibility monists. Pluralists furthermore argue that some kinds of moral responsibility demand a more advanced kind of free will and/​or more moral competence than others; it is therefore possible for agents to be morally responsible in some sense whereas not in another. The purpose of this chapter is to both explain and to some extent critically assess the theories and distinctions made by the most influential pluralists.

74   Sofia Jeppsson Before moving on, it is important to define what, exactly, makes responsibility X a different kind of responsibility from responsibility Y. Even if some disagreements between compatibilists and incompatibilists about moral responsibility are merely verbal ones, it certainly seems as if such disagreements can be substantial as well. Unless given really good arguments to the contrary, we should therefore assume that such substantial disagreements sometimes take place; we should thus allow for the possibility that a compatibilist and an incompatibilist are talking about the very same thing. If so, responsibility X does not count as a different kind of responsibility from Y solely because X has different or more extensive requirements on part of the agent. Rather, responsibility X should be considered a different kind of responsibility from Y only if X licenses a different kind of treatment of the agent than Y does, or provides a different justification for that treatment. It might, for instance, be the case that responsibility X gives us a pro tanto reason to angrily blame a wrongdoer whereas responsibility Y merely gives us reason to calmly criticize her, or that responsibility X licenses backward-​looking blame whereas responsibility Y merely forward-​looking. It is possible to argue that responsibility X and responsibility Y are different kinds of responsibility in the above explained sense, while holding that they have the same requirements for free will, moral competence, and so on, so that whenever an agent has X for an action she also has Y. It is not always obvious whether a given philosopher is a monist in the sense that she believes there is only one kind of moral responsibility, or if she believes responsibility X and responsibility Y to be distinct kinds although coextensive. Angela Smith, for instance, seems to argue for monism in the first sense in her 2012 paper, but has previously been interpreted as a monist in the second sense by, among others, Neil Levy (2005). The focus of this chapter, however, will be on the kind of pluralism that allows for the possibility that agents have one kind of moral responsibility while lacking another, and I will label people monists who deny this possibility. Finally, it is important to note that the terminology in the debate between pluralists and monists is inconsistent. Attributability, answerability, and accountability—​all those terms are used differently by different philosophers in different books and papers. This will become evident as I delve into different pluralist theories.

2.  A Brief Overview of Pluralism Philosophers usually become pluralists about moral responsibility in order to handle one of two problems. First, there is the problem that moral responsibility skepticism gives rise to. If no one can deserve to be praised or blamed, punished or rewarded, how are we to organize society? How are we to interact with one another? We cannot just let wrongdoers off the hook for everything. This prompts the skeptic to distinguish a less demanding kind of moral responsibility from the one that does not exist,

Accountability, Answerability, and Attributability    75 and say that this less demanding kind both exists and suffices to underpin our most important moral responsibility practices. Second, there is the problem of agents who suffer from certain kinds of mental impairments, extremely harsh and impoverished backgrounds, and the like. We often have conflicting intuitions about the moral responsibility of such agents. One way to handle such cases and explain those intuitions is to say that such agents are morally responsible for what they do in one sense but not in another. J. J. C. Smart (1961) made an early attempt to save and justify some important moral responsibility practices, despite an explicit skepticism about moral responsibility in any more substantial sense. (Smart is sometimes labeled a compatibilist, but considering how much he is skeptical about, it arguably makes more sense to label him a skeptic; see also Pereboom 2014: 2–​3.) Smart argues that there are two different ways to understand our moral responsibility practices: a clearheaded and pragmatic one, and a metaphysically confused one. He does not believe that anyone can ever deserve praise or blame, reward or punishment. After all, everything we do is ultimately the result of factors over which we lack control. Whether our actions are deterministically caused or some pure randomness is at play as well, we do not have the kind of free will that would be required in order to deserve various responses to what we do (Pereboom 2014: 293–​294). Still, we should recognize that it can be useful both to grade people’s character as better or worse, and to praise and blame, reward and punish in order to affect people’s behavior in a desirable direction (Pereboom 2014: 302–​306). Some moral responsibility practices can thus still be justified. Smart even writes, not just about the practice of holding responsible, but about people being responsible for what they do, when such grading and such reinforcing behavior is appropriate (Pereboom 2014: 305). Nevertheless, it is unclear whether it makes sense to talk about moral responsibility here. After all, we can grade and evaluate the temperament of a horse or a dog as well, and praise and blame them in an attempt to influence their future behavior. Yet we do not say that horses and dogs are morally responsible for what they do. We can grant that blame directed at humans is different in some ways from blame directed at dogs and horses—​we appeal to moral reasons when blaming humans, rather than just saying a stern “bad!” and wagging our fingers, and similarly with praise. Still, if the sole justification for blame and praise is to influence behavior in a more desirable direction, that difference seems inconsequential. The “moral” seems to fall out of “moral responsibility”. In later years, moral responsibility skeptics have advanced more nuanced theories, presenting us with a thicker and more attractive picture of the kind of moral responsibility that exists, most notably Derk Pereboom, to whom I will return later in this chapter. Gary Watson, in his influential 1996 paper “Two Faces of Responsibility,” explicitly attempts to distinguish between two different kinds of moral responsibility, where one is more demanding than the other—​and he does so in order to be able to handle cases of wrongdoers with serious psychological issues rather than to solve the problems of moral

76   Sofia Jeppsson responsibility skepticism.1 Responsibility, Watson argues, has one face of aretaic evaluation or attributability and one of accountability. In his discussion of attributability, Watson takes off from Susan Wolf ’s discussion in her 1990 book Freedom Within Reason. Before settling on her own “reason view,” according to which moral responsibility requires an ability to appreciate and act on the Right and the Good, Wolf discusses several other moral responsibility theories, among them one she calls “the real self view” (Wolf 1990: ch. 2). According to the real self view, we are morally responsible for actions that flowed from our “real self ”—​from our values rather than just our desires. Wolf notes that this “real self ” might have come about in a way that is intuitively responsibility-​undermining, and that people might act from their “real selves” despite lacking moral competence. She therefore accuses real self views of giving a much too superficial account of moral responsibility (Wolf 1990: 40–​41). This superficiality does not seem like a drawback to everyone. Neil Levy (2005) argues that there seems to be conceptual space for agents who are bad without being blameworthy—​in any sense at all. Something like the real self view, Levy writes, captures a kind of moral badness that is still distinct from moral responsibility proper. Watson, alternatively, wants to defend the thesis that what the real self view provides us with really is a kind of moral responsibility. When someone, say, hurts another person, and his doing so stems from his values, it makes sense to evaluate him from an aretaic standpoint and label him cruel. Doing so constitutes blame of a kind; it is a way of holding someone morally responsible which lies at the heart of our moral responsibility practices (Watson 1996: 234–​235). Still, Watson writes, aretaic evaluations of this kind do not capture everything about our moral responsibility practices. Sometimes we hold agents liable to social sanctions, such as blame to their face (Watson 1996: 236–​239). We hold them accountable for their actions. Accountability is the kind of moral responsibility that has been at issue in the traditional debate between compatibilists and incompatibilists. Whereas Smart is certain that traditional moral responsibility would require an impossible kind of free will, Watson is not so sure, but he does agree that an agent can only be blameworthy for an action in the accountability sense if she could have avoided doing what she did (Watson 1996: 239). When an agent is accountable for a wrongful action we are justified in subjecting her to social sanctions that are unpleasant to her, and doing so would be unfair unless she could have avoided doing what she did. We are furthermore justified in making demands on an accountable agent’s behavior, but once again, such demands make no sense unless it is possible for the agent to meet them (compatibilists and incompatibilists can, of course, debate back and forth whether we are to understand avoidability in a way that makes it compatible with determinism or not).

1  Michael Zimmerman made a similar distinction between ”liability” and ”appraisability” in his 1988 book An Essay on Moral Responsibility. These terms, however, never really caught on, and the view did not become quite popular until Watson.

Accountability, Answerability, and Attributability    77 For all these reasons, agents need more in the way of normative competence and control over their actions for accountability than for attributability. An agent can be morally responsible for his actions in the attributability sense even if he is not responsible in the accountability sense. Suppose, for instance, that an agent was brutally abused during his entire childhood, and grew up to commit heinous crimes. If this agent is so damaged by the abuse that he cannot abstain from his crimes, and cannot live up to any moral demands, he is not, according to Watson, accountable for what he does. But it is still possible that we are justified in labeling him cruel, and thus in holding him responsible in the attributability sense (Watson 1996: 240–​241). David Shoemaker (2015), building on Peter Strawson’s quality of will theory, argues that there are three kinds of moral responsibility: attributability, accountability, and answerability. He focuses on agents with neuropsychiatric and psychiatric conditions rather than deprived childhoods, but argues, like Watson, that an agent can have one kind of moral responsibility for his action while lacking another. I have, so far, focused on philosophers who want to divide moral responsibility into several kinds, either as a way of handling skepticism or as a way of handling the case of impaired agents. A third possible motive, however, is to brush aside philosophical disagreements as merely verbal. There is, for instance, a lively debate in philosophy about whether direct moral responsibility requires that the agent consciously chose what to do, or whether it is sufficient that the action in some sense expressed who the agent is. Levy and McKenna (2009: 118) suggest that these disagreements are frequently merely verbal; the former philosophers discuss accountability and the latter attributability. They cite, among others, Angela Smith, who writes that conditions might be more demanding if we are to impose social sanctions on someone than merely judging that she is blameworthy (Smith 2008: 378). Later, however, Smith explicitly denies that she merely talks about attributability when she calls conscious choices unnecessary for responsibility (Smith 2012: 576). Influential libertarian Robert Kane believes that compatibilists about moral responsibility and determinism must ultimately fall back on some more or less consequentialist analysis of moral responsibility (1998: part one). Such moral responsibility might very well be compatible with determinism, but ultimate moral responsibility, as Kane is interested in, requires, he argues, libertarian free will. Now, even if Kane were right, there would be room for a substantial disagreement between him and the compatibilists; they could still disagree about which kind of freedom or free will was important or worth wanting. I do not believe there is reason to accept, though, that compatibilists in general discuss a consequentialist or watered-​down concept of moral responsibility. Influential compatibilists like Fischer and Ravizza (1998), for instance, seem to me to discuss the same kind of non-​consequentialist responsibility as Kane does. It should be clear, by now, that we do not merely have monism on the one hand and pluralism on the other. There are a variety of pluralist theories, disagreeing with each other as well as with the monists. Things would certainly be simpler if moral responsibility was one thing only. So what are the arguments in favor of monism?

78   Sofia Jeppsson

3.  Smith’s Moral Responsibility Monism It is much more common to simply assume moral responsibility monism than to explicitly argue for it. It is fair to say that the majority of philosophers who have discussed moral responsibility throughout the years have been monists. Pluralists are still more the exception than the norm. Thus, it is the philosophers who believe that there are more than one kind of moral responsibility who feel that they have the burden of proof on their side, and actually have to argue for their view. Angela Smith, however, explicitly argues for monism. Smith builds her theory of moral responsibility primarily on T. M. Scanlon’s responsibility theory in his 1998 book What We Owe to Each Other. Scanlon argues, like Smart, that there is no such thing as desert (although he believes this to be true independently of which kind of free will exists, Scanlon 1998: 274–​275). Unlike Smart, he argues that influenceability is neither necessary nor sufficient for moral responsibility. Moral responsibility depends, rather, on whether the action resulted from a judgment-​sensitive attitude of the agent’s (Scanlon 1998: 270–​274).2 Smith further develops this theory, and forcefully argues for moral responsibility monism. Moral responsibility is, at its heart, about placing demands on other people and holding them answerable for flouting these demands. When we angrily blame others, it is because they flouted a particularly serious demand, not because they are morally responsible in a different sense than the regular one, and we likewise only engage in aretaic grading of people whose behavior they can reasonably be held answerable for. Smith argues for monism against the background of a paper of Shoemaker’s where he lays out an early version of his so-​called “tripartite theory of moral responsibility” (Shoemaker 2011). In this paper, Shoemaker interestingly claims that even our interactions with regular, unimpaired agents make sense only if we assume that there are more than one kind of moral responsibility. Shoemaker and Smith agree that in order to be responsible for an action in the answerability sense, it must bear a rational relation to your judgments. When this is the case, it makes sense to demand that the agent explains herself (Shoemaker: 605; Smith 2012: 576–​577). This is not supposed to be a particularly demanding requirement. Smith is quite explicit in claiming that the scope of things for which we can be answerable is pretty wide (Smith 2012: 577–​578). When I think to myself, at the bar with my friends, that I ought to call it a night and go home, but akratically end up having another beer instead, my having another beer still bears a rational relation to some of my evaluative judgments; I judge, for instance, that beer is good and hanging out with my

2  Scanlon (2008: ch. 4) would later develop a theory of blame according to which blaming is a way to recognize that someone has done something to impair our relationship with her, but that theory need not concern us here.

Accountability, Answerability, and Attributability    79 friends is nice. It therefore also makes sense to ask me why I had another beer instead of going home. However, whereas Smith argues that when an agent is answerable, she is also morally responsible full stop, there are no other kinds of moral responsibility; Shoemaker wants to argue that accountability and attributability are distinct kinds. Accountability, according to Shoemaker, has the same rational relations requirement as answerability. The difference between them is first, interestingly enough, which kind of wrongdoing the agent is responsible for: whether he has merely failed to live up the hopes that the other party had for their relationship, which would make his responsibility answerability; or failed to satisfy the basic demands of said relationship, which would make it accountability. Shoemaker’s examples are of a husband who either fails, year after year, to buy his wife an anniversary gift that she actually appreciates (answerability), or who cheats on her (accountability). Secondly, accountability, but not answerability, licenses angry blame (Shoemaker 2011: 621–​622). Attributability is distinguished from both answerability and accountability in that it does not require that the rational relationship requirement is satisfied (Shoemaker 2011: 607–​609). This tripartite theory is definitely less convincing than the one Shoemaker lays out in his later book. Regarding the supposed distinction between answerability and accountability, Smith convincingly argues that this is merely a difference between being morally responsible for serious and less serious wrongs (Smith 2012: 586–​589). If someone were to violate the basic demands of our relationship, this is a more serious case of wrongdoing than if someone fails to fulfill the hopes I have. I might therefore be justified in being actually angry in the first case, but merely disappointed in the latter. Still, there is no need to posit that there are two different kinds of responsibility at play here in order to explain my having different reactions in the two scenarios. Besides, it is also possible and not at all out of the ordinary to be a little bit angry, not just disappointed, with someone who commits a less serious wrong. It is not the case that we only ever get angry with people who violate basic relationship demands. When it comes to attributability, furthermore, Smith claims that depending on how we fill in the details, it either collapses into answerability, or it is not truly a kind of moral responsibility at all. Smith considers two examples of Shoemaker’s, intended to show that we can be responsible in the attributability sense for something despite the fact that it does not bear a rational relation to our judgments. Firstly, Shoemaker discusses irrationality, as when an arachnophobic both fears spiders and judges them to be harmless. Shoemaker argues that this irrationality is attributable to the arachnophobic, despite the fact that she is not answerable for it, since she does not judge irrationality to have any positive value. Smith’s reply is that the arachnophobic can be answerable both for her (emotional and implicit) judgment that spiders are dangerous and for her judgment that they are harmless, and since her irrationality is constituted by the combination of these two judgments, she is thereby answerable for her irrationality as well (Smith 2012: 579–​580). Secondly, Shoemaker points to cases where we deeply care about something contrary to our judgment (and these cases need not involve any kind of impairment or mental illness on the part of the carer); for instance, a parent might care about his son

80   Sofia Jeppsson despite the fact that the son grew up to be a serial killer and the parent judges that he is a horrible human being. Smith’s reply here is, firstly, that she doubts the plausibility of the case as stated. The parent might think it horrible that his son is a serial killer, but unless he has also experienced some good sides of his son’s personality and judges those sides valuable, it is implausible to suppose that he cares deeply about him. Thus, Smith argues, in a plausible case, his care and actions stemming from his care will bear a rational relation to some of his evaluative judgments. Smith does say, though, that whether a case with a deeply caring parent who does not judge at all that the son for whom he cares has any good sides whatsoever is plausible or not is an empirical psychological question, and she therefore does not want to put too much weight on that response to Shoemaker. What she does strongly object to, however, is his thesis that pure attributability really is a kind of moral responsibility (Smith 2012: 581–​584). If a parent were to care for a serial killer son despite not judging any of his qualities whatsoever being at least somewhat good, he would not be morally responsible for so caring. Shoemaker insists that attributability judgments of the kind “he is cruel,” “she is arrogant,” “he is generous,” and so on go beyond mere description; they explain why agents act as they do (even when the agent cannot provide reasons for why she did what she did when asked) and have important implications for whom we want to interact with and why (Shoemaker 2011: 612–​615). Smith argues that this explanation as to why attributability is a kind of moral responsibility seems to imply that an agent is, for instance, morally responsible in this sense for OCD behavior. The fact that an agent suffers from OCD can explain why she acts the way she does, and have important implications for whether we want to interact with her (perhaps her OCD makes normal friendships next to impossible to maintain). The same can be said about certain other symptoms, such as severe depression. And yet, intuitively, agents are not morally responsible in any sense (or if they are, it is to a very small degree rather than to the full degree that the theory of attributability seems to imply) for actions that stem from severe mental illnesses. Smith’s suggestion is that aretaic appraisals are just a different way of holding an agent answerable, or morally responsible full stop, for what she does. We do not say that the OCD sufferer who dare not leave the house for fear of being contaminated by germs acts “cowardly”; we reserve that label for behavior that is motivated by fear to an unjustifiable extent and is such that her behavior bears a rational relation to her judgments (Smith 2012: 584–​586). Interestingly, Smith suggests in passing that “responsibility as eternal comeuppance,” Heaven and Hell responsibility, might indeed be a different kind, possibly one that cannot exist, but she argues that our actual existing moral responsibility practices can all be explained in terms of answerability (Smith 2012: 588).3 As long as we stick to Earthly matters, there is no need to posit the existence of other kinds of moral responsibility. 3  Fischer and Tognazzini (2011) similarly argue that whether an agent can deserve eternal damnation is a separate issue, not to be confused with ordinary responsibility attributions.

Accountability, Answerability, and Attributability    81 As interesting as questions about Heaven and Hell responsibility are, however, I will not delve into them in this chapter. What Smith has shown in her discussion with Shoemaker is that as long as we stick to regular agents, without special impairments, and as long as we steer clear of skepticism, pluralism is hard to motivate. Let us now turn to the issue of whether pluralism is needed to handle impaired agents, like autistics and psychopaths.

4.  Responsibility and Impaired Agents In his 2015 book, where Shoemaker lays out a more developed version of his tripartite theory of responsibility, the focus is on mentally impaired agents, and how the tripartite theory can explain the ambivalence we might feel regarding their responsibility; impaired agents, Shoemaker argues, are frequently responsible in one or two senses while lacking responsibility in another. His ambition is to motivate and build his theory on real-​life cases rather than philosophical thought experiments. However, there is still a tendency, throughout the book, to focus on extreme cases at the expense of perhaps more common ones. Shoemaker works in Peter Strawson’s quality of will tradition here, although he argues that the will can be divided into a character, a judgment, and a regard part. We can react to one or two of those “will parts” even if the third one is missing. Now, I wrote in the beginning of this chapter that we should distinguish between the kind of reactions that a certain kind of responsibility licenses, and what that kind of responsibility requires on part of the agent. For instance, a compatibilist and incompatibilist about desert-​entailing moral responsibility can agree that when an agent is morally responsible in this sense, backward-​ looking blame and praise can be justified, but whereas the compatibilist believes that an agent can be morally responsible in this sense as soon as she is reasons-​responsive and has a certain moral competence, the incompatibilist believes that she needs libertarian free will as well. In Shoemaker’s book, however, the agential requirements and the kind of judgments and expressions of attitude that the different kinds of responsibility licenses are tied very closely together. According to Shoemaker, when an agent expresses her quality of character in action, she is responsible for the action in the attributability sense. When she expresses her quality of judgment, she is responsible in the answerability sense, and when expressing her quality of regard, she is accountable. Actions that are attributable to an agent primarily give rise to disdain or admiration, answerable actions to regret or pride (and third-​person analogues of those) and accountable actions to anger or gratitude (Shoemaker 2015: 35). Shoemaker then goes through a number of different impairments, stating for each one which kinds of moral responsibility agents with the impairment can have and which are beyond them. For the purposes of this chapter, I will focus on autism and psychopathy.

82   Sofia Jeppsson Starting with autistic people, they are said by Shoemaker to lack a capacity for regard for other people, and therefore they can only be morally responsible in the attributability and answerability senses, but not in the accountability sense.4 Their lack of accountability responsibility explains why we ought not to get angry with autistics who, say, make offensive remarks that hurt other people’s feelings. It is interesting to compare this neat view on autistics—​they lack this capacity and so they cannot have this responsibility but they can have that—​with the far messier picture painted by Richman and Bidshari (2018). Richman and Bidshari claim that depending on which moral responsibility theory is correct, which theory of autism is correct (since there are conflicting scientific theories regarding what the core of the condition is), and finally which difficulties this particular individual has and the details of the situation in which she finds herself, autistics might or might not exempt an individual from moral responsibility. It strikes me as plausible that if we really want to discuss real-​life cases rather than “philosophy autistics,” “philosophy psychopaths,” and so on, conclusions as complicated and messy as Richman and Bidshari’s are likely to come out, rather than a neat statement according to which people with a certain impairment can have responsibility X while they cannot have responsibility Y. Overall, it is not obvious that we need to postulate the existence of different kinds of moral responsibility in order to handle cases of autistic agents, since there might be alternative solutions available to explain why we often should not get angry with an autistic person who, for example, makes an offensive remark. If someone suffers an impairment which makes it difficult for her to understand that making a certain remark is likely to cause hurt feelings in another, we might argue that she is excused due to non-​culpable ignorance about the fact that what she said was harmful. Non-​culpable ignorance is a familiar excuse which can be invoked by moral responsibility monists as well as pluralists, and the monist might have other resources at hand as well. It is thus not quite obvious that we need to invoke different kinds of moral responsibility in order to handle autism. Psychopaths, according to Shoemaker, can likewise have answerability and attributability responsibility but not accountability, since psychopaths too lack a capacity for regard for others. Here, Shoemaker jumps into a very lively philosophical debate. Clearly many people have the intuition that psychopaths who hurt others are properly labeled “cruel” despite having an arguably responsibility-​undermining impairment, but the intuition is not universal. Nelkin (2015: 366–​368) questions whether someone can really be cruel if she does not merely disregard the feelings of other people, but lacks the very capacity needed for understanding that someone else’s distress is a reason against hurting her (she leaves it open whether psychopaths in real life completely lacks this capacity or not). After all, there is a difference between disregarding the feelings of others and simply being unable to take those feelings into account. It is 4  Claims about empathy deficits and ”lack of regard” in autistic people are highly controversial. See Anna Stenning (2020) for a critical discussion of the research and the moral conclusions drawn.

Accountability, Answerability, and Attributability    83 thus not obvious either that the case of psychopaths is best handled through a pluralistic moral responsibility theory. However, even though Nelkin disagrees with Shoemaker about psychopaths, she is open to the possibility that there are agents who can properly be labeled “cruel” although lacking accountability for their actions (Nelkin 2015: 383). An example might be a racist person who was raised in a fully racist environment, like an 18th-​century slave owner. Even if the psychopath lacks moral responsibility period, we might want to hold on to the judgment that there is a sense in which the 18th-​century slave owner is morally responsible for keeping slaves and treating them badly, and another sense in which he is not. Unlike the psychopath, he might properly be said to show contempt and disregard for his slaves, so that he is properly labeled “cruel,” but if he never really had an opportunity to question his racist values or develop different ones, we might still want to say that there is an important sense in which he is not morally responsible for what he did. It might be that even if agents with neuropsychiatric impairments can be handled by monism, we need pluralism to make sense of agents like the slave owner. Ultimately, whether it makes sense to label a psychopath “cruel” or not (if we imagine, for the sake of argument, that she is completely incapable of regard for others, not just somewhat impaired), might come down to brute intuitions. It is not obvious to me that this is a question that can ever be fully resolved. The intuitions of philosophers seem to point in different directions, and we might not be able to resolve things by calling on the intuitions of the folk either. Even if we establish that most people tend to call real-​life psychopaths “cruel”, it would be hard to determine whether this is because most people believe that attributability judgments do not require a capacity for regard for others, or because most people suspect that real-​life psychopaths have this capacity to some extent. Even if our intuitions regarding psychopaths who hurt others are truly conflicting, there might be strategies for handling them available to the monist. Neil Levy (2014) argues that due to their impairments, psychopaths are not quite capable of realizing the special wrongness involved in hurting persons. Due to their impairment, psychopaths cannot be morally responsible (supposedly not in any sense) for hurting other persons, although they can be morally responsible for hurting animals (including cases where the animal in question is a human being). Special mention should also be made of Matt Talbert (2012), who argues that it makes perfect sense to blame psychopaths for their wrongdoing, since blame is a way of standing up for oneself and retaining one’s self-​respect. Arguably, Talbert is talking about a different kind of blame, blame-​as-​standing-​up-​for-​oneself, than, say, the more commonly discussed blame-​as-​the-​dealing-​out-​of-​deserved-​social-​punishments or blame-​as-​deterrence-​from-​bad-​behaviour, and a correspondingly different kind of moral responsibility. Talbert himself does not suggest that he discusses any special kind of moral responsibility, but if different kinds of moral responsibility are individuated on the basis of which kind of response they license and how that response is justified, as I have suggested, it might be considered one.

84   Sofia Jeppsson

5.  Responsibility, Free Will Skepticism, and Basic Desert If Shoemaker is an heir to Watson, Derk Pereboom is rather an heir to Smart, although the kind of moral responsibility he argues exists is of a much deeper and more sophisticated kind than the one found in Smart’s theory. Pereboom does not distinguish between accountability, answerability and attributability, but between answerability and moral responsibility in the basic desert sense. Is the latter the same kind of moral responsibility as accountability? That depends on how accountability relates to desert. Both Shoemaker and Michael McKenna think it possible to discuss accountability and what it requires on part of the agent without delving into the issue of desert; these philosophers believe that the connection between the two, if there is one, can be saved for later (Shoemaker 2015: 222; McKenna 2012: 171–​172). After delving into the issue, however, Dana Nelkin tentatively concludes that at least accountability blameworthiness implies that the blameworthy person also deserves to be blamed, in the sense that blaming her is fitting (Nelkin 2016a). (The fact that doing so is fitting, furthermore, might provide one with a reason to blame her, depending on further factors.) If so, it makes sense to equate Pereboom’s “moral responsibility in the basic desert sense” with accountability. A further connection between the two is that moral responsibility in the basic desert sense, on Pereboom’s view, licenses angrily blaming wrongdoers even if no good consequences are likely to come out of it, and the same holds true for accountability on Watson’s and Shoemaker’s views. I will therefore, from now on, use “accountability” instead of the longer and more cumbersome “moral responsibility in the basic desert sense”, on the assumption that they come to the same thing. However, we might ask whether it can really be fitting to angrily blame someone for what she did, in a backward-​looking sense not dependent on any good consequences coming out of so blaming her, if the world is deterministic. Pereboom is a hard incompatibilist about accountability. In a deterministic world, no one deserves to be blamed for what they did, and we ought never to be angry with wrongdoers. In order to provide an argument for his answer, he presents us with the four-​case manipulation argument, featuring Professor Plum, manipulated in various ways into murdering Ms White for selfish reasons (Pereboom 2001: 110–​117; 2014: 74–​ 79). Plum is stipulated to fulfill any moral responsibility condition suggested by popular compatibilist theories; he is responsive to reasons, he acts in character, he is morally competent and so on. However, in the first case presented to us in the thought experiment, Plum is manipulated to act as he does by neuroscientists using radio-​like technology. In the second case, the scientists created him at birth in such a way as to ensure that he would grow up to be rationally egoistic and thus decide to murder Ms. White when he stood to gain enough from doing so. In the third case, there are no scientists, just a community bringing Plum up to be selfish since this community values selfishness.

Accountability, Answerability, and Attributability    85 Finally, in the fourth case, we have ordinary determinism. Pereboom now states that the difference between each case is morally irrelevant. Therefore, if Plum lacks accountability in case 1 when he is directly manipulated, he lacks it in every other case as well, including case four, ordinary determinism. Pereboom’s conclusion is that accountability does not exist in a deterministic universe. (Since Pereboom believes that it makes no difference to the argument whether we add a little randomness here and there, he also concludes that accountability does not exist under plausible versions of indeterminism either; he is not just a hard determinist, but a hard incompatibilist. Pereboom 2001: chs. 2–​3; 2014: chs. 2–​3). John Fischer initially seemed to bite the bullet really hard when it comes to Professor Plum, claiming that he is morally responsible throughout. In a paper with Neil Tognazzini (2011), however, Fischer backpedals a little bit, and invokes distinctions between different kinds of moral responsibility. In the paper, Fischer and Tognazzini want to elaborate further on Watson’s distinction between attributability and accountability, by enumerating a large number of questions necessary to determine whether someone was morally responsible at all for a wrongful action, in which way she was responsible, and whether anyone has the standing to blame or punish her. Fischer does not want to claim, he writes, that Plum is blameworthy in the accountability sense throughout all the manipulation cases—​merely in the much weaker attributability sense. At least in case one, the manipulation is “too invasive” to preserve accountability blameworthiness, and Plum is therefore excused for the murder, even though it is attributable to him (Fischer and Tagnazzini: 400–​404). Fischer thus denies that there are no morally relevant differences between the various cases, even though he does not elaborate on what it is that makes manipulation more or less invasive, and when manipulation undermines accountability and when it does not (something which would be needed in order to provide a full answer to the challenge presented by Pereboom’s thought experiment). Pereboom, conversely, consistently claims that there are no morally relevant differences between the cases; Plum lacks accountability throughout, but he has moral responsibility in the answerability sense, also throughout (see Pereboom 2001: 156–​157; 2013; 2014: 132–​134, for the distinction between accountability and answerability, and 2014: 136, for the claim that Plum is answerable). A similar distinction between accountability and answerability (although with different terminologies) can be found in Parfit (2011: 261 and ch. 11), Kane (1998) and Kelly (2002). Since I have previously discussed Shoemaker’s tripartite theory of which “answerability” is part, it should also be pointed out that Pereboom does not mean quite the same thing as Shoemaker. For Shoemaker, “answerability” is narrowly concerned with the quality of an agent’s judgment. “Answerability” for Pereboom seems to be concerned with regard for others as well. An agent is answerable for what she did in Pereboom’s sense when she was responsive to reasons, so that criticism for wrongful behavior (and commendation for exemplary behavior) can have the beneficial effects of improving moral understanding, protecting victims by discouraging future wrongdoing, and improving relationships (Pereboom 2001: 156–​157; 2013: 51; 2014: 132–​134. See also Kelly 2002: 194). It seems clear from these goals that Pereboom’s “answerability” encompasses more than Shoemaker’s

86   Sofia Jeppsson “answerability” does; it is hard to see how someone who was only capable of judgment, not regard, could have meaningful relationships or any real moral understanding. It is important for Pereboom that answerability, which Plum has, is a pretty substantial kind of moral responsibility. A number of philosophers argue that life would be terribly bleak if we ceased to believe that we have moral responsibility for what we do, and also began acting as if moral responsibility skepticism was true (e.g., Strawson 1962/​2013; Wolf 1981). And it’s easy to see how an important part of morality and even of our self-​conception would be lost if the only kind of moral responsibility we had was the shallow kind advocated by Smart. Pereboom wants to paint a positive picture of life under skepticism, and part of this project is to drive home that answerability, the kind of moral responsibility that exists, is thick and deep enough to be satisfying. However, this requires a bit of a balancing act from Pereboom’s side. If answerability is conceived of as thick enough, as obviously moral enough, it becomes less intuitively plausible that manipulated Plum can truly be morally responsible in this sense either. We might ask whether manipulated Plum is a true moral agent in a thick enough sense for his actions to be labeled “right” and “wrong” rather than just “desirable” and “undesirable.” We can ask whether it is possible to truly reason with Plum and truly criticize him, rather than just pragmatically say things that we think might influence his behavior for the better (Jeppsson 2016: 701–​705). And if manipulated Plum is not a real moral agent, and Pereboom is right when he says that there are no morally relevant differences between the cases, the same is true of us, and skepticism would be as bleak as philosophers like Strawson and Wolf have claimed. In order to show that skepticism does not threaten anything important, Pereboom thus has to make answerability thick enough to be satisfying, but not so thick that manipulated Plum is unlikely to have it.

6.  To Sum Up Summing up this chapter, philosophers usually embrace pluralism for one of two reasons. First of all, we have the issue of agents with various impairments, harsh childhoods, and the like that seem to affect their moral responsibility in some way, without bringing them all the way out of the moral responsibility ballpark. Some philosophers argue that such agents are morally responsible in one or two senses while lacking responsibility in another. Second, there are moral responsibility skeptics who deny that skepticism is impossible to live by; they want to argue that there is a kind of moral responsibility that still exists, and suffices to ground our most important moral responsibility practices. Pluralism is arguably more crucial for the skeptics. If we assume monism, and combine it with skepticism, it is hard to see how anything resembling our current life would be possible. Of course, a philosopher could consistently claim that the only thing properly labeled moral responsibility is something which does not exist, and proceed to use different terms for what does exist. Still, any skeptic who wants to paint a positive picture

Accountability, Answerability, and Attributability    87 of life under skepticism must argue that something which is recognizable as a kind of moral responsibility does exist and suffices to underpin at least some moral responsibility practices. When it comes to impaired agents, it is less obvious that pluralism is needed. One might instead argue that difficulties can diminish moral responsibility without rendering it nonexistent (Nelkin 2016b; Wolf 1990: 86–​87), and that people who have various impairments that make acting morally more difficult for them therefore have diminished moral responsibility for their wrongdoings, without invoking different kinds of moral responsibility. One might also argue that when an impairment like, say, an inability to feel proper regard for other people, becomes severe enough, it undermines moral responsibility completely, and that this is intuitively plausible upon reflection, even if our initial reaction to some such cases was more ambivalent. For some impaired agents, the excuse of non-​culpable ignorance might be relevant. Of course, as we saw in the discussion of Shoemaker’s 2011 paper, a philosopher might argue that pluralism is needed even to just make sense of our ordinary responsibility practices for ordinary agents. Michael Zimmerman (2015) argues that different kinds of moral responsibility are individuated by which reactive attitude they make appropriate, and that there are therefore as many kinds of moral responsibility as there are reactive attitudes. Even though I, too, argue that kinds of moral responsibility are individuated by which kind of treatment they license for the wrongdoer, I believe we have to group treatments into fairly big and rough groups. Otherwise, we are simply stuck with too many different kinds of moral responsibility for them to be useful in responsibility discussions. Even if we ultimately ought to embrace pluralism and the existence of two or three different kinds of moral responsibility, surely there comes a point when the positing of additional kinds merely becomes unwieldy and downright confusing.

References Fischer, John and Tognazzini, Neal. 2011. “The Physiognomy of Responsibility.” Philosophy and Phenomenological Research 82: 381–​417. Fischer, John Martin and Ravizza, Mark. 1998. Responsibility and Control: A Theory of Moral Responsibility. New York: Cambridge University Press. Jeppsson, Sofia. 2016. “Accountability, Answerability and Freedom.” Social Theory and Practice 42: 681–​705. Kane, Robert. 1998. The Significance of Free Will. New York: Oxford University Press. Kelly, Erin. 2002. “Doing Without Desert.” Pacific Philosophical Quarterly 83: 180–​205. Levy, Neil and McKenna, Michael. 2009. “Recent Work on Free Will and Moral Responsibility.” Philosophy Compass 4(1): 96–​133. Levy, Neil (2014). Psychopaths and blame: The argument from content. Philosophical Psychology 27 (3):351–​367. McKenna, Michael. 2012. Conversation and Responsibility. Oxford: Oxford University Press. Parfit, Derek. 2011. On What Matters. Volume One. Oxford: Oxford University Press. Pereboom, Derk. 2001. Living Without Free Will. Cambridge: Cambridge University Press.

88   Sofia Jeppsson Pereboom, Derk. 2013. “Free Will Skepticism and Criminal Punishment,” in the Future of Punishment, edited by Thomas Nadelhoffer: 49–​78. New York: Oxford University Press. Pereboom, Derk. 2014. Free Will, Agency and Meaning in Life. New York: Oxford University Press. Nelkin, Dana Kay. 2015. “Psychopaths, Incorrigible Racists, and the Faces of Responsibility.” Ethics 125: 357–​390. Nelkin, Dana Kay. 2016a. “Accountability and Desert.” Journal of Ethics 20: 173–​189. Nelkin, Dana Kay. 2016b. “Difficulty and Degrees of Moral Praiseworthiness and Blameworthiness.” Noûs 50(2): 356–​378. Richman, Kenneth A. and Bidshahri, Raya. 2018. “Autism, Theory of Mind and the Reactive Attitudes. Bioethics 32: 43–​49 Scanlon, T. M. 1998. What We Owe to Each Other. Cambridge MA: The Belknap Press of Harvard University Press. Scanlon, T. M. 2008. Moral Dimensions: Permissibility, Meaning, Blame. Cambridge, MA: Belknap Press of Harvard University Press. Shoemaker, David. 2011. “Attributability, Answerability and Accountability: Toward a Wider Theory of Moral Responsibility.” Ethics 121: 602–​632. Shoemaker, David. 2015. Responsibility from the Margins. Oxford: Oxford University Press. Smart, J. J. C. 1961. “Freewill, Praise and Blame.” Mind 70: 291–​306. Smith, Angela M. (2008). Control, responsibility, and moral assessment. Philosophical Studies 138 (3):367–​392. Smith, Angela. 2012. “Attributability, Answerability, and Accountability: In Defense of a Unified Account.” Ethics 122 (3): 575–​589. Stenning A. 2020. “Understanding empathy through a study of autistic life writing. On the importance of neurodivergent morality.” Neurodiversity Studies. A new critical paradigm. Edited by Hanna Bertilsdotter Rosqvist, Nick Chown and Anna Stenning, pp. 108–​124. London: Routledge. Strawson, Peter. 1962/​2013. “Freedom and Resentment.” Reprinted in The Philosophy of Free Will. Edited by Russell, Paul and Deery, Oisin. New York: Oxford University Press. Talbert, Matthew. 2012. “Moral Competence, Moral Blame and Protest.” Ethics 16: 89–​109. Watson, Gary C. 1996. “Two Faces of Responsibility.” Philosophical Topics 24: 227–​248. Wolf, Susan. 1981. “The Importance of Free Will.” Mind 90: 386–​405. Wolf, Susan. 1990. Freedom Within Reason. New York: Oxford University Press. Zimmerman, Michael J. 1988. An Essay on Moral Responsibility. Totowa Zimmerman, Michael J. 2015. “Varieties of Moral Responsibility.” In The Nature of Moral Responsibility: New Essays. Edited by Randolph Clarke, Michael McKenna and Angela Smith. Oxford: Oxford University Press.

PA RT I I I

DI M E N SION S OF R E SP ON SI B I L I T Y

CHAPTER 5

Re sp onsibilit y for Ac ts and Om is si ons Randolph Clarke

We commonly think that we are morally responsible not just for actions and (some of) their consequences, but also for omitting to act, refraining or abstaining from action, forbearing action, and (some) consequences of these. A parent’s omission to remove a child from a parked car can result in the child’s death, something for which the parent might be to blame. One can be praiseworthy for refraining from retaliating for an offense. And responsibility for omitting to attend, inquire, or investigate often figures crucially in cases of negligence. A comprehensive theory of responsibility, then, will include in its coverage omitting and refraining (I’ll generally abbreviate the list to these) and their upshots. If omitting or refraining were in every case performing an action, then we might expect that the conditions that are individually necessary and jointly sufficient for responsibility for action would apply straightforwardly to cases of omitting or refraining.1 But (as I’ll argue) the antecedent of this hypothetical isn’t true. It is to be expected, then, that what is required for responsibility for omitting or refraining differs from what is required in the case of action. Just what the differences are is a relatively unexplored question in philosophical work on responsibility. The answer, I believe, is complicated by the fact that cases of omitting or refraining are quite varied, and an account of agents’ responsibility in these cases will have to reflect this variety.

1  However, as I observe in section 3, even when such an identity holds, there might be a difference in what is required for responsibility.

92   Randolph Clarke

1.  Some of the Variety In many cases in which an agent does not do a certain thing, the agent intentionally does not do that thing. These can be cases of refraining, abstaining, or forbearing. This class of cases itself varies along several dimensions. For one thing, some of these cases might be called episodic, while others are what we might call periodic. An example of the former: Ann refrains from reaching for a chocolate when she sees some set out on a table at a reception. An example of the latter: Bob abstains for years (thirty, so far) from eating meat. Episodic cases of intentionally not doing a certain thing vary in their relation to action. In some of these there is, arguably, a simple identity of the agent’s not doing the thing in question with her performing an action. A mime stands on a sidewalk, frozen in mid-​ gesture, for many minutes refraining from moving in any noticeable way. Maintaining the pose requires the exertion of effort, making fine adjustments to muscle tension, in response to careful self-​monitoring, for the maintenance of balance and position. The mime is active in doing this; she performs an action of holding her pose. What is the mime’s intentionally not moving, her refraining from moving? There is a good case for saying that this is simply her action of holding the pose, but now negatively described, described in terms of something it is not (it is not an instance of moving).2 The refraining may fairly be said to take place when and where the act of holding still does, to be just as static and as difficult as the act is, and to have that act’s causal features—​caused by all and only the things that cause the act and causing all and only the things that the act causes. Another kind of case in which there may be said to be an identity has, in conception, a teleological structure that cases of simple identity lack. This second kind includes cases of what Bruce Vermazen (1985: 103) calls “displacement refraining,” doing one thing in order to prevent oneself from doing another. If Ann doubts her self-​control when she sees the chocolates on offer, she might shove her hands into her pockets or grasp them tightly behind her back as a means of keeping herself from reaching for a chocolate. She might, in performing such an action, intentionally bring it about that she doesn’t reach. In this kind of case, the agent conceives of the action she performs as something done in order to prevent herself from performing the action from which she refrains. Here, too, we might identify the agent’s refraining with an action, in this case the action by performing which she intentionally doesn’t reach.3 2  If this is correct, then expressions such as “the mime’s refraining from moving” can be and sometimes are used referentially, with success, to pick out actions. Writers affirming this view include Clarke (2014: 21–​28), Davidson (1985), Schaffer (2012), and Varzi (2006, 2008). Moore denies it, issuing the following challenge: “Imagine picking out objects by the properties they do not have” (2009: 438). Of course, we can easily do so, e.g., referring to Eris as the most massive non-​planet directly orbiting the Sun. 3  Identifying Ann’s refraining from reaching with (say) her action of shoving her hands into her pockets requires that sometimes when one A-​s by B-​ing we can identify one’s A-​ing with one’s B-​ing. Several views of act individuation allow such identification. On this point, see Clarke (2014: 24–​25).

Responsibility for Acts and Omissions    93 But episodic cases of intentionally not doing something are not all of either of these two kinds. Ann might have no lack of confidence in her self-​control; indeed, she need not experience any temptation to reach for a chocolate. She might still intentionally not reach for one. When she does, she might remain near the table, her arms at her sides. It should not be thought that her refraining must be an act of keeping her arms at her sides, for she might not perform any such action. Her arms, we may suppose, were hanging freely at her sides before she decided not to reach for a chocolate, and whatever was causing them to do so then may be all that is causing them to do so after she has so decided. It might be that she no more needs to keep her arms from moving toward the chocolate than she needs to keep the chocolate from moving toward her arms. In a case of this sort, it is doubtful that there is any action that the agent is performing that is (identical with) her refraining from reaching. It should not be thought that refraining can generally be identified with an agent’s decision not to do what she refrains from doing, for at least two reasons. First, decisions—​ even decisions not to do certain things—​are generally not conceived of teleologically (by the agents who make them), as means to doing the things decided on. Hence the consideration that, with respect to cases of displacement refraining, favors identifying the refraining with some overt action is not present here to favor identifying the refraining with the making of the decision. Second, one can intentionally not do a certain thing without having decided not to do it.4 Even if refraining from A-​ing requires having an intention not to A,5 one can come to have such an intention other than in making a decision. Decisions resolve practical uncertainty, uncertainty about what to do. But (in a variant of the case) Ann might have had no uncertainty about whether to reach for a chocolate; she might have come to intend not to do so as soon as she saw them, without any question arising about whether to reach. Cases of periodic refraining or abstaining, too, frustrate attempts to find some action that is identical with the agent’s intentionally not doing something. Certainly no action that Bob has performed in the last 30 years is his abstaining over that period from eating meat. Some fusion of actions—​say, all of the acts of eating that Bob has performed during this period—​might be proposed as what his abstaining is, but Bob can claim to have abstained for the 30-​year period, not just on the occasions when he ate. (Perhaps without his policy of abstention he would have eaten meat on some other occasions.) If credit is due, then Bob (like an addict in a 12-​step program) gets it for the whole period. On many occasions when someone omits to do a certain thing, the agent does not intentionally not do that thing. Indeed, in many of these cases, the agent isn’t even aware

4  Pace Bach (2010: 52–​54) and Moore (1979), who take deciding not to A to be necessary for refraining from A-​ing. P. Smith (1986: 14) similarly holds that refraining requires performing some mental action in order to prevent oneself from doing the thing that one refrains from doing. 5  Although I’m inclined to think that intentionally not A-​ing does require having an intention with some relevant content, I do not think that it requires having an intention not to A. On this point, see Clarke (2014: 65–​7 1).

94   Randolph Clarke that she isn’t doing the thing in question. We may call these cases of unwitting omission.6 Again, these may be episodic or periodic. An example of the first: Carla has the duty of turning on the pool pump on evenings when an overnight freeze is forecast (a precaution against burst pipes), and she routinely fulfills it; but one evening, despite knowing that a freeze is forecast, she forgets to turn on the pump and omits to do so. An example of the second: David has a standing plan to pull weeds in the garden for at least an hour each month, but he omits to do so in June, not realizing until the month is over that he has neglected the garden. An unwitting omission is a failure to exercise agency; it is not itself an exercise of agency. Not being such, it isn’t an action. It is thus wrongheaded to seek an action performed by the agent in question with which such an omission might be said to be identical. Indeed, it is not clear that there is any entity of any kind that is an unwitting omission. Certainly people unwittingly omit to do this or that; but that truth does not obviously imply that when one unwittingly omits to do a certain thing, there is something that is one’s omission.7

2.  A Familiar Distinction and Two Conditions It is common in discussions of moral responsibility to distinguish between, on the one hand, direct or basic responsibility and, on the other, indirect or derivative responsibility. An agent bears the latter for something if she bears responsibility for it that derives from her responsibility for something else. The derivation is, in part, explanatory: when one is indirectly responsible for something, one is responsible for it at least partly because one is responsible for something else. Direct responsibility for something is responsibility that is not in this way derivative.8

6 

As “omit” and “omission” are used in ordinary English, only rarely when an agent unwittingly does not do a certain thing do these terms apply. A moment ago, without realizing it, you were not knocking the Moon from its orbit, but it would be a sharp departure from ordinary usage to say that you omitted to do so or that the case is one of omission. Exactly what distinguishes the cases in which the terms are applicable is a matter of some controversy. There are two features of this case that might be thought to disqualify it: you aren’t able to knock the Moon from its orbit, and there is no reason why you should do so. Here I’ll stick to cases in which, as I see it, the terms (with their ordinary meanings) do apply. (Thanks to Michael Zimmerman for suggesting that I address this point.) For further discussion, see Clarke (2014: 28–​33). 7  If there is nothing that is an unwitting omission, what should we say about the claim, “An unwitting omission is a failure to exercise agency”? One thing we might say is that the proposition expressed by it does not quantify over omissions. The issue arises as well with talk of absences and lacks. Certain things are absent or lacking on certain occasions, but it is not clear that there is any entity that is an absence or a lack. For more on this point, see Clarke (2014: ch. 2). 8  The definitions allow that one can bear both direct and indirect responsibility for something, for one can have some responsibility for that thing that derives from one’s responsibility for some other thing

Responsibility for Acts and Omissions    95 It is generally thought that intentional actions, or at least some kinds of these, are among the things for which we can be directly responsible. A state of the world apart from one’s body that results from one’s intentional action, it is thought, is something for which one can be at most only indirectly responsible. What about omitting to act or refraining for action? Can an agent be directly responsible for either of these? Two kinds of conditions are generally thought to be necessary for direct responsibility for something. One is a control condition. In the case of intentional action, this condition is usually thought to require that the action is freely performed. Exactly what this freedom comes to, even in the case of intentional action, is a contested matter. (For example, there is debate about whether it requires an ability to do otherwise.9) A theory of responsibility for omitting or refraining ought to say whether a control condition for direct responsibility can be satisfied in these cases and, if so, whether it is the same as or different from the control condition required in cases of action. The second condition is an epistemic condition. This is usually thought to concern belief or awareness of what one is doing, and it is often said to concern as well knowledge or awareness of the moral status of one’s conduct. But whether it is actual belief or awareness that is required, or instead a capacity for one or the other of these things can fulfill the requirement, is a point of contention.10 Evidently, in cases of unwitting omission, in which the agent is unaware and lacks a belief that she is omitting to do the thing in question, a condition requiring actual belief or awareness would go unsatisfied.11 Indirect responsibility is likewise thought to require satisfaction of at least two conditions. Something for which one is indirectly responsible must, in the first place, result from or be a consequence of something else for which one is responsible. It is a disputed matter whether the consequence must be caused by the thing in question, or whether relations (or quasi-​relations) of other kinds can satisfy this portion of the requirement.12 Further contested is whether the agent must have been able, by acting

and some responsibility for it that does not derive from one’s responsibility for anything else. (Thanks to Michael Zimmerman for help in formulating this point.) 9 

Much of the debate stems from Frankfurt (1969). The discussion is extensive; for a collection of essays focusing on it, see Widerker and McKenna (2003). 10  For essays on the epistemic requirement, see Robichaud and Wieland (2017). 11  The two conditions do not seem to be entirely independent. Freely performing an action arguably requires having some awareness of what one is doing, and hence some epistemic requirement might be included in the control condition. But part of an epistemic requirement for responsibility, such as knowledge or awareness (or a capacity for one or another of these things) of the moral status of one’s conduct, might be thought to be additional. It is sometimes argued that moral responsibility requires, further, a capacity to respond to (specifically) moral reasons. Here I’ll set aside this matter by assuming that agents in all the cases discussed have any moral capacity required for responsibility. 12  For discussion, see Clarke (2014: 51–​54), Dowe (2001), Moore (2009: ch. 18), and Sartorio (2016: 46–​50). The mention here of “quasi-​relations” is prompted by the consideration that if on some occasion when an agent omits to do a certain thing there is nothing at all that is her omission, it will be wrong to speak of an outcome’s relation to the omission, since there is no relation without its relata. This issue is discussed further in section 6.

96   Randolph Clarke otherwise, to prevent the outcome in question.13 An epistemic requirement is also generally thought to apply: responsibility for a consequence can derive from responsibility for an antecedent only if, at the time of that antecedent, the agent believed or was aware, or was capable of believing or being aware, that the consequence in question would likely ensue. Several questions arise concerning responsibility for consequences of omissions since, as we have observed, omissions are sometimes unwitting and, further, it is a disputed matter whether omissions can be causes.14

3.  Responsibility for Intentionally Not Doing a Certain Thing When an agent’s refraining from A-​ing is simply identical with her performing a certain action, we might think that if the agent is directly responsible for performing the action, then she is directly responsible for refraining. An account of direct responsibility for action, it may seem, applies straightforwardly to direct responsibility for refraining in such a case. Suppose that it is not (as some say) only mental actions of certain kinds (decisions, volitions, etc.) for which one can be directly responsible, but that one can also be directly responsible for overt bodily actions, such as moving about in certain ways or holding still. Then the mime might be directly responsible for her action of holding still. Suppose that she is. Since her refraining from moving is her action of holding still, it might seem to follow that she is directly responsible for refraining from moving. The inference is invalid. For one thing, given that commonly a single action can be described in a variety of ways, the epistemic requirement appears to concern not actions but actions-​under-​descriptions. One might be aware of what one is doing under one description but unaware of it under another description. An attribution of responsibility employing the first description might then be true while one employing the second description is false. Responsibility ascriptions thus seem to be intensional: substitution of co-​referring expressions can fail to preserve truth-​value. Still, it is easy enough to see that the epistemic condition for the mime’s direct responsibility for not moving can be satisfied: when she is holding still, she is aware that she is refraining from moving, and we may suppose that she is aware as well of any moral significance her refraining from moving has. So far so good. There is a second worry about the inference considered three paragraphs back. Examples due to Harry Frankfurt (1969)—​Frankfurt cases, they are called—​have

13  Fischer and Ravizza (1998: ch. 4) require a kind of sensitivity of the outcome to one’s bodily conduct, but they do not require that one was able to prevent the outcome. 14  Dowe (2004) and Moore (2009: 444–​451) argue that omissions cannot be causes; Schaffer (2004, 2012) argues that they can be.

Responsibility for Acts and Omissions    97 convinced many that an agent can be directly responsible for an action of A-​ing even if she could not have avoided A-​ing. But some theorists convinced of this view nevertheless maintain that one can be directly responsible for not A-​ing only if one was able to A. If both of these claims are correct, then the control condition incorporates an asymmetry: direct responsibility for an action does not require an ability not to perform that action, whereas direct responsibility for not performing a certain action requires an ability to perform that action.15 Suppose that although the mime acted freely in holding still, some entity was poised to intervene and ensure that she not move had she not freely held still. She was unable to move. If there exists the asymmetry in question, then the mime might be directly responsible for holding still but not directly responsible—​indeed, not responsible—​for not moving. Does this possibility show that the proposed identity of the mime’s refraining with her holding still is mistaken? Perhaps not; perhaps, instead, we find here another way in which the intensionality of responsibility ascriptions shows up.16 In any event, it seems plain that even if there is the purported asymmetry in control conditions, in cases of (what appear to be) simple identity, an agent can be directly responsible for refraining from doing a certain thing. (Frankfurt cases are, after all, quite rare!) Cases of displacement refraining, too, seem to allow for direct responsibility for refraining. Suppose that Ann had shoved her hands into her pockets in order to—​and thereby did—​prevent herself from reaching for a chocolate. If her refraining from reaching is her act of shoving her hands into her pockets, then she might satisfy not only the control but also the epistemic condition for direct responsibility for so refraining. For she might freely shove her hands into her pockets while remaining able (and having the opportunity) to reach; and she might be aware that by shoving her hands into her pockets she is thereby refraining from reaching, and aware as well of any moral significance of her refraining. What of the version of Ann’s case in which she performs no act as a means to refraining? She might be directly responsible for deciding not to reach, but that decision is not her refraining, and in any event she might refrain from reaching for a chocolate

15  Fischer (1985–​1986) and Fischer and Ravizza (1991) argued for the asymmetry. Van Inwagen (1983: 161–180) advances the second half of it and judges the first half probably correct. Fischer and Ravizza (1998) subsequently rejected the asymmetry, as have Byrd (2007), Clarke (1994, 2011), Frankfurt (1994), and McIntyre (1994). Fischer (2017) and Sartorio (2005) each argue for a different asymmetry (from the one characterized in the text here) between responsibility for actions and responsibility for omissions (as does Clarke [1994, 2011]). Note, further, that other writers (e.g., Ginet [2003], Widerker [1995]) reject the asymmetry because they reject the first half. 16  Fischer (2017: 156) argues that in cases in which an omission is identical with an action, responsibility for the omission does not require an ability to do otherwise, because responsibility for an action does not require such an ability. The argument implicitly rejects the intensionality of responsibility ascriptions that is considered in the text here. Perhaps the rejection is correct, but the issue merits attention.

98   Randolph Clarke without making a decision not to reach. (She might intentionally not reach without having made a decision not to reach.) Whether she makes the decision or not, Ann’s refraining is an exercise of a capacity to do things intentionally. In this respect it is like performing an intentional action. Further, refraining in such a case is “basic” for the agent in a way that many intentional actions are: it is something that the agent can do without having to do anything else as a means to doing it. Thus, there is reason to think that just as an agent can be directly responsible for a (basic) intentional action, Ann can be directly responsible for refraining from reaching for a chocolate, even when her so refraining isn’t an action. It is easy enough to see how refraining in a case of this last kind might satisfy the epistemic condition for direct responsibility. But how might it satisfy the control condition? It might be thought that this condition is satisfied just in case the agent freely refrains, but what does freely refraining come to, when refraining isn’t performing an action? Consider: An agent freely refrains from A-​ing just in case she intentionally doesn’t A while remaining free to A.

Even if this is correct as far as it goes,17 it raises two further questions: what does intentionally not doing a certain thing come to, particularly when this isn’t identical with performing any action; and what is it to remain free to A, when one intentionally doesn’t A? There are proposals concerning the first of these questions, but there is no consensus on the matter.18 This is one point at which further work on agency is needed if we are to have a comprehensive theory of responsibility. Perhaps the second question might be addressed as follows. When an agent intentionally doesn’t A, she is on that occasion free to A if she is then able to and has the opportunity to A.19 Having both the ability and the opportunity to A on that occasion, the agent can—​in J. L. Austin’s (1979: 229) all-​in sense—​A then. If we have sufficient conditions for direct responsibility for an omission in this kind of case, one in which the omission is not (identical with) an action, then it is not the case that moral responsibility is always grounded in responsibility for action.20 And yet, we have seen, we might get this result even though the control condition for direct responsibility concerns freedom of action. For in a case of omission, that condition might be satisfied not by acting freely but by remaining free to act. 17  Consideration of Frankfurt cases might lead some to think that the proposal overstates what is necessary for freely refraining. It might then be restated as a conditional (rather than a biconditional) offering only a sufficient condition. (Thanks to Michael Zimmerman for raising this issue.) 18  For some examples, see Clarke (2014: ch. 3), Sartorio (2009), and Shepherd (2014). 19  Having the ability mentioned here should be understood to require (among other things) having what it takes mentally to do the thing in question. An agent with ordinary motor control might lack this kind of ability to touch a spider if she has a pathological fear of spiders. (Michael Zimmerman prompted this clarification.) 20  Contrary to what is claimed by Fischer (2017: 159).

Responsibility for Acts and Omissions    99 So far we have considered episodic cases of intentionally not doing a certain thing. What about periodic cases? Bob, we imagined, has abstained for 30 years from eating meat. During some of that time he has been asleep, and when he is asleep, he isn’t able then to eat meat. We might still say that he intentionally hasn’t eaten meat for 30 years, and the periodic inability doesn’t seem to undermine his responsibility for abstaining. Even periods of unexpected incapacitation need not undermine this; if he was accidentally knocked unconscious for some period during the 30 years, it may still be that he has abstained for 30 years. Might he be responsible for abstaining for 30 years but not responsible for abstaining during such a period of unconsciousness? Cases of this sort are underexplored. It is far from obvious what account is to be given of agents’ responsibility in them.

4.  Responsibility for Unwitting Omission: Direct (or Nearly So) Surprisingly, cases of unwitting omission have (of late) received quite a lot of attention, perhaps singled out because they are, prima facie, the least like cases of action, and hence are thought to present the toughest challenge for a theory of responsibility meant to cover omitting and refraining as well as action. They are indeed challenging. Consider, first, an episodic case, such as Carla’s. Despite being committed to turning on the pool pump on evenings when an overnight freeze is forecast, one evening, although she has heard that a freeze is forecast, she forgets to turn on the pump and omits to do so. Writers are divided on whether an agent can be directly responsible for her omission in a case of this kind. One kind of view on which an agent can be responsible for such an omission, without her responsibility for it having to derive from responsibility for any action or other omission, includes commitment to a general theory of responsibility known variously as Attributionism or Responsibility as Answerability.21 (I’ll use the latter name.) On this theory, direct responsibility does not require voluntariness; agents can be, and commonly are, responsible for non-​voluntary attitudes such as desires, emotional states, cares, and concerns, without their responsibility for these attitudes having to derive from responsibility for any actions or omissions to act. Proponents of this theory need not eschew a control condition altogether; as they often observe, the attitudes in question are “judgment-​sensitive,” generally responsive to our evaluative judgments or 21  Proponents of Responsibility as Answerability include Hieronymi (2014), Scanlon (1998: ch. 6), A. Smith (2005, 2015), H. Smith (2011), and Talbert (2017). A. Smith (2017) and Talbert (2017) apply the view to cases of unwitting omissions, and H. Smith (2011) applies it to cases of wrongdoing done from ignorance, in some of which the ignorance is due to unwitting omission. All three of these authors endorse the requirement, for blameworthiness, that the omission in question reflect morally objectionable attitudes (see the discussion in the text below).

100   Randolph Clarke judgments about what reasons we have.22 Coming to have such an attitude, then, is commonly an exercise of reason, even though not an intentional action. The attitudes and the actions that reflect them are then said to be attributable to agents as a basis for moral responses to these agents, who are said to be answerable for these attitudes and actions, in that it makes sense to ask their reasons for them. Unwitting omissions, too, can stem from and reflect the attitudes in question, and, it is said, one can be similarly answerable for them. Thus, on this theory, one can be as directly responsible for an unwitting omission as one can be for an action.23 The requirement that an unwitting omission reflect one’s attitudes, if one is to be in this way responsible for it, might be thought to be (at least part of) a control condition. But what does it come to? In the case of action, reflecting the agent’s attitudes is in part a causal matter: the attitudes in question are causes of the action, or of changes partly constituting the action.24 But are unwitting omissions caused? Recall that it is not clear that there is any entity at all that is one’s omission, when one unwittingly omits to do a certain thing. If there isn’t, then there is nothing to serve as a causal relatum here.25 It might be proposed that one’s attitudes can explain an unwitting omission even if they can’t cause it. Perhaps Carla’s interest in the Grateful Dead documentary she was watching on Netflix explains her forgetting, and thus her omitting to turn on the pool pump. An informative account would have to spell out what explanation comes to here (if it isn’t causal explanation). In the case of action, attitudes that play a crucial causal role commonly rationalize the action, and the action’s reflection of them might be thought to be not just a causal matter but an expressive one: the action is not mere evidence of those attitudes, it is an expression of them. But attitudes that explain an unwitting omission need not similarly rationalize that omission. Carla’s omission to turn on the pump, even if explained by her interest in the documentary, isn’t in this way an expression of that interest. Her interest in the documentary is not, and is not thought by her to be, a reason not to turn on the pump. (Neither is the documentary, or the fact that it was available on Netflix.) For the theory of responsibility in question here, then, there is work to be done in spelling out what the reflection of attitudes comes to in cases of unwitting omission. 22  “Judgment-​sensitive” is Scanlon’s (1998: 20) designation of the attitudes in question. As he puts it, “these attitudes are ‘up to us’—​that is, they depend on our judgment as to whether appropriate reasons are present” (22). 23  See, e.g., A. Smith (2017) and Talbert (2017). It is an interesting question for what, on this theory, agents should be thought to be directly responsible. It might be thought to be the attitudes reflected by actions or omissions. If so, that would seem to make our responsibility for actions or omissions not quite direct (hence “Or Nearly So” in the title of this section). Scanlon’s claim that these attitudes are “up to us” (quoted in the preceding note) suggests that we are directly responsible for judgments that partly constitute or rationalize these attitudes. 24  I remain neutral here between two ways of characterizing action: as caused in a certain way by certain things, or as a causing (of certain things, by certain things, in a certain way). My formulation is nevertheless committed to the idea that some of an agent’s attitudes play a key causal role when she acts. 25  It might still be said that we can have causation, albeit causation that is not relational. Lewis (2004) proposes this kind of view of causation involving absences.

Responsibility for Acts and Omissions    101 Suppose that unbeknown to Carla, the pool pump was broken; even if Carla had flipped its switch, the pump would not have operated. She wasn’t able to turn it on. On the assumption that she is not responsible for the pump’s being broken, she is not then responsible for its not operating that night; and it seems just as clear that she is not responsible, either, for not turning it on (even if she is responsible for not trying to turn it on). If this is correct, then the control condition for direct responsibility in a case of this kind might require something—​an ability to do otherwise—​that (if Frankfurt is right) is not required for direct responsibility for action.26 What about the epistemic condition? Carla isn’t aware that she is omitting to turn on the pump; she is thus unaware of an aspect of the moral significance of her conduct. Still, if having a capacity for such awareness can suffice for responsibility, Carla might nevertheless satisfy the requirement. Here we will want an account to say more about what such a capacity is and what having it comes to. Some proponents of Responsibility as Answerability reject the epistemic condition altogether.27 But suppose that Carla had been rendered incapable on this occasion of thinking to turn on the pump. If she is not responsible for the incapacity, it is doubtful that she is responsible for her omission, even if it does reflect her interest in the documentary. It seems unreasonable to expect someone to do something if she isn’t even capable of thinking to do it. Theorists who embrace Responsibility as Answerability commonly hold that an agent is blameworthy for such an omission only if it reflects attitudes that are morally objectionable (or some objectionable complex of attitudes). Theirs is a quality of will view, and on it one is blameworthy for conduct only if it reflects ill will or a lack of good will. If Carla’s omission reflects no objectionable will, she is not to blame for not turning on the pump. Consider cases in which a parent forgets to drop off her infant at day care, forgets the infant in the car when she parks at work, and discovers the infant in the car at the end of the work day, dead from heatstroke. The parents who do this are typically decent people, loving parents and, except on this occasion, attentive to their children. Afterward they are devastated, usually blaming themselves for the child’s death and commonly suffering severe depression. Perhaps mercy, as well as the recognition that we, too, could make such a mistake, precludes our blaming them at all. But it is not clear that they are blameless; they certainly don’t think they are.28 “I just forgot” explains that their misdeed wasn’t of another kind, but it isn’t an excuse.

26  An alternative diagnosis of Carla’s lack of responsibility for not turning on the pump in this case might note that (with the pump broken) Carla’s not trying to turn it on doesn’t cause (or quasi-​cause) its not coming on. (Carolina Sartorio mentioned this alternative in correspondence.) On quasi-​causation, see section 6. 27  See, e.g., H. Smith (2017: 99–​102). 28  Pace Talbert’s suggestion (2017: 24, n. 10) regarding a similar case, agent-​regret is not the attitude that is fitting for these parents to have. That attitude might be fitting for a parent whose infant fell victim to SIDS after the parent put her down for a nap. The latter parent’s conduct (we may imagine) was not in any way faulty, and she could not reasonably have been expected to act so as to prevent the death. In

102   Randolph Clarke A different kind of view that allows for direct responsibility for unwitting omissions embraces a version of the control condition on which it concerns freedom of the kind exercised in intentional action. Of course, an unwitting omission isn’t an exercise of this kind of freedom, but the agent might nevertheless be free to perform the action that she omits, where being free to do this requires having the ability and opportunity to do that thing intentionally. Accounts of this kind focus on agents’ capacities, abilities, and opportunities to explain how agents might satisfy both control and epistemic requirements for direct responsibility in such cases. Our agency relies on our having capacities to notice the need to act, think to act when needed, and remember to act as we intend. Commonly we do these things without taking any action to ensure that we do; commonly no such action is needed. But we can also take action to ensure such things, for example, asking ourselves whether there are conditions we might have failed to notice, or whether there is something we are forgetting to do. Unusual circumstances can excuse failures to act that stem from failures to think to do, remember to do, or see the need to do certain things. We might excuse Carla if she had received a call that evening informing her that her father had just passed away. But absent such circumstances, we expect agents to see the need to do things that they ought to do, when they have the capacity to do so. The expectation is normative: absent excusing conditions, agents with obligations to do certain things and capacities to think to do these things should think to do them. On this view, if such agents fail to do so, they can be at fault for omitting the actions in question.29 Carla, we may suppose, is capable of remembering on this occasion to turn on the pump. And we may suppose that she is able and has the opportunity to turn it on—​there is nothing wrong with her motor control, she has access to the switch, and the pump is in good working order. She is free to turn it on, and capable of remembering to do so and of realizing that, in continuing to watch the documentary, she is wrongly omitting to do so. Absent an excusing condition like the disturbing call, she may be to blame for

contrast, the conduct of the parent who leaves a child in a hot car is plainly faulty, and she certainly can reasonably have been expected to act so as to prevent the death. 29 

I advance a view of this kind in Clarke (2014: ch. 7, 2017a, 2017b). Sher (2009, 2017) advances a similar account, one to which my own is much indebted. Two differences: first, Sher does not require that the agent be able to do the thing that she omits to do. (In fact, he holds [2017: 12] that on the occasion of an unwitting omission, the agent is incapable of recognizing and responding to the reasons to perform the action in question. But a failure to recognize and respond does not entail the incapacity; further, it seems to me that an agent subject to such an incapacity would not be directly responsible for the omission.) Second, he requires that the agent’s failure to recognize the wrongness of her conduct be caused by “the interaction of some combination of his constitutive attitudes, dispositions, and traits” (2009: 88). (As noted in the text earlier, the requirement might be stated in terms of explanation rather than causation.) If the causal (or explanatory) requirement is needed, it can easily be added to the view as stated in the text here. Accounts of responsibility for negligent conduct, which commonly stems from omissions to advert, that similarly appeal to agents’ capacities are advanced by Hart (1968), Sverdlik (1993), and Raz (2011: ch. 13).

Responsibility for Acts and Omissions    103 her omission, and for the damaged pipes should they burst. Her responsibility for the omission, on this view, can be direct, not having to derive from responsibility for any other thing. Do agents in cases of unwitting omission generally have these capacities and abilities? Common sense says they do, and we may reasonably take common sense to be innocent until proven guilty here. What is the evidence for the prosecution? First, it might be objected that although Carla has a general capacity to remember to act as she intends, she isn’t capable of so remembering on this occasion.30 For she is watching TV and, it might be said, she is not capable of thinking of turning on the pump while so engaged. It would not be at all unrealistic to imagine that Carla routinely remembers to fulfill her obligations, including safeguarding the pool pipes, even when she watches TV in the evening, and even when she watches things as interesting as the Grateful Dead documentary. Barring general skepticism about unmanifested capacities, there might be no reason to think that Carla lacks the capacity in question on this occasion.31 Second, it might be objected that we can reasonably expect people to do things only if they can do them rationally, where what an agent can do rationally is understood to be a function of what the agent takes her reasons to be. Carla, it might be said, can’t reasonably be expected to turn on the pump on the evening in question, because she could only do so irrationally.32 But Carla knows that she has reason to turn on the pump, for she knows that a freeze is forecast. If prompted, she can easily identify the reason. Further, common sense says that she is capable of being aware on this occasion of such a reason. Capable of being so aware, she is able to act on a reason of which she is aware in turning on the pump on this occasion. She is thus able to perform that action rationally, in the indicated sense. Third, it might be objected that direct responsibility for an omission requires that the agent have an opportunity, at the time in question, to perform the omitted action, and that in a case of unwitting omission the agent lacks that opportunity. Having an opportunity to do something that one omits to do, it might be said, requires actually being aware that one is not doing that thing.33 If having promised to call my daughter I omit to do so simply because I forgot, I shouldn’t tell her that I didn’t have the opportunity to call. Lacking the opportunity would be lacking access to a phone, or being tied up with more important matters, or some similar circumstance. The objection might be taken as stating that actual awareness that one is omitting to do a certain thing is a requirement for its being the case that one all-​in can do that thing. But if one has the capacity to think to do it on this

30 

Nelkin and Rickless (2017: 127) raise this worry about a similar case. There might be something in the circumstances that caused Carla not to remember to turn on the pump (if not-​rememberings can be caused). However, “caused not to do” does not imply “rendered incapable of doing.” More would need to be said to support the denial of the capacity on this occasion. 32  Levy (2011: 127) advances an objection of this kind regarding cases of this sort. 33  Nelkin and Rickless (2017: 119) raise this objection to the account at issue here. 31 

104   Randolph Clarke occasion, and one is able to and has the opportunity to do it, then it appears that one all-​ in can do it. Note, further, that if it isn’t true that Carla all-​in can turn on the pump on the evening in question, then there might be a surprising implication concerning her obligation: if “ought” implies “can,” then, if she cannot, it isn’t true that she ought. We could be relieved of obligations, then, simply by forgetting to do what we had been supposed to do! That hardly seems to be so. Finally, it might be objected that an agent such as Carla cannot be blameworthy for her omission unless it reflects objectionable attitudes of hers. If it does not, then we should not take offense or be indignant toward her for it; and if no such reactive attitude is appropriate, then (the objection goes) she is not blameworthy.34 Of course, an agent can be responsible for something that is morally deficient without being blameworthy for it. But there may be a case for blameworthiness even when passionate reactive attitudes are inappropriate. Even if Carla, rather than forgetting, had decided not to bother with the pump on the evening in question, resentment or indignation might be overblown. Still, we might think she is blameworthy for risking burst pipes, and she might later blame herself. If she is to blame, it is for failing to turn on the pump when she should have, not for having bad attitudes. Periodic cases of unwitting omission do not appear to present any additional challenges for an account of responsibility. When David omits to pull weeds in the garden in June, there is no particular interval during June at which we must locate his omission. Any period of weed pulling an hour or more in length—​even a fusion of shorter periods summing to an hour or more—​would have fulfilled his commitment. Still, it seems that if any account can provide for direct responsibility for Carla’s omission, it can provide the same for David’s.

5.  Responsibility for Unwitting Omissions: Indirect Several writers maintain that responsibility for an unwitting omission can only be indirect, deriving from responsibility for some action or other omission. Of course, defenders of the possibility of direct responsibility can allow that sometimes responsibility for unwitting omissions is indirect; the issue is whether it can only be such. Defense of an affirmative answer will need, first, to show that no account of direct responsibility for such omissions is adequate. A further task is to provide an account of indirect responsibility. Indirect responsibility for an unwitting omission might be said to derive, in some cases at least, from responsibility for a prior decision not to take precautions. Carla 34 

Talbert (2017) raises this objection.

Responsibility for Acts and Omissions    105 might have decided not to create a reminder on her phone when she learned that a freeze was forecast, or when she began to watch the documentary. She might be responsible for making that decision, and, it might be proposed, her responsibility for her unwitting omission to turn on the pump, on the occasion in question, derives from her responsibility for making the decision. The unwitting omission might be said to result from the earlier decision. As has been noted, a question arises here about whether this can be a matter of causation. It is not clear that in a case of this sort there is any entity at all that is the agent’s omission, and if there is not, then no causal relatum is the omission. Spelling out what “resulting from” comes to here is part of identifying a control condition that can be satisfied for indirect responsibility in such a case. One question that must be addressed regarding the control condition is whether, for indirect responsibility, the agent must have had the ability and opportunity to do what she omitted to do. It is usually thought that derivative responsibility requires satisfaction of an epistemic requirement. In the case in question, it might be proposed, at the time of the earlier decision, the agent must have been aware that in so deciding she risked the later omission; or, alternatively, it might be said, she must at least have been capable of such awareness at the time of the decision. A requirement of actual awareness will commonly be unsatisfied; if Carla made the decision we are imagining, she might have judged when she did that no reminder was needed. Further, if the epistemic requirement can be satisfied here by possession of a capacity for awareness of risk, it is curious why a capacity for awareness can’t satisfy the epistemic requirement for direct responsibility for the subsequent unwitting omission.35 A second challenge for this view is that such a prior decision might be something for which the agent is in no way blameworthy. Carla might know that she has never before forgotten to turn on the pump when she knew that a freeze was forecast. She might have reasonably judged that failing to create a reminder created no significant risk. Still, she forgot on the occasion in question. Despite her earlier blameless conduct, we (and she) might think she is to blame for the omission, and for the damaged pipes, should they burst. Third, of course, Carla might not have considered taking any such measures to ensure that she would fulfill her obligation regarding the pool pump. And if in fact she has never before failed in her obligation, it might have been entirely sensible of her not to consider such measures. There is then no earlier decision from her responsibility for which her responsibility for the omission might be said to derive.36 The strategy under consideration might be supplemented with the idea that responsibility for unwitting omissions derives, in some cases, from responsibility for prior

35  Miller (2017) argues for the related point that there is an inconsistency in holding both that blameless ignorance excuses and that tracing requires foreseeability but not actual foresight. 36  Several writers note such limitations on the strategy of deriving responsibility for an unwitting omission from responsibility for a prior decision. See, e.g., Clarke (2017a) and Nelkin and Rickless (2017).

106   Randolph Clarke witting omissions. For example, it might be said, Carla need not have decided not to create a reminder. She need only have omitted earlier to create the reminder while aware then that she was so omitting, and while believing that in so omitting she was omitting to do something that would significantly reduce the likelihood of a subsequent omission to turn on the pump.37 The epistemic requirement for direct responsibility here is actual awareness, and for the derivation of responsibility it is actual belief in the likelihood of the outcome. As before, an account will be needed of how unwitting omissions result from other omissions. But there is a further problem if, as some proponents of this view have it, lack of awareness that one is omitting to A deprives one of the opportunity to A. If while watching TV Carla lacks the opportunity to turn on the pump, and if she isn’t to blame for lacking that opportunity, she hardly seems to be to blame for not turning on the pump. It might be suggested that she can be indirectly to blame for lacking the opportunity, since (it may be said) the lack results from her earlier omission to create a reminder. But must we then require that at the time of the earlier omission, she believed that it would result in her later lack of opportunity? It is doubtful that agents commonly have such beliefs; even if I believe that something I am doing now is likely to result in my not doing something else later, I would not ordinarily believe that what I am doing now will likely deprive me of the opportunity to do that other thing later. A further strategy for locating the basis of derivative responsibility for unwitting omissions points to prior actions that contributed to character traits that, in turn, resulted in these omissions. For example, if Carla cares little about fulfilling her duties, she might be said to be responsible for her omission on this occasion (in part) because she is responsible for earlier character-​forming actions that resulted in her lack of concern. Of course, even agents without such bad traits sometimes unwittingly omit to do things they should do. Moreover, even when agents with such bad traits have contributed, via prior actions, to their bad character, they might well have lacked awareness at the times of the earlier actions that the subsequent omissions would result.38 There are many cases of unwitting omission whose agents we commonly take to be responsible but in which, if direct responsibility is ruled out, plausible requirements for responsibility are not satisfied. If responsibility for an unwitting omission can only be indirect, then agents are responsible in such cases far less often than we think. A view

37 

Nelkin and Rickless (2017) advance an account of responsibility for unwitting omissions that appeals to responsibility for an earlier witting omission of this sort. As they see it, awareness that one is omitting to do a certain thing is a requirement for having the opportunity to do that thing. Thus, when an agent unwittingly omits to do something, she lacks the opportunity to do the thing then. I raised objections to this claim in section 4. In the text to follow here, I discuss a problem for their view of indirect responsibility for an unwitting omission. 38  Again, if a capacity for awareness, without actual awareness, can suffice (as far as the epistemic requirement is concerned) for this derivation of responsibility for the subsequent omissions, it is curious why a capacity for awareness can’t satisfy the epistemic requirement for direct responsibility for unwitting omissions.

Responsibility for Acts and Omissions    107 that denies the possibility of direct responsibility for unwitting omissions is in this respect highly revisionist.

6.  Responsibility for Consequences of Omissions We commonly think that agents are responsible for (some) consequences of omissions—​ whether intentional or unwitting—​for which they are responsible. If the mime’s not moving for quite some time delights a group of schoolchildren, then the mime might be responsible not just for refraining from moving but also for the resulting delight. If the parent is responsible for omitting to remove the child when she parks her car for the day, she might be responsible, as well, for the death of the child. Results of omissions can be as foreseeable and foreseen as can results of actions, so it is no mystery how an epistemic condition for responsibility for the former can be satisfied. What about a control condition? We’ve seen reason to wonder whether omissions are causes, particularly in cases in which there appears to be no action by the agent that is her omission. Even if they aren’t, they might be “quasi-​causes” of outcomes. Quasi-​causation is characterized in terms of causation and counterfactuals of causation. For example, Carla’s omission to turn on the pool pump quasi-​causes the bursting of the pipes if i) Carla omits to turn on the pump; ii) the pipes burst; and iii) there is some condition x such that a) x causes the bursting of the pipes; and b) had Carla turned on the pump, her doing so would have prevented the bursting of the pipes by causally interacting with a process due to x.39 The freezing temperature might be the condition in question. Prevention, in turn, can be understood as quasi-​causation of the non-​occurrence of an event of some type, in this case, a bursting of the pipes.40 Might simple counterfactual dependence of an outcome on an omission be the link that is required, and suffices as far as linking is concerned, for responsibility for that 39 

40 



Modeled on Dowe (2001: 222). Carla’s turning on the pump prevents the bursting of the pipes if

i) Carla turns on the pump; ii) the pipes don’t burst; and iii) there is some condition x such that a) Carla’s turning on the pump causally interacts with a process due to x; and b) had Carla not turned on the pump, x would have caused the bursting of the pipes. (Modeled on Dowe [2001: 221].)

108   Randolph Clarke outcome?41 Such counterfactual dependence is not a necessary condition. Imagine that Carla had neglected a day earlier to purchase a part needed to repair the pool pump. Having forgotten today her earlier neglect, she sees the weather report and plans to turn the pump on this evening. But she forgets to do so, and she omits to hit the switch. Had she purchased the part and repaired the pump, we may suppose, still the pipes would have burst, for she would have forgotten to turn on the pump. And had she hit the switch, still the pipes would have burst, for the pump would not have operated, not having been repaired. Nevertheless, if Carla can be responsible for the damaged pipes in the initial version of the case, she can be responsible for that outcome in this version. Note that, in this case, Carla is responsible for the fact that, on the evening in question, the bursting of the pipes does not counterfactually depend on her omission to hit the switch. Perhaps, then, a linkage required for responsibility for an outcome is that either the outcome counterfactually depends on one’s omission or, if it does not, one is responsible for the fact that it does not.42 If we can have only quasi-​causation or counterfactual dependence, and not causation, linking omissions with outcomes, this fact does not itself lessen our responsibility for results of omissions.43 One can offend by saying a certain thing that one should not say, or by not saying a certain thing that one should. In the first kind of case, one offends by acting, in the second, by omitting to act. In some pairs of cases, the duties that are violated might be equally stringent, and the harm that results equally bad and equally maliciously intended. It can be equally foreseeable and equally preventable. The agents can be equally blameworthy for it.44

References Austin, J. L. 1979. “Ifs and Cans.” In Philosophical Papers, 3rd edition, eds. J. O. Urmson and G. J. Warnock, 205–​232. Oxford: Clarendon Press. Bach, Kent. 2010. “Refraining, Omitting, and Negative Acts.” In A Companion to the Philosophy of Action, eds. Timothy O’Connor and Constantine Sandis, 50–​57. Oxford: Blackwell. Byrd, Jeremy. 2007. “Moral Responsibility and Omissions.” Philosophical Quarterly 57: 56–​67. Clarke, Randolph. 1994. “Ability and Responsibility for Omissions.” Philosophical Studies 73: 195–​208. Clarke, Randolph. 2011. “Omissions, Responsibility, and Symmetry.” Philosophy and Phenomenological Research 82: 594–​624. Clarke, Randolph. 2014. Omissions: Agency, Metaphysics, and Responsibility. New York: Oxford University Press. 41 

Moore, who argues (2009: 444–​451) that omissions cannot be causes, maintains (451–​452) that counterfactual dependence is necessary for responsibility for outcomes that ensue from omissions. 42  Sartorio (2004) advances a view on which even the weakened condition suggested here is not required. 43  Moore (2009: 448) argues that the commonly lesser responsibility for omissions (than for actions with equally bad outcomes) is explained by the causal difference that actions cause their outcomes whereas omissions do not. 44  Many thanks to John Martin Fischer, Dana Nelkin, Sam Rickless, Carolina Sartorio, and Michael Zimmerman for comments on an earlier draft of this chapter.

Responsibility for Acts and Omissions    109 Clarke, Randolph. 2017a. “Blameworthiness and Unwitting Omissions.” In The Ethics and Law of Omissions, eds. Dana Kay Nelkin and Samuel C. Rickless, 63–​83. New York: Oxford University Press. Clarke, Randolph. 2017b. “Ignorance, Revision, and Commonsense.” In Responsibility: The Epistemic Condition, eds. Philip Robichaud and Jan Willem Wieland, 233–​251. Oxford: Oxford University Press. Davidson, Donald. 1985. “Replies to Essays I-​IX.” In Essays on Davidson: Actions and Events, eds. Bruce Vermazen and Merrill B. Hintikka, 195–​229. Oxford: Clarendon Press. Dowe, Phil. 2001. “A Counterfactual Theory of Prevention and ‘Causation’ by Omission.” Australasian Journal of Philosophy 79: 216–​226. Dowe, Phil. 2004. “Causes Are Physically Connected to their Effects: Why Preventers and Omissions Are Not Causes.” In Contemporary Debates in Philosophy of Science, ed. Christopher Hitchcock, 189–​196. Malden, MA: Blackwell. Fischer, John Martin. 1985–​86. “Responsibility and Failure.” Proceedings of the Aristotelian Society 86: 251–​270. Fischer, John Martin. 2017. “Responsibility and Omissions.” In The Ethics and Law of Omissions, eds. Dana Kay Nelkin and Samuel C. Rickless, 148–​162. New York: Oxford University Press. Fischer, John Martin and Mark Ravizza. 1991. “Responsibility and Inevitability.” Ethics 101: 258–​278. Fisher, John Martin and Mark Ravizza. 1998. Responsibility and Control: A Theory of Moral Responsibility. Cambridge: Cambridge University Press. Frankfurt, Harry G. 1969. “Alternate Possibilities and Moral Responsibility.” Journal of Philosophy 66: 829–​839. Frankfurt, Harry G. 1994. “An Alleged Asymmetry Between Actions and Omissions.” Ethics 104: 620–​623. Ginet, Carl. 2003. “In Defense of the Principle of Alternative Possibilities: Why I Don’t Find Frankfurt’s Argument Convincing.” In Moral Responsibility and Alternative Possibilities: Essays on the Importance of Alternative Possibilities, eds. David Widerker and Michael McKenna, 75–​90. Aldershot: Ashgate. Hart, H. L. A. 1968. “Negligence, Mens Rea, and Criminal Responsibility.” In Punishment and Responsibility: Essays in the Philosophy of Law, 136–​157. Oxford: Clarendon Press. Hieronymi, Pamela. 2014. “Reflection and Responsibility.” Philosophy and Public Affairs 42: 3–​41. Levy, Neil. 2011. Hard Luck: How Luck Undermines Free Will and Moral Responsibility. Oxford: Oxford University Press. Lewis, David. 2004. “Void and Object.” In Causation and Counterfactuals, eds. John Collins, Ned Hall, and L. A. Paul, 277–​290. Cambridge, MA: MIT Press. McIntyre, Alison G. 1994. “Compatibilists Could Have Done Otherwise: Responsibility and Negative Agency.” Philosophical Review 103: 453–​488. Miller, Daniel J. 2017. “Reasonable Foreseeability and Blameless Ignorance.” Philosophical Studies 174: 1561–​1581. Moore, Michael S. 2009. Causation and Responsibility: An Essay in Law, Morals, and Metaphysics. Oxford: Oxford University Press. Moore, Robert E. 1979. “Refraining.” Philosophical Studies 36: 407–​424. Nelkin, Dana Kay and Samuel C. Rickless. 2017. “Moral Responsibility for Unwitting Omissions: A New Tracing View.” In The Ethics and Law of Omissions, eds. Dana Kay Nelkin and Samuel C. Rickless, 106–​130. New York: Oxford University Press. Raz, Joseph. 2011. From Normativity to Responsibility. Oxford: Oxford University Press.

110   Randolph Clarke Robichaud, Philip and Jan Willem Wieland, eds. 2017. Responsibility: The Epistemic Condition. Oxford: Oxford University Press. Sartorio, Carolina. 2004. “How to be Responsible for Something Without Causing It.” Philosophical Perspectives 18: 315–​336. Sartorio, Carolina. 2005. “A New Asymmetry Between Actions and Omissions.” Noûs 39: 460–​482. Sartorio, Carolina. 2009. “Omissions and Causalism.” Noûs 43: 513–​530. Sartorio, Carolina. 2016. Causation and Free Will. Oxford: Oxford University Press. Scanlon, T. M. 1998. What We Owe to Each Other. Cambridge, MA: Harvard University Press. Schaffer, Jonathan. 2004. “Causes Need Not be Physically Connected to Their Effects: The Case for Negative Causation.” In Contemporary Debates in Philosophy of Science, ed. Christopher Hitchcock, 197–​216. Malden, MA: Blackwell. Schaffer, Jonathan. 2012. “Disconnection and Responsibility.” Legal Theory 18: 399–​435. Shepherd, Joshua. 2014. “Causalism and Intentional Omissions.” American Philosophical Quarterly 51: 15–​26. Sher, George. 2009. Who Knew? Responsibility Without Awareness. New York: Oxford University Press. Sher, George. 2017. “Unintentional Omissions.” In The Ethics and Law of Omissions, eds. Dana Kay Nelkin and Samuel C. Rickless, 3–​16. New York: Oxford University Press. Smith, Angela M. 2005. “Responsibility for Attitudes: Activity and Passivity in Mental Life.” Ethics 115: 236–​271. Smith, Angela M. 2015. “Responsibility as Answerability.” Inquiry 58: 99–​126. Smith, Angela M. 2017. “Unconscious Omissions, Reasonable Expectations, and Responsibility.” In The Ethics and Law of Omissions, eds. Dana Kay Nelkin and Samuel C. Rickless, 36–​60. New York: Oxford University Press. Smith, Holly M. 2011. “Non-​ Tracing Cases of Culpable Ignorance.” Criminal Law and Philosophy 5: 115–​146. Smith, Holly M. 2017. “Tracing Cases of Culpable Ignorance.” In Perspectives on Ignorance from Moral and Social Philosophy, ed. Rik Peels, 95–​119. New York: Routledge. Smith, Patricia G. 1986. “Ethics and Action Theory on Refraining: A Familiar Refrain in Two Parts.” Journal of Value Inquiry 20: 3–​17. Sverdlik, Steven. 1993. “Pure Negligence.” American Philosophical Quarterly 30: 137–​149. Talbert, Matthew. 2017. “Omission and Attribution Error.” In The Ethics and Law of Omissions, eds. Dana Kay Nelkin and Samuel C. Rickless, 17–​35. New York: Oxford University Press. van Inwagen, Peter. 1983. An Essay on Free Will. Oxford: Clarendon Press. Varzi, Achille C. 2006. “The Talk I Was Supposed to Give . . .” In Modes of Existence: Papers in Ontology and Philosophical Logic, eds. Andrea Bottani and Richard Davies, 131–​151. Frankfurt: Ontos Verlag. Varzi, Achille C. 2008. “Failures, Omissions, and Negative Descriptions.” In Meaning, Intentions, and Argumentation, eds. Kepa Korta and Joana Garmendia, 61–​75. Stanford, CA: CSLI Publications. Vermazen, Bruce. 1985. “Negative Acts.” In Essays on Davidson: Actions and Events, eds. Bruce Vermazen and Merrill B. Hintikka, 93–​104. Oxford: Clarendon Press. Widerker, David. 1995. “Libertarianism and Frankfurt’s Attack on the Principle of Alternative Possibilities.” Philosophical Review 104: 247–​261. Widerker, David and Michael McKenna, eds. 2003. Moral Responsibility and Alternative Possibilities: Essays on the Importance of Alternative Possibilities. Aldershot: Ashgate.

CHAPTER 6

De grees of Resp onsi bi l i t y D. Justin Coates

1.  Introduction The traditional debate about moral responsibility concerns the question of where to draw the line between responsible and non-​responsible agency: which agents can be praiseworthy and blameworthy for their actions and which cannot be? Young children, people suffering from certain kinds of cognitive, affective, or behavioral disorders, and non-​ human animals are all agents who are plausibly thought to fall below the minimal threshold of responsible agency. These agents altogether lack or only have insufficiently developed versions of the agential capacities that are required to be morally responsible for one’s actions. As such, they do not deserve praise or blame for what they do. On the other hand, mature adults are widely thought to be above the threshold.1 And so, the thought goes, most psychologically and developmentally normal adults are deserving of praise or blame for how they conduct themselves.2 There are, of course, more difficult cases. Early adolescents, psychopaths, addicts, and those laboring under great stress are all “close calls” in the sense that it’s just not clear as to whether agents of these sorts meet the minimal threshold conditions on responsible agency. These sorts of agents seem in some ways to be morally responsible for their actions. However, they also seem in some other ways to not quite meet the minimal standards on responsible agency.3 But how we regard someone, what attitudes we’ll 1  Because they are generally skeptical of anyone’s being morally responsible, responsibility skeptics deny that adults typically meet this threshold. For some skeptics (see Derk Pereboom 2001), it’s because the bar for morally responsible agency is too high for any actual human agents to meet. For other skeptics (see Galen Strawson 1994), it’s because the bar for morally responsible agency would require powers that no possible agent (human or otherwise) could possess. 2  Throughout the chapter, when I speak of agents’ actions or conduct, I mean to be as inclusive as possible, such that some omissions and even some mental attitudes can count as being part of this class. 3  See Chapter 33 (John Doris and Dominic Murphy on duress in wartime) of this volume for some of the difficulties that (some of) these classes of agents pose for theories of moral responsibility. The

112   D. Justin Coates adopt towards them, and what modes of interaction we’ll open ourselves up to crucially depends on whether or not they are in fact morally responsible. The question of what separates responsible from non-​responsible agency—​a question made all the more vivid by the fact that how we answer it has significant normative implications for how we should treat the millions of agents who fall into one of these difficult cases—​is thus an important one. Yet even if we are able to ascertain just where the threshold between responsible and non-​responsible agency lies, we can’t leave it there.4 After all, agents who meet the minimum standards on responsible agency are not all the same with respect to what they deserve. This is true, to take one obvious kind of case, when comparing two morally responsible agents, one of whom performs a qualitatively worse action than does the other. If Randall steals $20 from you and Pearl steals $2000 from you, then these two agents are deserving of very different responses, since the content of Pearl’s morally responsible action is significantly worse. But cases of this sort surely do not exhaust the scenarios in which we have reason to adjust the degree to which we hold someone responsible. Nor do they really get to the heart of the matter, since it’s perfectly sensible to say of both Randall and Pearl that they are each fully morally responsible for their actions. Indeed, the explanation of why these agents deserve different things is a function of how bad the thing they are morally responsible for doing is and not how morally responsible they are for doing it. The real problem that degrees of responsibility present for general theories of moral responsibility thus lies in cases in which we hold fixed the action type (and also, perhaps, hold fixed the circumstances in which the action is performed and the foreseeable consequences of the action).5 For even in such cases, it still seems possible that two morally responsible agents can differ significantly in what they deserve. To see this, let’s first consider Blaire, a friend who’s suffering from depression but who has promised to help you move.6 Since a general lack of motivation—​even in the face of question of the moral responsibility of children and young adolescents is not one that has been taken up in any systematic way, though philosophers working on these issues have tended to follow P. F. Strawson (1974) in thinking that children at least are exempt from moral responsibility. A notable exception to this trend is David Brink (2004). 4  In addition to the difficult cases I’ve already mentioned, the notoriously thorny debates between compatibilists and incompatibilists concerning the significance of causal determinism provide a further reason to think that this will prove to be no small thing. 5  Some are skeptical that moral responsibility is scalar in this way. And although I see no reason to deny that it is scalar—​our practices certainly seem to behave as if it is and I know of no dispositive theoretical grounds for such a denial—​I think much of what I say in this chapter can be adopted by those who are hostile to thinking of moral responsibility in scalar terms. After all, no one who is not a moral responsibility skeptic would deny that praiseworthiness and blameworthiness come in degrees. So for those doubtful about a scalar conception of moral responsibility, I invite you to appropriate what I say here about degrees of moral responsibility for a theory of degrees of praiseworthiness and blameworthiness. 6  These cases are similar to ones I developed with Philip Swenson (Coates and Swenson 2013), but unlike those, I here consider the implications for praiseworthy as well as blameworthy agency. For thoughtful engagement with those cases, see Nelkin (2016) and Tierney (forthcoming).

Degrees of Responsibility    113 a judgment that the act in question really is what you should do—​is symptomatic of clinical depression, it’s natural to think Blaire has done something quite impressive simply by keeping her promise. Indeed, Blaire seems to be more praiseworthy for mustering what little energy she has in an effort to help you than another friend, Aida, would be for doing the same thing, since helping you would be no trouble at all for Aida. Now in this case, the fact that it’s easy for Aida to help you doesn’t mean that Aida is praiseworthy only to a small degree for her action. She need not be downgraded in the degree to which she’s praiseworthy simply because she finds no difficulty in keeping her promises. Instead, facts about Blaire—​in particular, that keeping the promise took a lot of effort on Blaire’s part and that although her effort wasn’t likely to succeed, she overcame the odds—​lead us to regard her as being an exceptional case. It’s in light of these facts that we come to regard Blaire as an agent who really does deserve more praise than is ordinarily made fitting by praiseworthy behavior of this sort. Alternatively, consider a teenager, Emma, working her first job. Frustrated by the tedium of her job, Emma finds herself sorely tempted to make some easy money. It’s plausible that Emma would be less blameworthy for stealing customers’ credit card numbers than Mila, an adult with lots of customer service experience but who also finds herself deeply bored with her job, would be for engaging in the exact same kind of fraud. We mitigate blame in Emma’s case not because we’re skeptical that she’s morally responsible but because she’s “got a lot of growing up to do.” Only then, can we reasonably hold her to the normative standard in question with the same alacrity and the same intensity that’s found in expressions of blame directed towards Emma’s adult counterparts (like Mila) for their type-​identical actions.7 In other words, the fact that she’s still developing the very capacities that underwrite her status as a morally responsible agent doesn’t get Emma off the hook completely. But it does seem to rationalize importantly different responses in otherwise similar cases of wrongdoing. On a plausible theory of act-​individuation, Blaire doesn’t do anything better than Aida does, and Mila doesn’t do anything worse than Emma does.8 The fact that these agents deserve different responses doesn’t obviously depend on the relative badness of their actions (as it does in the case of Randall and Pearl). It seems, then, that agents can

7 

In “Freedom and Resentment” (1974), P. F. Strawson considers pleas of the sort that we (or third parties) avail ourselves of when confronted with blame that we don’t regard as altogether fitting. Strawson’s own treatment of these pleas leaves out the complexities that arise when thinking about diminished or enhanced moral responsibility. This is an unfortunate omission, since the thought that someone has “a lot of growing up to do” is a genuine basis for somewhat altering one’s attitudes toward agents like Emma, even though it doesn’t fit nicely into either of Strawson’s categories. A full account of these pleas must supply not only explanations of excusing and exempting conditions but also of mitigating and enhancing conditions as well. 8  One should not, I think, offer such a fine-​grained account of act-​individuation so as to include what an agent must overcome in order to perform the act in question into one’s account of what action has been performed. If one rejects this advice, then precisely the same issues that I’ll consider in this chapter arise in the context of assessing what exactly it was that the agent did rather than in assessing the degree to which an agent is morally responsible. There is no escape.

114   D. Justin Coates be more or less responsible for their actions—​that they can be morally responsible to different degrees.9 It’s of course important to know the minimal threshold for morally responsible agency, but it is, I submit, just as important to know what determines the degree to which agents are morally responsible for their actions. What this means is rather straightforward. Any adequate theory of moral responsibility must account for the fact that agents can be more or less morally responsible for their actions. To do this, I think that we have to answer two distinct questions. The first of these questions concerns what exactly it means to say that one agent is more (or less) responsible for her actions than some other agent who has also performed an action of that type or than she would be for performing the same actions in some relevantly similar set of alternative circumstances. Put differently, it’s just the question of what we mean when we say that Blaire is more praiseworthy for keeping her promise than Aida is or that Mila is more blameworthy for committing fraud than Emma is. The second of these questions concerns the conditions that can affect or alter the degree to which an agent is more (or less) morally responsible for what she’s done. This is ultimately the question of what explains why Emma is less blameworthy than Mila for committing credit card fraud. This won’t be all there is to say about degrees of responsibility, but it is the bare minimum that adequate theories of moral responsibility owe us.

2.  Being More (or Less) Morally Responsible Before going any farther, we need to address the question of what exactly it is to be morally responsible for some action. Only then can we hope to understand what exact property it is that we’re ascribing to someone when we say of her that she is more (or less) morally responsible for her action.

2.1  The Desert-​Entailing Conception of Moral Responsibility There is no one thing people are referring to when they discuss moral responsibility. Perhaps the most common conception of moral responsibility, however, is the desert-​ entailing conception of moral responsibility. Concerning this conception of moral

9 

Here you might be thinking that the agents in question merely differ in the degree to which they are praiseworthy and blameworthy rather than in the degree to which they are morally responsible. I admit this is a possibility, but it seems to me that the explanation for why we enhance praise in Blaire’s case and mitigate blame in Emma’s case is because facts about their responsibility-​grounding agential capacities make it harder for them to comply with the norms in question.

Degrees of Responsibility    115 responsibility—​a conception that traces back to (at least) the work of Spinoza—​Derk Pereboom offers the following statement: For an agent to be morally responsible for an action is for this action to belong to the agent in such a way that she would deserve blame if the action were morally wrong, and she would deserve credit or perhaps praise if it were morally exemplary. The desert at issue here is basic in the sense that the agent, to be morally responsible, would deserve the blame or credit just by virtue of having performed the action, and not, for example, by way of consequentialist considerations (Pereboom 2001, xx).10

So what separates the desert-​entailing conception of moral responsibility from other conceptions is that on this view an agent is not morally responsible simply because it’s good or beneficial to praise or blame her for what she does; rather she is morally responsible because she deserves to be praised or blamed, even if praise or blame aren’t themselves good or beneficial in the circumstances. Of course, Pereboom’s characterization of what it is to be morally responsible is not without its critics.11 But for present purposes, we can set these worries aside, since Pereboom has quite clearly adequately characterized something that many of us care about, even if in so doing, he’s characterized the phenomenon too narrowly.

2.2  Two Ways to Be More (or Less) Morally Responsible The desert-​entailing conception of moral responsibility fits nicely with the idea that agents can be more or less morally responsible for their actions. A simple extension of this conception of moral responsibility might hold that agents are more morally responsible for their actions when they are more deserving of praise or blame and that they’re less morally responsible for their actions when they’re less deserving of praise or blame. This seems right, as far as it goes, but this simple extension is ambiguous in a significant way. To get a better handle on precisely the way that this sort of talk is ambiguous, let’s return to the case of Emma and Mila. Recall that Emma and Mila each commit credit card fraud, but that because Mila is an adult and Emma is a teenager, it seems that Mila is in some sense more morally responsible for her crime than Emma is for hers. But how is this connected to what they each deserve? One possibility is that in attributing to Emma and Mila different degrees of responsibility, we’re saying that Mila is more deserving of blame for what she’s done than Emma is. Another possibility is that we’re saying that

10 

For further elucidation of these ideas, see Michael McKenna (2012) and Derk Pereboom (2014). Of note is that in his most recent restatement of this view, Pereboom adds that not only is basic desert something that obtains independently of consequentialist considerations, it also obtains independently of contractualist ones as well. 11  See James Lenman (2006) and Manuel Vargas (2013) for two examples.

116   D. Justin Coates Mila is deserving of more blame for her action than Emma is for hers. And despite the apparent similarities in these two possibilities, the properties of being more deserving and of being deserving of more are importantly different.12 To say that someone is more deserving of something suggests that it is intrinsically better to give that person whatever it is that they deserve.13 What this means is that solely from the point of view of desert it’s better to give Mila what it is that she deserves than it is to give Emma what it is that she deserves. Alternatively, we could say that in such a case there are weightier desert-​based reasons to blame Mila than there are to blame Emma. On the other hand, to say that someone is deserving of more of something (be it praise or blame) is to say something about how much praise or blame is deserved. Now unlike monetary rewards or prison sentences, deserved praise and blame resist easy quantification. But this is not to say that we can’t say something meaningful about deserving differing amounts of praise or blame. For example, the blame we bestow on someone for acting poorly can vary in intensity or duration. If we regard someone as being deserving of a great deal of blame, then we might be more reticent to extinguish our blaming attitudes. We might also experience the blaming attitudes more intensely. And this, in turn, might more powerfully motivate us to express these attitudes. What this means is that if an agent is deserving of more blame then she is deserving of more intense blame, or of being targeted with blame for longer, or having more direct expressions of blame being leveled at her. On the other hand, if an agent is deserving of less blame, then she is deserving of less intense blame, or blame that doesn’t last as long, or blame that’s less intrinsically motivating. The degree to which an agent is deserving of praise or blame seems to depend on these two independent variables: how weighty the desert-​based reasons there are to praise or blame and how much praise or blame the agent deserves. These two properties can combine in ways that explain what exactly it is that people deserve in light of their actions.14 Recall Blaire, who helped you even though it was quite difficult for her to overcome temptation that had its source in her depression. Blaire is very praiseworthy in the sense that you should be more effusive with your praise (i.e., she is deserving of more praise), but she is also very praiseworthy in the sense that the desert-​based reasons to praise her in the way that she deserves are quite weighty. To see this, suppose you’re talking to a mutual friend about your move. If you fail to praise Aida, who helped you, but for whom helping was very easy, then you’ve missed a chance to give credit where credit’s due. This is bad, given the fact that your conversation was about your move. Yet

12 

For more on this distinction, see Robert J Hartman (2017) and D. Justin Coates (2019). Talk about “intrinsic value” is often contested. However, the point I’m making here can be given other metaethical glosses—​e.g., one can make the point not in terms of value but instead in terms of fittingness. Although I find it more salutary to talk of value or reasons, I want the view I sketch here to be neutral with respect to these delicate metaethical issues. 14  I expand upon this point with the help of desert graphs—​an apparatus for understanding desert developed by Shelly Kagan (2012)—​in greater detail in Coates (2019). 13 

Degrees of Responsibility    117 this failure seems amplified if you were to fail to praise Blaire. A possible explanation for this, I think, is that the desert-​based reasons to praise Blaire are weightier than the desert-​based reasons to praise Aida. It seems, in some sense, better to praise Blaire than to praise Aida (though plausibly good to praise each). As such, failure to act on those reasons absent weighty countervailing reasons is correspondingly greater in Blaire’s case than it is in Aida’s. Having these two variables to play with allows us to conceptually map the different degrees to which agents might be morally responsible for their actions. That is, this framework (at least when it’s suitably worked out) helps us to clearly articulate what exactly we’re attributing to an agent when we claim that she is more (or less) responsible. However, it doesn’t yet tell us the conditions under which agents are more (or less) morally responsible for their actions. I turn to this question now.

3.  When Are Agents More (or Less) Responsible? What, then, are the conditions under which agents are more (or less) morally responsible for their actions? In particular, how do differences in the intrinsic properties of an agent or in the circumstances that agents find themselves affect the degree to which they are morally responsible?

3.1  Three Proposals: Quality of Will, Epistemic Position, and Control There are three common ways of explaining why one agent is more (or less) morally responsible than another (or than she herself would be were circumstances different) for performing type-​identical actions. At first glance, these explanations appear to be quite distinct from one another. As we’ll see, however, they actually rely on a shared assumption about the connection between how hard it is to comply with legitimate normative standards and the degree to which an agent is morally responsible for her compliance (or non-​compliance, as the case might be). The first proposed explanation appeals to the quality of agents’ wills. On this proposal, agents with better (or worse) qualities of will can differ in the degree to which they are morally responsible for performing some action. If this is right, then we’ll need an account of what “quality of will” comes to exactly. One such account appeals exclusively to the intrinsic desires that are manifested in an agent’s action. Another way of understanding quality of will emphasizes the quality of the reasons that move an agent to action. On these accounts, the degree to which an agent is morally responsible depends,

118   D. Justin Coates inter alia, on the degree to which one’s action manifests intrinsic desires of the relevant sort or on the degree to which one’s action manifests better (or worse) reasons. Or both!15 A second proposal appeals to an agent’s epistemic position with respect to the moral status or her action. Many agents are ignorant of the moral status of what they do. Some of these agents are culpably ignorant. Ignorance of this sort plausibly doesn’t excuse, but it does seem that an agent who knowingly does evil would be more responsible for what they’ve done than someone who does evil due to culpable ignorance (compare, e.g., the reactions one has to Iago’s villainy to Othello’s tragic rage). So perhaps culpable ignorance mitigates responsibility.16 Other agents, however, are apparently non-​culpably ignorant. What should we say about these individuals’ degree of moral responsibility? Does non-​culpable ignorance excuse, as some have suggested? Or does it merely mitigate? A third proposal posits that the degree to which an agent is morally responsible for her action depends at least in part on the degree to which that agent has control over what she does. On one prominent theory, having control over one’s action amounts to the capacity to guide one’s behavior in light of what reasons one recognizes oneself to have. If this is right, then improvements to or deficiencies in one’s ability to recognize reasons or to translate those reasons into action will affect the degree to which you control that action. However, if control depends more broadly on the quality of one’s opportunity, where this is determined not only by the agent’s own rational capacities but also by the circumstances that she finds herself in, then the degree to which she is morally responsible depends on a more holistic assessment of her situation. In either case, differences in the degree to which agents have control over their actions do seem relevant to the degree to which those agents are morally responsible for what they do.

3.2  Unifying the Proposals: The Significance of Difficulty Rather than viewing these proposals as being radically at odds with one another, I’m inclined to think that each provides us with a partial account of the conditions under which agents are more (or less) morally responsible for their actions. This might be taken to suggest that there’s no one thing underlying the conditions that affect the degree to which an agent is morally responsible. However, I don’t think that’s right. The explanations for why agents are more (or less) responsible that these proposals offer are more unified than it appears at first glance. The common thread in each of these proposals is that they each get their appeal in light of the following thought: how responsible you are for what you’ve done depends 15 

As we’ll see, these proposals are, for the most part, consistent with one another. So it might be that it is only when they’re each given their due that we’ll have a suitably fleshed-​out explanation of why it is that some agents are more (or less) morally responsible for their actions. 16  Or alternatively, perhaps knowingly doing evil enhances the degree to which an agent is morally responsible.

Degrees of Responsibility    119 on how difficult it was for you to comply with legitimate normative expectations. To wit: if it’s easier for me to do the right thing, then I show worse quality of will for doing the wrong thing than I would if it had been hard to comply with legitimate norms. That’s what explains why I’m more blameworthy in the former case than I would’ve been in the latter. Or, if I’m ignorant of some important moral fact, but it would have been very hard for me to have come to the truth, then I am less blameworthy for acting wrongly than if I could have very easily arrived at the moral facts but didn’t concern myself to do so. Similarly, if circumstances make it harder for me to control myself in the way that’s expected of me, then I seem to be more praiseworthy for doing the right thing precisely because in so doing, I overcame a significant difficulty. This straightforward explanation of the connection between degrees of responsibility and praiseworthiness and blameworthiness seems to rely on what Holly Smith (1991) has called the battle citation model of moral praise, and on a related model of moral blame. On the battle citation model, “an agent is creditable for performing a right action if and only if a morally good desire won a hard battle in the war against temptation to perform the wrong act,” (Smith 1991, 281-​82). Since this model posits that what it is in virtue of which an agent deserves moral credit or praise is that he or she has won a hard or difficult battle, it seems plausible on the very same grounds that how deserving of praise the agent is depends (at least in part) on just how difficult it was for her to win the battle. If beating the temptation in question constituted defeating “overwhelming odds” then it stands to reason that the agent accrues more credit or praiseworthiness than if it was merely a hard-​fought-​but-​still-​manageable battle against some moderately recalcitrant desire. After all, mustering the effort required to overcome an extremely powerful temptation reveals that you are deeply committed to the good. In other words, overcoming a difficult temptation reveals that you have a good will. Difficulty thus seems to enhance praiseworthiness. By parity of reasoning, how difficult it is to comply with normative standards is also relevant to how blameworthy an agent is. Suppose that you fail to comply with legitimate standards in a case in which it would’ve been quite difficult for you to have done the right thing. Here, difficulty seems to mitigate rather than enhance blameworthiness precisely because when it’s very hard to comply with legitimate norms—​norms that exist at the edge of what we can reasonably expect of one another—​agents seem less blameworthy when they fail to comply with them. One explanation of this is that if it’s easy for me to comply with a normative standard, then failure to do so will typically reveal that I do not care about meeting that standard. On the other hand, failure to comply with difficult but legitimate demands does not necessarily show me to be indifferent about meeting the standard in question. Smith’s battle citation model of moral praiseworthiness and the corresponding model of moral blameworthiness provide us with a conceptual basis for the slogan that difficulty determines degrees.17 In other words, it provides us with a framework for understanding 17 

In her work, Smith considers other models of moral praise (and corresponding models of moral blame) that might be tied to something other than the difficulty of compliance. There is much in Smith’s

120   D. Justin Coates how the difficulty of complying with legitimate norms regularly explains our judgment of others as being more (or less) morally responsible for their conduct. This does not mean, of course, that it’s only difficulty that determines the degree to which an agent is morally responsible for what she does. But appeals to difficulty do a lot of explanatory work in this domain. Indeed, the three evidently disparate proposals that I started with all seem to agree on, and are perhaps even based on, this basic idea. As such, to properly evaluate these proposals, we’ll need to see exactly how difficulty affects the quality of an agent’s will, how culpable she might be for ignorance, or the degree to which she can be said to control her action.

3.3  Quality of Will One explanation of why two agents who perform type-​identical actions could be morally responsible for doing so to different degrees is that one agent’s action expresses more good or ill will than did the other’s. This seems to follow from the widely held view that praiseworthiness and blameworthiness track, inter alia, quality of will.18 After all, if it’s the case that I must display good or ill will in order to be praiseworthy or blameworthy for my action, then how much good or ill will I display will render me more (or less) deserving of praise or blame for that action. And since what it is to be morally responsible for some action is to be deserving of praise or blame for that action, then it seems like the degree to which I display good or ill will in my actions is directly relevant to the degree to which I am morally responsible for those actions. Nomy Arpaly and Tim Schroeder (2014) have recently proposed an account along these lines. On their view, (complete) good will is “an intrinsic desire for the right or the good, correctly conceptualized,” and (complete) ill will is “an intrinsic desire for the wrong or the bad, correctly conceptualized,” (Arpaly and Schroeder 2014, 162).19 From this they build scalar theories of praiseworthiness and blameworthiness. An agent is praiseworthy for some action, on their view, to the extent that that action manifests good will (as they’ve defined it in terms of intrinsic desires for the right or the good). On the other hand, an agent is blameworthy for some action to the extent that that action manifests ill will (again, as they’ve defined “ill will”).20 There’s something to this thought. Agents who more fully manifest intrinsic desires for the good do seem more praiseworthy for what they’ve done than agents who only partially manifest desires for the good. And agents who more fully manifest intrinsic

discussion to recommend but because the other models that she considers do not correspond as directly to issues that are tied to degree of responsibility, I will not consider them here. 18 

In the contemporary debate, most trace this thesis to P. F. Strawson (1974), but the general thought goes back at least to Adam Smith (1759/​1976). 19  In this same passage, Arpaly and Schroeder also define partial good will and partial ill will, but for our purposes that level of detail is unnecessary. 20  Arpaly and Schroeder (2014): 170.

Degrees of Responsibility    121 desires for the bad seem correspondingly more blameworthy than agents who manifest a mixture of desires or who appear to be reluctant in going along with something bad. This might not provide a general explanation of the conditions under which agents are more (or less) morally responsible, but it does appear to have explanatory power in a circumscribed range of cases. Yet despite its initial plausibility, Dana Kay Nelkin (2016) has raised an important objection to accounts like Arpaly and Schroeder’s that take degrees of responsibility to be explained by more (or less) good or ill will. Nelkin invites us to: Consider a case in which a person is blamelessly drunk at a party. In this state, she gleefully and loudly shares a scandalous secret, told to her in confidentiality about another guest. Suppose that this person would never have done this if she had been sober, and she values keeping promises and respecting others’ privacy. Still, she harbors an intrinsic desire for the other guest to suffer embarrassment [i.e., she has ill will], and it is this desire on which she acts. This seems a case in which the action manifests significant ill will. She also has good will, which normally would overcome the ill. But she is drunk and it does not, (Nelkin 2016, 363-​64).

From this case Nelkin goes on to argue for a more general point, which is simply that your action can manifest ill will even if you’re excused, such that you’re not morally responsible at all for that action.21 As a result, it cannot be the quality of the blameless drunk’s will that explains the degree to which she is responsible, since in this case, she has both a poor quality of will and is not morally responsible at all. This is a powerful challenge to Arpaly and Schroeder, and I think Nelkin is right in her assessment of the case as she describes it. But I also think she’s wrong (or at least, too quick) to draw the general conclusion that she does. That is, Nelkin’s right to conclude that differences in good and ill will as described by Arpaly and Schroeder are insufficient to account for the degree to which agents are morally responsible in cases of this sort. But this, I submit, should lead us to doubt Arpaly and Schroeder’s account of good and ill will rather than the more general claim that the quality of an agent’s will affect the degree to which she is morally responsible for the actions that manifest that will. After all, on one natural interpretation of the case that Nelkin describes, the blameless drunk doesn’t actually display ill will at all. It’s true, of course, that she’s mean-​spirited and maybe even malicious. But young children can be mean (and maybe even malicious) without manifesting ill will in the sense at stake.22 So manifesting good and ill will

21 

Nelkin frames the point slightly differently, since what she’s after is an account of difficulty of the sort that might affect the degree to which an agent is praiseworthy or blameworthy for her action. But in the case of the blameless drunk, she’s less responsible because it would’ve been difficult for her to comply with legitimate expectations even though that difficulty doesn’t seem to affect the degree to which she displays ill will in this case. 22  This thought follows from a roughly “Strawsonian” conception of quality of will. Although P. F. Strawson (1974) never explicitly defines what he takes good or ill will to be, it’s clear that he regards young children as being unable to display good or ill will of the sort that the reactive attitudes respond to (they

122   D. Justin Coates in one’s actions thus requires some further connection between the intrinsic desires for the good or the bad and the mechanisms that translate these desires to action. It’s only when these mechanisms have some further property—​e.g., that they are suitably reasons-​responsive, or that the intrinsic desires for the good or the bad reflect a suitable understanding of what’s at stake—​that the manifestation of one of the relevant intrinsic desires in an agent’s action counts as manifesting good or ill will. It’s therefore possible to maintain the initial thought that differences in the degree to which agents are morally responsible are explicable by appeal to differences in the quality of their wills. But to clearly see how quality of will might still matter, let’s reconsider the cases of Blaire and Aida. Unlike the blameless drunk or a young child, these two agents plausibly meet the minimum threshold on morally responsibility. This means that the mechanisms issuing in their actions suitably instantiate whatever property (e.g., reasons-​responsiveness) that’s necessary for the manifestation of one of the relevant intrinsic desires to count as good or ill will. In these cases, Nelkin’s worries won’t obviously apply. Both Blaire and Aida are connected to their actions in the “right” way; they both possess the kind of control that’s necessary for moral responsibility. Yet Blaire seems more responsible for keeping her promise than Aida does. The explanation for this is just that Blaire has manifested an even better quality of will towards you than Aida has (or perhaps any agent could reasonably be expected to), since given her severe depression, the alternatives were more (rationally) tempting for Blaire than they were for Aida.23 More (rationally) tempting, and yet, Blaire overcomes that temptation. As Smith helpfully puts it, Blaire won the battle, and this triumph speaks especially well of her quality of will.24

are exempt from moral responsibility, after all). But no one with young children would deny that they could be mean—​after being very gently scolded for continually ignoring a day care employee, my three-​ year-​old daughter recently informed my wife that she was, “just a little mean sometimes, mommy.” 23  Along these lines, Hannah Tierney (2019) has recently argued that the degree to which an agent is morally responsible depends in part on “the quality of the reasons” for which she acts. Tierney takes this to show that the degree to which an agent is morally responsible depends on more than how difficult it is (or would be) for an agent to comply with legitimate normative standards. In some cases, Tierney tells us, it depends on the “quality of [an agent’s] reasons,” where this concerns the moral, epistemic, or “popular” status of those reasons. I think that Tierney’s right to think that degrees of responsibility depend (at least in part) on the quality of an agent’s reasons, but I’m not sure that the quality of reasons can be understood without appeal to some conception of difficulty. After all, how weighty a reason is can (and often does) depend on how difficult it is to comply with that reason. 24  This isn’t to say that there are no control-​relevant differences between the degree to which Blaire is morally responsible and the degree to which Aida is morally responsible. I’ll argue in section 3.5 that there are. But whatever one thinks about the connection between control and the degree to which an agent is morally responsible, once we correct for the problem in Arpaly and Schroeder’s view, there’s room left for the quality of an agent’s will to matter for the degree to which she is morally responsible.

Degrees of Responsibility    123

3.4  The Epistemic Condition Whether I am morally responsible at all for performing an action depends not only on whether I meet the control condition on moral responsibility, but also on whether I meet some further epistemic condition on responsibility. It stands to reason, then, that differences in agents’ epistemic statuses can affect the degree to which those agents are morally responsible for their actions. One factor that affects whether an agent meets the epistemic condition on moral responsibility is her attention to and awareness of morally relevant facts.25 These are the building blocks on epistemic and moral conscientiousness, which serve to protect agents from the epistemic and moral mistakes that we’re readily prone to make. Yet agents can attend to moral considerations to varying degrees. And unfortunately, even when an agent is especially attentive to morally relevant considerations she might nevertheless fail to “put it together” in a way that leads her to recognize the moral truth. In such a case, she seems (at least) less blameworthy for acting in a way that’s contrary to how she ought to act than she would be had she simply ignored those considerations. Conscientiousness of this sort seems to mitigate blameworthiness because genuinely conscientious agents display good will even when they get it wrong; and this tells against them being deserving of blame to the same degree that non-​conscientious wrongdoers are deserving of blame.26 But if conscientiousness mitigates in ordinary cases, then maybe in exceptional cases—​cases in which even maximal conscientiousness wouldn’t lead the agent to the right conclusion—​it fully exculpates. Gideon Rosen, to name one example, takes up this thought, claiming that “[g]‌iven the intellectual and cultural resources available to a second millennium Hittite lord, it would have taken a moral genius to see through the wrongness of chattel slavery,” (Rosen 2003, 66). Such an agent, Rosen concludes, is blameless: “in my view it makes no sense to hold this injustice against the perpetrator when it would have taken a miracle of moral vision for him to have seen the moral case for acting differently,” (Rosen 2003, 66). Here, Rosen seems to be thinking that because no amount of careful reflection (and indeed, nothing short of a miracle) would have led the Hittite lord to sensible views about slavery, he’s excused for his wrongdoing. This is startling, since Hittite lords knowingly and intentionally treated their slaves with brutal cruelty. But if Rosen’s exculpatory conclusion simply follows from the thought that some degree of conscientiousness can mitigate, then maybe we should regard that idea to be dubious.

25  This thought calls to mind Neil Levy’s (2014) “consciousness thesis,” which holds that an agent is morally responsible for her action only if she is conscious of the facts that give that action its moral significance. 26  This part of the explanation of why epistemic conscientiousness mitigates fault is related to Tierney’s quality of reasons thesis, which concerns not only the moral reasons for which agents act but also the epistemic reasons that they regard as action-​guiding.

124   D. Justin Coates In fact, however, there’s much to be said in response to Rosen’s claims that moral ignorance is an excusing (rather than a merely mitigating) condition. First, it’s doubtful that most Hittite lords were even minimally conscientious with respect to the question of chattel slavery. As such, if they really are blameless (or weaker: if they are significantly less blameworthy than contemporary slaveholders would be27), it can’t be because they’re working hard to arrive at the truth, if only to an imperfect degree. More generally, however, it’s not clear why we should be so quick to reject the Hittite lords’ responsibility in cases like this. Just because it’s difficult—​nearly impossible!—​for them to know that slavery is wrong doesn’t mean they cannot be morally responsible for that action, at least to some degree. Difficulty might determine degree, but not in the limiting case. That is, difficulty isn’t per se exculpatory. Alex Guerrero (2017) develops a version of this thought by distinguishing between three types of difficulty: skill-​related difficulty, effort-​related difficulty, and difficulty in trying. On this taxonomy, we can say that it might be difficult for the ancient slave owner to have gotten it right due to some skill-​related difficulty that’s inherent to solving the moral problem at issue. Here we’re imagining that figuring out that slavery is wrong is somehow akin to figuring out a complicated mathematical problem. Alternatively, it might be difficult for the ancient slave owner to have gotten it right due to some effort-​ related difficulty that arises in the psychology of the slave owner. Maybe in this case, it would’ve just been very hard for the ancient slave owner to give the question the attention required in order to figure out the answer, even though, were he to have done so, arriving at the correct answer would have been easy. A third possibility is it’s difficult for the slave owner to get it right because it would have never even occurred to him to try to figure out the moral status of slavery.28 However, Guerrero claims that whether one thinks that difficulties of these kinds mitigate ultimately depends on the theory of moral reasoning one accepts. If you accept a theory according to which moral reasoning is relatively straightforward, then it’s doubtful that skill-​related difficulties are really present in the case at hand. If, on the other hand, you accept a theory according to which successful moral reasoning typically requires a great deal of cognitive sophistication and expertise, then, Guerrero claims, it’s doubtful that skill-​related difficulties would mitigate. On this point he does admit that it’s plausible that coming to believe novel moral truths might have skill-​related difficulties, but he counters this with the equally plausible claim that difficulties of this sort can’t be what explains why Hittite lords fail to believe what’s true, since they don’t even try to figure things out. Yet what Guerrero says here is consistent with the view that failing to try something that’s hard to do mitigates less than trying and failing would. As

27 

See Erik Loomis (2015) for an up-​to-​date account of the role of slave labor in contemporary supply chains. Unlike ancient slave owners, contemporary ones inhabit a world where egalitarian values are widespread (if not in practice, at least in principle). Accordingly, they can’t pretend not to know that the people they’re enslaving matter and should be treated accordingly. 28  Guerrero rightly notes that it also could’ve been difficult in all three of these ways.

Degrees of Responsibility    125 a result, if arriving at moral truths does involve skill related difficulties, then Guerrero hasn’t yet given us a reason to deny that it mitigates (at least to some small degree). Later, Guerrero claims that to the degree the Hittite lord faces effort-​related difficulty or difficulty in trying, those forms of difficulty only arise as a “result of the agent’s moral attitudes/​character,” (213) and so don’t mitigate the degree to which the agent is morally responsible.29 But if Rosen goes too far in denying that the Hittite lord is blameworthy for owning slaves, Guerrero overcorrects on this point. The fact that some effort-​related difficulty, or difficulty in trying, results from an agent’s moral attitudes or character will fail to mitigate only if that agent is fully morally responsible for those attitudes or character. But plausibly, if an agent is less than fully responsible for those attitudes or character, then difficulties that arise from them would mitigate the degree to which she is blameworthy for failing to arrive at the right conclusion. To see this, consider a scenario in which you faced an effort-​related difficulty that was due to your moral attitudes or character—​say, a case in which it’s hard for you to overcome prejudiced attitudes that you have as a result of being brought up in a cult that brainwashes children from a young age. Here, your prejudice makes it hard for you to supply the effort needed to see that everyone is equally deserving of respect. That prejudice is itself an element of your moral character. And yet, it’s not one that you are morally responsible for, since it’s uncontroversial that one is not morally responsible for attitudes one has as a result of brainwashing. Now, since the prejudice in question only makes it difficult for you to muster the effort needed to come to a morally salutary picture of the world, it doesn’t wholly undermine your responsibility for failure to do so. But surely the fact that you were brainwashed matters normatively. Surely, it’s relevant to our assessment of your responsibility that the moral character trait that explains the effort-​related difficulty is one that you can’t be held responsible for. We wouldn’t, after all, regard you as being just as responsible for expressions of your prejudice as equally odious expressions of prejudice coming from the cult leader whose prejudice is one that can be traced back to a character trait for which he is responsible. But this is analogous to the position that the Hittite lord is in. After all, if he is fully blameworthy for his attitudes, then we’d have reason to treat him in precisely the same way that we have reason to treat a contemporary who argues for the moral permissibility of chattel slavery. Yet it’s doubtful that we do have reason to treat these two agents in exactly the same way. If I were to meet (roughly) a peer who apologized for slavery, I’d be inclined to simply blame him. On the other hand, if I met a time-​traveling Hittite lord I’d feel a strong pull towards at least trying to educate him.30 This, of course, doesn’t mean 29 

Guerrero admits that this claim rests on a controversial conception of agency and responsibility. I agree with this, though I suspect Guerrero and I come down on different sides of the controversy. 30  To make this point a bit differently, consider what would be required for you to give these two agents the benefit of the doubt. Plausibly, each of them would exhaust your generosity fairly quickly, but it also seems plausible that you’d have reason to withdraw your generosity even more quickly in the case of the contemporary slavery apologist. The explanation for this: there are weightier desert-​based reasons to blame him than there are to blame the Hittite lord. But as I argued in section 2.2, this just is what it is for the former agent to be more blameworthy than the latter.

126   D. Justin Coates that he’s not blameworthy to some degree, but it does suggest that the desert-​based reasons to blame the Hittite lord are weaker than the desert-​based reasons to blame the contemporary defender of slavery. There is a final reason to think that the Hittite lord’s blameworthiness is to some degree diminished. Suppose that the Hittite lord did engage in the kind of reflection that epistemically and morally conscientious agents are prone to, and that after a lot of careful thought, he came to believe that slavery was immoral. It seems plausible that in this scenario, the Hittite lord would be more praiseworthy for coming to accept a moral outlook that proscribes slavery than, say, contemporary liberals are for having the same moral outlook. After all, moral knowledge that is hard won is more laudable than the same moral knowledge learned at your parent’s knee. But if the difficulty of this undertaking didn’t mitigate blameworthiness in the case in which he failed to reflect, as Guerrero maintains, it’s doubtful that it could enhance praiseworthiness in the case in which the Hittite lord did reflect and after a grueling bout of self-​assessment came to believe that he was doing something quite horrific. So if we want to maintain the plausible claim that in this case the Hittite lord is more praiseworthy for getting it right (and acting on that belief) than I am (or would be) for doing the same, we must also accept the refrain I keep returning to: that (all together now) difficulty determines degrees. What this means is that although I heartedly agree with Guerrero that Rosen too easily lets ignorant wrongdoers off the hook, I also think that how hard it can be to get the facts right (in any of the senses of difficulty Guerrero articulates) is directly linked to the degree to which agents are responsible.

3.5  Control According to one widespread view, agents are morally responsible only if they control their actions.31 But what does control of the relevant sort consist in? Minimally, it requires that an agent is suitably reasons-​responsive. After all, if an agent is incapacitated in a way that renders the mechanisms that issue in her actions non-​receptive to reasons or non-​reactive to those reasons, then she is not normatively (or rationally) competent in the way that is the basis for well-​functioning moral agency.32 But it can be more or less difficult to recognize reasons for (or against) some course of action. So too, it can be more or less difficult to translate that recognition into action. And the degree to which an agent controls her action plausibly depends on how easy or hard it is for the mechanisms issuing in her actions to issue in a right action. So far, so good. But if we

31  The notion of control at issue here is connected to freedom of the sort that is putatively required for moral responsibility. 32  Here I’m following John Martin Fischer and Mark Ravizza (1998) by talking about the control-​ grounding properties of the mechanisms that issue in an agent’s action rather than the agent herself. If this worries you, then I think almost everything I say can be translated into an idiom that appeals only to the control-​grounding properties of agents themselves without remainder.

Degrees of Responsibility    127 want to secure this explanation, then we’ll need to say more about what it is in virtue of which it might be more or less difficult for the mechanisms to issue in actions that are suitably responsive to sufficient reasons for action. In recent work with Philip Swenson (2013), I’ve argued that we can do just this by building on extant accounts of control. We start with an influential account of guidance control due to John Martin Fischer and Mark Ravizza (1998). According to this account, agents possess the kind of control required for moral responsibility only if the mechanisms that issue in their actions are “moderately reasons-​responsive.” For Fischer and Ravizza, this amounts to the claim that: A mechanism of type K is moderately responsive to reason to the extent that holding fixed the operation of a K-​type mechanism, the agent would recognize reasons (some of which are moral) in such a way as to give rise to an understandable pattern (from the viewpoint of a third party who understands the agent’s values and beliefs), and would react to at least one sufficient reason to do otherwise (in some possible scenario). That is, a mechanism is moderately responsive to reason insofar as it is “regularly” receptive to reasons (some of which are moral), and at least weakly reactive to reasons (Fischer and Ravizza 1998, 243-​244).

Of particular note here is the idea that a mechanism is weakly reactive to reasons just in case that mechanism would issue in the action for which there is sufficient reason in at least one possible world.33 This seems plausible as a bare minimum. After all, the mechanisms issuing in my actions—​mechanisms that include practical reasoning, habit, instinct, etc.—​can’t count as being responsive if there’s no possible world in which, given sufficient reason to do otherwise, they do so. If they failed to have this property, then we would evidently lack the very capacity that makes us rational agents. Since Fischer and Ravizza’s analysis works well as an account of the bare minimum that’s required for morally responsible agency. Swenson and I extend it in the following key way. If one is weakly reactive to reasons in virtue of what happens in one possible world, the degree to which an agent is weakly reactive to reasons depends on the comparative similarity of the possible world in which the mechanism reacts to the sufficient reason to do otherwise. If that world is closer, the thought goes, then there is more overlap in what it and the actual world share. So if we can hold more things fixed between these two worlds, then it’s easier in some meaningful sense for the agent’s mechanism to issue in the right action. And if it’s easier to get it right, then an agent seems more blameworthy for getting it wrong. We motivate this by considering a case not unlike that of Blaire and Aida. However, in our original case, the two agents, Marcia and Thomas, fail to keep their promise to help 33 

Swenson and I also discuss reasons-​receptivity and offer a model of how that might vary in a way that explains variances in the degree to which agents are morally responsible. For our current purposes, however, I think it’s fine focus exclusively on extending Fischer and Ravizza’s account of weak reasons-​ reactivity, since what we say about receptivity is, to a large degree, isomorphic to what we say about reactivity.

128   D. Justin Coates you. Marcia fails because she is suffering from severe but non-​debilitating depression. And even though it would’ve been easy for Thomas to keep his promise, he fails simply because at the time he needed to leave to go help you, he was moved to sit around at home and watch reruns instead. In light of these cases, Swenson and I claim: In our view, to say that it is “more difficult” or that it is “harder” for Marcia to keep her promise is to say that in the relevant sense, the world in which she does so is less accessible from the actual world. Of course, what makes worlds more or less accessible is a matter of comparative similarity . . . And we reductively analyze the notion of “difficulty” in terms of comparative similarity, (Coates and Swenson 2013, 638–​39).

It’s more difficult for agents like Marcia to respond to good reasons because the worlds in which they do so are more distant. They therefore have less control over their actions, and are, accordingly, less responsible for those actions. Here again we see difficulty determining degrees. Here it does so because how easy or hard it is for these agents to comply with legitimate normative standards affects the degree to which they possess the kind of control that’s necessary for morally responsible agency. The account that Swenson and I offer seems to get it right in a wide range of cases in which, intuitively, agents are more (or less) morally responsible for their actions. Despite that, it can’t serve as an explanation of why agents have differential degrees of control. Nelkin (2016), in particular, has pointed out two important problems for our view. The first is that it’s generally a mistake to appeal to modal properties to explain things like difficulty. More plausibly, the modal properties merely co-​vary with difficulty because, for example, the fact that it’s hard for a depressed agent to be moved to act on the reasons she judges herself to have is what grounds the modal properties in question (viz., that it’s a distant world in which given sufficient reasons to do otherwise, the mechanism in question issues in an action guided by such reasons). In other words, it’s the difficulty of being moved by good reasons that explains the modal properties, not the other way around. This point is almost surely correct. We should have been content to identify the modal differences that track difficulty of the sort that’s relevant to how well an agent controls her action, as even this would provide us with what we would need to figure out when to mitigate or enhance the degree to which we should hold someone responsible. The second objection Nelkin has to the view is that it’s overly narrow. Swenson and I explain difficulty of the sort that’s relevant to control wholly in terms of the modal properties of agent’s own reasons-​responsive mechanisms. However, even if we hold fixed all the intrinsic properties of these mechanisms, the control-​related difficulties an agent faces might vary according to the circumstances the agent finds herself in.34 And if this is right—​and I think it’s plausible that it is—​then the degree to which an agent is morally

34  Perhaps the best recent treatment of how an agent’s circumstances can enhance or diminish her control is due to Manuel Vargas (2013). See especially Vargas’s treatment of an agent’s “moral ecology” (Vargas 2013: 243ff.).

Degrees of Responsibility    129 responsible might depend, at least in part, on the quality of the opportunity she has to exercise her control. For Nelkin, this means that difficulty depends on two variables: the control-​grounding mechanisms and the control-​grounding circumstances.35 So if one’s control-​grounding mechanisms are more developed, or if one finds oneself in circumstances that better facilitate good decision-​making, then it will be easier for you to recognize and respond to good reasons for action. If, on the other hand, one’s control-​grounding mechanisms are diminished or impoverished in some way, or if one finds oneself in circumstances that undermine good decisions, then it’ll be harder for you to recognize and respond to good reasons. In the former case, you have more control over what you do—​you have, in Nelkin’s terminology, a better quality of opportunity. In the latter case, you have less control over what you do, or again, a worse quality of opportunity. This more comprehensive account of control is better equipped to explain why having more or less control affects the degree to which an agent is morally responsible: the quality of opportunity an agent has grounds difficulty facts of the sort that determine degrees.

4.  Conclusion Agents can be more (or less) morally responsible for their actions. As a result, an adequate theory of moral responsibility owes us (i) an account of what this means and (ii) an explanation why some agents are more (or less) morally responsible. In this chapter I’ve sketched how one might start going about each of these tasks. But it would be foolish (or maybe, foolishly self-​deceived) to pretend that I’ve worked out the details all the way here.36 More work is therefore needed.

References Arpaly, Nomy and Tim Schroeder. 2014. In Praise of Desire (Oxford: Oxford University Press). Bradford, Gwen. 2015. Achievement (New York: Oxford University Press). Brink, David O. 2004. “Immaturity, Normative Competence, and Juvenile Transfer: How (Not) to Punish Minors for Major Crimes,” Texas Law Review vol. 82, pp. 1555–​1585.

35  Nelkin would resist characterizing both these elements in terms of control, but I do so because I think, in light of what’s clearly correct about this objection, we should conclude that the nature and degree of control an agent possesses depends at least in part on facts that are extrinsic to the agent. 36  There are at least two glaring omissions in the treatment of these issues here. The first of these simply concerns the nature of difficulty itself. For the most part, I’ve left difficulty unanalyzed in this paper. Elsewhere, Alex Guerrero, Dana Nelkin, and Gwen Bradford (2015) have done some good work in the service of this goal, but more work is needed. The second omission is that I’ve ignored other determinates of degrees. It’s doubtful that only difficulty serves as the basis for an agent being more (or less) morally responsible. So we need an account of these other determinates, whatever they are.

130   D. Justin Coates Coates, D. Justin. 2019. “Being More (or Less) Blameworthy,” American Philosophical Quarterly 56 (3):233–​246. Coates, D. Justin Coates and Philip Swenson. 2013. “Reasons-​Responsiveness and Degrees of Responsibility,” Philosophical Studies vol. 165, no. 2, pp. 629–​645. Fischer, John Martin and Mark Ravizza. 1998. Responsibility and Control (Cambridge: Cambridge University Press). Guerrero, Alex. 2017. “Intellectual Difficulty and Moral Responsibility,” in Responsibility: The Epistemic Condition, eds. Philip Robichaud and Jan Willem Wieland (Oxford: Oxford University Press), pp. 199–​218. Hartman, Robert J. 2017. In Defense of Moral Luck: Why Luck Often Affects Praiseworthiness and Blameworthiness (New York: Routledge). Kagan, Shelly. 2012. The Geometry of Desert (New York: Oxford University Press). Lenman, James. 2006. “Compatibilism and Contractualism: The Possibility of Moral Responsibility,” Ethics vol. 117, no. 1, pp. 7–​31. Levy, Neil. 2014. Consciousness and Moral Responsibility (Oxford: Oxford University Press). Loomis, Erik. 2015. Out of Sight (New York: The New Press). McKenna, Michael. 2012. Conversation and Responsibility (New York: Oxford University Press). Nelkin, Dana K. 2016. “Difficulty and Degrees of Moral Praiseworthiness and Blameworthiness,” Nous vol. 50, no. 2, pp. 356–​378. Pereboom, Derk. 2001. Living Without Free Will (Cambridge: Cambridge University Press). Pereboom, Derk. 2014. Free Will, Agency, and Meaning in Life (New York: Oxford University Press). Rosen, Gideon. 2003. “Culpability and Ignorance,” Proceedings of the Aristotelian Society vol. 103, no. 1, pp. 61–​84. Smith, Adam. 1759/​1976. The Theory of Moral Sentiments (Oxford: Oxford University Press). Smith, Holly M. 1991. “Varieties of Moral Worth and Moral Credit,” Ethics vol. 101, no. 2, pp. 279–​303. Strawson, Galen. 1994. “The Impossibility of Moral Responsibility,” Philosophical Studies vol. 75, no. 1–​2, pp. 5–​24. Strawson, P. F. 1974. “Freedom and Resentment,” Freedom and Resentment and Other Essays (London: Methuen & Co. Ltd), pp. 1–​25. Tierney, Hannah. 2019. “Quality of Reasons and Degrees of Moral Responsibility,” Australasian Journal of Philosophy 97 (4):661–​672. Vargas, Manuel. 2013. Building Better Beings (New York: Oxford University Press). Watson, Gary. 1996. “Two Faces of Responsibility,” Philosophical Topics vol. 24, no. 2, pp. 227–​248.

CHAPTER 7

Grou p Resp onsi bi l i t y Christian List

A central feature of our social world is our practice of assigning responsibility to one another for our actions. We treat one another as responsible agents, and on this basis, we praise one another for some actions and blame one another for others. The practice of assigning responsibility underlies both civil and criminal law. This raises the question of what the loci of responsibility are: Who or what are the entities capable of bearing responsibility? Are individual human beings the only such entities, or are groups, especially organized collectives, appropriate targets for the assignment of responsibility as well? Given the influential roles that many collectives play in the social world, from states and corporations to courts and committees, this question is of great importance for morality and politics as well as for the law. The aim of this chapter is to describe and defend the view that certain organized collectives—​namely, group agents—​can be held responsible for their actions, over and above their individual members, and that group responsibility is not reducible to individual responsibility.1 This view has some important implications. There are some things that happen in the social world, including some harms, that cannot easily be attributed to any single individual or even any number of individuals. If individual human beings were the only entities capable of bearing responsibility, then we might not be able to identify anyone who bears sufficient responsibility in such cases. We might be faced with “responsibility gaps” or “responsibility voids”: situations where some significant harm has occurred, but there is a shortfall of attributable responsibility.2 If, alternatively, there are responsible group agents to whom the actions in question can be attributed, we may be able to avoid such responsibility gaps. And indeed, many jurisdictions recognize some forms of corporate civil and even criminal liability, holding certain collective 1 

The literature on the responsibility of collectives is large, and I cannot provide a comprehensive review here. See, among others, Feinberg (1968), Held (1970), French (1984), May (1987), Erskine (2001), Copp (2006), Pettit (2007), List and Pettit (2011, ch. 7), Pasternak (2013), and Tollefsen (2015). For an overview, see Smiley (2017). 2  See, e.g., Braham and van Hees (2011, 2018), Collins (2017, 2018), Duijf (2018a), and Buell (2018).

132   Christian List entities responsible over and above their individual members. If the notion of group responsibility can be defended, then this legal practice stands on a firm footing. The chapter is structured as follows. In section 1, I will introduce the problem of group responsibility in more detail and put this chapter’s question into sharper focus. In section 2, I will say more about the conditions an entity must meet in order to count as “fit to be held responsible.” While adult human beings are normally assumed to meet those conditions, it is much less clear whether collectives could do so too. In sections 3 and 4, I will introduce the idea of group agency and explore whether group agents can be fit to be held responsible. In section 5, I will look at the relationship between individual and group responsibility and outline the case for the irreducibility of group responsibility. In section 6, finally, I will ask whether the view that certain groups are responsible agents also commits us to the view that those groups should be given rights of their own, and I will give a qualified negative answer. An important disclaimer is due: My focus in this chapter is on the question of whether certain groups can be loci of responsible agency in their own right; it is not on what responsibilities individuals have in the context of their participation in collective actions or their membership in certain groups. I will touch on the latter question only in passing. In taking the present approach to the subject of group responsibility, I am building on my earlier work on group agency, especially my joint work with Philip Pettit.3 Others would undoubtedly approach the subject from different angles, but given space constraints, I cannot cover the subject in its entirety here.

1.  The Problem of Group Responsibility Our lives are profoundly affected by the activities of many collectives. These include not just firms and commercial corporations, but also states, other governmental and nongovernmental organizations, courts, healthcare providers, universities, expert panels, clubs, and many informal groups. All of these make a difference to people’s lives through their decisions and actions, and sometimes cause harms, either by accident or as more or less predictable externalities or side effects of what they do. Examples readily come to mind: from the Deepwater Horizon oil spill in the Gulf of Mexico, caused by BP’s drilling operations; through the environmental degradation caused by big industry; to the recent Windrush scandal in the United Kingdom, where the rights of many legal residents were infringed due to a change in the country’s immigration policies. As is widely acknowledged, when harms result from the actions of collective entities, it is

3  See List and Pettit (2011) and List (2018a,b). See further Pettit (2007), from whom I am borrowing the language of “fitness to be held responsible.” The present approach has many precursors in the literature; recall footnote 1. The chapter draws significantly on List (2018b), which, for the academic record, should be regarded as prior.

Group Responsibility   133 not always easy to identify one or several individuals to whom all the relevant responsibility can be attributed. It is possible that most, or even all, of the individual members of the collectives in question have acted in ways that were reasonable, given their circumstances, and are therefore individually blameless or only blameworthy to a small extent, and yet their individual contributions have jointly produced a very harmful outcome. In this way, the sum total of individual responsibility may fall short of the full amount of responsibility that we might intuitively find appropriate for the harm caused. There seems to be a shortfall of attributable responsibility: a partial or even complete “responsibility gap” or “responsibility void.”4 An illuminating example, given by Philip Pettit, is that of the Herald of Free Enterprise, a car ferry in the English Channel which sank on its way from Zeebrugge to Dover in 1987.5 Almost two hundred people died. An inquiry found that the operating company, Townsend Thoresen, was extremely sloppy and had very bad safety standards. As a commentator put it, “[f]‌rom top to bottom the body corporate was infected with the disease of sloppiness.”6 Nonetheless, strikingly, nobody was found legally responsible to an extent that seemed commensurate with the harm that had occurred. The same commentator noted: “the primary requirement of finding an individual who was liable . . . stood in the way of attaching any significance to the organizational sloppiness that had been found by the official enquiry.”7 Because the ferry company was not considered a locus of responsibility, there was no agent to whom the relevant responsibility could be assigned. The harm was treated, in effect, as if it had been the result of a natural disaster or bad brute luck, not the result of anyone’s wrongdoing. One view about responsibility, which we may call “individualism,” would vindicate this outcome. According to an individualist view about responsibility, only human beings, especially adult human beings, are appropriate targets for the assignment of responsibility; collectives are not. On such a view, it is appropriate to ask how various individuals have contributed to a harmful outcome and, if negligence, recklessness, or wrongdoing can be established, to hold them individually accountable. But, on the individualist view, the collective itself cannot be held responsible over and above its members. If, even after the most careful accounting, the sum total of individual responsibility falls short of the amount of responsibility that would seem commensurate with

4  Recall the references in footnote 2. Interestingly, responsibility gaps might be viewed as involving a violation of what is sometimes called “moral harmony.” In the words of Portmore (2018, p. 318), an account of morality is “harmonious” if “the agents who satisfy [it], whoever and however numerous they may be, are guaranteed to produce the morally best world that they have the option of producing.” Perhaps, then, responsibility gaps could occur only against the background of an insufficiently harmonious criterion of individually reasonable action. That said, the envisaged scenarios seem familiar enough from the perspective of common-​sense morality to be worth investigating. I am grateful to Hein Duijf for pressing me to address this point. 5  See Pettit (2007) and List and Pettit (2011, ch. 7). 6  See Colvin (1995), quoted in Pettit (2007). 7  Colvin (1995), quoted in Pettit (2007).

134   Christian List the harm done, then we would have to bite the bullet and accept that no further responsibility can be assigned. A competing view is “collectivism.” According to it, collectives may sometimes be held responsible in their own right, independently of the individuals involved. The key idea is that there can in principle be collective responsibility without any corresponding individual responsibility. In limiting cases, a collective entity could be held responsible for some outcome without any of its members bearing any individual responsibility. Of course, the details of when this is so, and what exactly it means, need to be spelled out further, and I will turn to them later. The collectivist view is sometimes motivated by the observation that unless we are prepared to assign responsibility to certain collectives in their own right, the occurrence of some undesirable responsibility gaps is inevitable in practice. However, from the premises that (1) on the individualist view, there may sometimes be responsibility gaps, and that (2) responsibility gaps are undesirable, we cannot derive the conclusion that (3) the collectivist view is correct. The argument from (1) and (2) to (3) would be a non-​sequitur. Premises (1) and (2) are insufficient to establish collectivism about responsibility. Perhaps it is simply an unfortunate fact about morality that there is sometimes a shortfall of individually attributable responsibility, and it is wishful thinking to expect that in every such case one could close the gap by identifying a responsible collective. Indeed, we can easily imagine cases of collectively caused harms for which it is implausible to hold anyone—​whether individuals or the collective—​responsible. Imagine, for instance, a freak accident in a town center, where, due to an unusual and unforeseeable confluence of pedestrian traffic, a stampede occurs and people get injured, but neither the pedestrians involved, nor the police, nor the local government did anything wrong. If this truly is a freak accident, then there is no responsibility to be assigned, given that there was nothing wrong with anyone’s conduct: no negligence or recklessness, let alone intentional wrongdoing. Furthermore, in this case, the relevant collective—​the various pedestrians—​seems far too amorphous to count as a plausible locus of responsibility.8 In fact, one line of reasoning would suggest that it is actually a misnomer to call this a responsibility gap. There simply is no responsibility here. It is no different from a natural disaster, as far as responsibility is concerned. We would not normally look for a

8 

On the distinction between organized and random collections, and on whether the latter can bear responsibility, see also Held (1970).

Group Responsibility   135 responsible agent in the case of a genuine natural disaster.9 Now, the proponents of an individualist view about responsibility might insist that the case of the Herald of Free Enterprise is similar. If, as we may stipulate for the sake of argument, there was no wrongdoing on the part of any individuals, then the ferry disaster might be like the stampede or a natural disaster. If we still want to argue that the ferry company should be held responsible as a collective entity, independently of the extent of any wrongdoing by relevant individuals, we need a more careful account of the conditions that an entity must meet in order to count as “fit to be held responsible.” Once we have specified those conditions, we are in a better position to adjudicate whether collectives could be bearers of responsibility in their own right.10

2.  The Conditions for Responsibility Adult human beings are paradigmatic examples of entities that we consider capable of bearing responsibility for their actions: They qualify as “fit to be held responsible,” as Philip Pettit puts it.11 Are adult human beings the only such entities, or could other entities also qualify? Obviously, we don’t want to answer this question just by stipulation. Rather, we would like to appeal to some general conditions for fitness to be held responsible. These will hopefully vindicate our assumptions in the paradigmatic human case, while helping us to clarify what is going on in other cases. For the purposes of this chapter, I will treat the following conditions as necessary and sufficient for fitness to be held responsible.12 9  I speak of “genuine” natural disasters here in order to set aside those disasters that have a plausible anthropogenic source, such as anthropogenic-​climate-​change-​induced disasters. The latter arguably do raise questions of responsibility. 10  Hein Duijf (in correspondence) has suggested the following more nuanced analysis of the (formally invalid) argument from (1) and (2) to (3). First, note that the claim that there may sometimes be responsibility gaps, as asserted by premise (1), implicitly relies on some background theory of responsibility. After all, it is claimed that individuals are not responsible, and yet that something has happened that would somehow warrant the attribution of responsibility. Now, if we further accept the claim that such responsibility gaps are undesirable, as asserted by premise (2), then we may well conclude that our background theory of responsibility is in some ways defective, and that we should revise it. At this point, several avenues are open to us. One would be to deny that there are any responsibility gaps. A second would be to revise the criteria for the attribution of individual responsibility. And a third would be to extend the loci of responsibility, by recognizing collectives as responsible agents. Finally, one might combine two or more of these avenues. Against the background of this analysis, the argument from (1) and (2) to (3) is still invalid, but premises (1) and (2) lend some support to the conclusion that something needs to be revised, even though accepting (3) is not the only possible way this revision could go. See further Duijf (2018b). 11  See Pettit (2007). 12  The conditions build on the ones in List and Pettit (2011, ch. 7), which, in turn, are variants of conditions in Pettit (2007).

136   Christian List Moral agency: The entity is an intentional agent and has the capacity to make normative judgments about its choices and to be guided by those judgments. Relevant information: The entity has the information needed for assessing its choices normatively, or it has at least reasonable access to such information. Free choice: The entity has the freedom and control required for making its choices.

To see that these conditions are indeed required for fitness to be held responsible, let us briefly go through them. The first condition—​moral agency—​is evidently necessary. Entities that do not qualify as intentional agents, such as rocks, armchairs, or hammers, do not even begin to qualify as candidates for bearing responsibility. But intentional agency alone is not enough. There are many entities, most notably nonhuman animals, that plausibly qualify as intentional agents without being moral agents. They exhibit systematic, goal-​directed behavior, but they lack the capacity to make judgments about what is right and wrong, permissible and impermissible, and to act on the basis of those judgments. And for this reason, we do not hold them responsible for anything. The second condition—​relevant information—​is equally necessary. If someone lacks the information—​or even reasonable access to the information—​required for a normative assessment of his or her choices, perhaps due to the constraints of his or her environment, then it would not be justified to hold him or her responsible (and certainly not fully responsible) for choices made in those circumstances. Finally, the third condition—​free choice—​captures the widely accepted idea that freedom and control, of some sort, are necessary for responsibility. If someone has no freedom or control in relation to his or her actions—​perhaps he or she lacks alternative possibilities or has no causal control over what he or she does—​then it would seem problematic to hold him or her fully responsible for those actions.13 Consistently with this, freedom-​or-​control-​ undermining constraints, whether external or psychological, are often thought to be responsibility undermining as well. While these considerations suggest that the three conditions are each necessary for fitness to be held responsible, I will also assume that they are jointly sufficient. It should be evident that it is very much a contingent question which entities meet these conditions. While adult human beings paradigmatically satisfy them—​or so we tend to believe—​we must not assume from the outset that some nonhuman entities 13  There are some complications here that I cannot go into, given space constraints. For instance, Frankfurt (1969) has argued that alternative possibilities are not needed for responsibility. I accept that one may debate the precise nature of the “free choice” requirement for responsibility, but I assume that at least some such suitably qualified requirement is needed. For further discussion, see also Alvarez (2009) and List (2019, pp. 168–​169).

Group Responsibility   137 could not do so as well. There is no conceptual reason, for instance, why future sophisticated robots could not also satisfy them; nor should we rule out the possibility that suitably organized collectives could do so. Of course, much depends on the details of specific cases. It would be implausible to suggest that the collection of pedestrians whose uncoordinated movements accidentally caused a stampede constitutes a responsible agent. Random collections are likely to fall short on each of the three conditions. But arguably, the satisfaction of the conditions is less implausible in the case of a suitably organized collective, such as a commercial corporation or a state. To see whether certain collectives can indeed meet them, we must turn to the idea of group agency.

3.  Group Agency A group agent, in brief, is an organized collective that constitutes a goal-​directed, intentional agent in its own right, over and above its individual members.14 Examples of group agents are firms and corporations, courts, churches, universities, governments, nongovernmental organizations, and even states. We ascribe beliefs and desires, and other intentional attitudes, to those entities and commonly treat them as agents. In fact, the law treats some of those entities as artificial persons with certain obligations and a legal status, distinct from the obligations and the legal status of their members. Now some people might object that the ascription of agency and attitudes to collectives is just metaphorical. In a much-​quoted passage, Anthony Quinton, for instance, writes: Groups are said to have beliefs, emotions and attitudes . . . But these ways of speaking are plainly metaphorical. To ascribe mental predicates to a group is always an indirect way of ascribing such predicates to its members . . . To say that the industrial working class is determined to resist anti-​trade union laws is to say that all or most industrial workers are so minded.”15

On Quinton’s view, any talk of group agency or group attitudes is just a shorthand for something that could be equally expressed by reference to the group’s members. However, as I will now explain, there are reasons for taking a realist view about group agency, according to which at least some groups can truly qualify as intentional agents. Perhaps Quinton is right that the industrial working class, like the group of pedestrians

14 

On group agency, see, e.g., French (1984), Rovane (1997), Pettit (2001, ch. 5; 2003), Tollefsen (2002, 2015), List and Pettit (2006, 2011), and Tuomela (2013). 15  See Quinton (1975, p. 17).

138   Christian List in my earlier example, is too amorphous to qualify as a genuine agent. But the situation is different in the case of suitably organized collectives, such as firms and corporations, as well as the other examples mentioned at the beginning of this section. Such collectives can be shown to meet the defining conditions for intentional agency, and they are in fact routinely treated as agents in the social sciences. Let me begin with the defining conditions for agency. One might be tempted to define group agency in some kind of sui generis way—​say, by describing how a collection of individuals must act together to qualify as a group agent—​but this approach would be rather ad hoc. It would leave open what it is that human beings, group agents, and perhaps other artificial agents (such as robots and AI systems) have in common. A more systematic approach is to introduce general conditions for intentional agency, not conditions tailor-​made for the corporate case, and then to ask whether certain groups could meet them. For present purposes, I will adopt a relatively simple definition. I will define an intentional agent as an entity, in some environment, that meets three conditions:16 Representation: The entity has representational states, which encode its “beliefs” about how things are. Motivation: The entity has motivational states, which encode its “goals” or “desires” as to how it would like things to be. Action: The entity has the capacity to interact with its environment in a way that systematically involves those states: the entity is responsive to information from its environment and “acts” in pursuit of its goals or desires in line with its beliefs.

Clearly, a great variety of entities can meet these conditions, from human beings and nonhuman animals to certain kinds of robots and AI systems. Each of these can be readily subsumed under the belief-​desire-​action structure described. Even a system as simple as a thermostat can be interpreted as meeting the conditions: It has representational states, representing the actual temperature of its environment; motivational states, representing a target temperature; and a capacity to interact with its environment, by measuring the temperature and regulating the heating so as to bring the actual temperature in line with the target. It is important to recognize that intentional agents lie on a spectrum, and that their capacities can range from very basic to extremely complex. Human beings, for instance, have a great variety of internal psychological states

16 

This definition, which expresses a version of the Humean belief-​desire conception of agency, draws on List and Pettit (2011, ch. 1) and List (2018a,b). On “belief-​desire-​intention agency,” see Bratman (1987).

Group Responsibility   139 that go beyond beliefs and desires, and their capacities give them moral agency, over and above mere intentional agency; more on this later. Given all this, the claim that certain organized collectives can meet the defining conditions for intentional agency should be fairly uncontroversial. Indeed, the theory of the firm in economics depicts firms and commercial corporations as belief-​desire agents, which rationally maximize expected profits in line with certain beliefs about their commercial environment.17 Likewise, the so-​called “realist” theory of international relations depicts states as belief-​desire agents, which rationally pursue their goals and objectives in line with their beliefs about the international world.18 In game-​theoretic models of the Cold War, for instance, the United States and the Soviet Union often featured as “rational players,” with certain strategic objectives and certain beliefs about how their opponent will behave. Their behavior was then explained in much the same way in which we would ordinarily explain the behavior of a rational individual, such as a consumer in the marketplace. Of course, those explanations are somewhat simplistic, but they illustrate that taking an “intentional stance” toward certain collectives—​explaining their behavior by viewing them as intentional agents—​is a reasonable explanatory strategy.19 A critic might say that all of this is consistent with interpreting the ascription of agency to the collectives in question as nothing more than an instrumentally useful move. Treating groups as if they were agents might be explanatorily useful, but that does not imply that group agency is a real phenomenon. There is, however, a fairly straightforward philosophy-​of-​science argument for realism about group agency:20 Premise 1: The ascription of intentional agency to certain collective entities is indispensable in (social-​)scientific explanations of their behavior. Premise 2: If the ascription of some property to an entity is indispensable in scientific explanations of that entity’s behavior, then we are—​at least prima facie—​warranted in assuming that the entity really has that property. (The warrant is only prima facie because it could still ultimately be defeated by other considerations.) Conclusion: We are—​at least prima facie—​warranted in assuming that the relevant collective entities really are intentional agents.

17 

See, e.g., Mas-​Colell, Whinston, and Green (1995). See Waltz (1979) and Snidal (1985). 19  On the “intentional stance,” see Dennett (1987). On taking that stance toward collectives, see also Tollefsen (2015). 20  For previous versions of this indispensability argument, see List (2018a,b). A similar argument is implicit in other works on group agency, including Tollefsen (2002) and List and Pettit (2011). 18 

140   Christian List The argument is certainly valid (the conclusion follows logically from the premises), and I will here assume that a case can be made for each premise. The argument is analogous to arguments for realism about other properties in the sciences. For instance, from the fact that the ascription of certain physical forces to certain entities is indispensable in physical explanations, we often derive a prima facie warrant for realism about those physical forces.21 The upshot is that, insofar as some of our best theories in the social sciences ascribe intentional agency to certain collectives, it is acceptable to treat group agency as a real phenomenon, at least prima facie. Before turning to the question of whether certain groups can also qualify as moral agents, I should distinguish group agency from two related phenomena: joint action and collective action. Joint action is the coordinated action of several individuals under a structure of shared or common intentions, for instance, when several individuals jointly implement a shared plan.22 Standard examples are carrying a piece of furniture downstairs together or going for a walk together. Collective action is the cooperative action of several individuals more generally, whether or not this is based on any shared or common intentions.23 Successfully cooperating in the provision of a public good, despite incentives to free ride, would be an example. Neither of these phenomena involves a group agent in its own right. To make sense of joint or collective action, we need not view the group itself as an intentional agent. The loci of agency are still the individuals.24

4.  Can Groups Be Fit to Be Held Responsible? We have seen that there is a case to be made for the claim that certain organized collectives qualify as intentional agents in their own right. This is so at least for those collectives for which the ascription of agency is explanatory indispensable, such as firms, corporations, and states, among others. By contrast, unorganized collectives such as the group of pedestrians in the example of the stampede would not warrant such an ascription. Now, if we restrict our attention to those groups that constitute intentional agents, the possibility that they might also qualify as moral agents becomes a live option. Although some of the most familiar cases of nonhuman agency, such as animal 21  This derivation implicitly relies on what is sometimes called “the naturalistic ontological attitude,” the view that science is our best guide to ontological questions—​questions about which entities and properties are real. See, e.g., Quine (1977) and Fine (1984). 22  On this phenomenon, see, e.g., Bratman (1999, 2014), Gilbert (1989), and Tuomela (2007). 23  On the problem of collective action, see Olson (1965) and Ostrom (1990). 24  Gilbert (1989) introduces the idea of a plural subject in her account of joint action. My view is that there are many cases of joint action where we do not need to postulate a group agent.

Group Responsibility   141 agency, fall short of moral agency, I have already pointed out that there is no conceptual reason why some nonhuman entities could not also qualify as moral agents. What distinguishes a moral agent from a mere intentional one is the capacity to make normative judgments about its choices—​judgments about what is right and wrong, permissible and impermissible—​and to be guided by those judgments. There is no reason why corporations and other organized collectives could not have institutional procedures and mechanisms that enable them to make such judgments and to let these judgments guide their actions. Whether an organized collective is like this depends entirely on how it is designed and structured.25 What about the other conditions for fitness to be held responsible? Recall that moral agency alone is not enough. The entity in question must also have access to the information needed for a normative assessment of its choices, and it must have the freedom and control required for making those choices. The access-​to-​information condition is relatively unproblematic. Just as individuals with the right cognitive capacities are capable of accessing and processing relevant information about their choices, provided their informational environment is not too constrained, so too are suitably organized group agents. For instance, a commercial corporation is no less well placed to access and process information about the likely consequences and negative externalities of its actions than any individual human being would be. The condition of free choice is trickier. Can group agents genuinely have freedom and control in making their choices? I cannot do justice to this question here, but I want to point out that, at least given the possibility of intentional group agency, a positive answer is not too far-​fetched. I take freedom and control to require, at a minimum, two things. First, the relevant agent must be able to choose between alternative possibilities. And second, when the agent makes a choice, the agent’s intentional state must be the difference maker of that choice: Whatever the agent does must not be some sub-​ agential fluke (such as an uncontrollable reflex or spasm). Can group agents meet these requirements?26 As far as the requirement of alternative possibilities is concerned, I see no reason why they could not. Group agents can certainly be faced with choices between alternative courses of action. Their choices need not be determined in a way that would have made it impossible for them to act any differently from how they did. When a corporation makes a risky decision that leads to some disaster, it may be perfectly reasonable to say that this corporation could have acted otherwise: The alternative, less risky course of action may have been entirely open to the corporation; perhaps its executive board even considered it but dismissed it. Second, group agents can also be said to have control over what they do. Just as—​in the case of individual agency—​the difference-​making cause of an individual’s action is 25  For recent works on corporate moral agency, see, e.g., Björnsson and Hess (2017) and Pasternak (2017). 26  I have previously discussed these questions in List and Pettit (2011, ch. 7) and in List (2019).

142   Christian List arguably that individual’s intentional state, so the difference-​making cause of a corporate action can plausibly be the group agent’s intentional state, for instance, as generated through its corporate decision-​making structure. Although the claim that group agents exercise causal control over their actions is vulnerable to the same kinds of philosophical challenges that arise for mental causation in the case of individuals, the available philosophical responses to those challenges in the individual case arguably carry over to the corporate one.27 I cannot do more than flag these issues here. At any rate, I want to note that there is, by now, a large philosophical literature that assumes that these obstacles can be overcome and that recognizes the possibility of corporate moral responsibility. Peter French’s work on corporate responsibility is a locus classicus.28 In sum, there are theoretical resources available for arguing that some group agents are fit to be held responsible. The recognition of corporate responsibility, in turn, may be combined with an appropriate penal code at the corporate level. The punishment for corporate wrongdoing may include, for instance, fines and other sanctions targeting the assets of the relevant corporate entities, as well as restrictions on their operating permissions. In the same way in which the punishment for an individual’s criminal wrongdoing may be the (perhaps temporary) removal of certain liberties, so the punishment for a corporate crime may be a certain kind of operating restriction or ban for the group agent. The present considerations suggest that society might protect itself against the risk of corporate responsibility gaps by legally requiring that group agents operating in high-​ stakes settings should be organized so as to meet the conditions for responsible moral agency. According to this proposal, for a powerful group agent to receive its corporate charter or operating license, it would have to demonstrate that it has procedures in place to ensure its fitness to be held responsible for its actions.29 In the same way in which universities and other organizations receive their accreditation only if they comply with certain legal requirements, so powerful group agents in general would be allowed to operate only if they are compliant with the conditions for fitness to be held responsible. The details of how this could work are complicated, but the present proposal would offer one way to avoid the kind of situation that arose in the example of the ferry disaster. Indeed, we can now see that if the ferry company truly fell short of the conditions for fitness to be held responsible, then—​according to the proposal just sketched—​the company should never have been permitted to run a ferry service.

27 

My reasoning relies on the account of mental causation in List and Menzies (2009). The solution to the corporate problem of mental causation suggested in List and Pettit (2011, ch. 7) is subtly different. On agential control as difference making, see also Himmelreich (2015). 28  See French (1984) and the other references in footnotes 1 and 2. 29  For a related proposal, involving a “developmental rationale” for holding group agents responsible, see List and Pettit (2011, ch. 7).

Group Responsibility   143

5.  How Does Group Responsibility Relate to Individual Responsibility? The claim that there can be group responsibility over and above individual responsibility—​and sometimes even in the absence of individual responsibility—​invites a natural objection. Group agents are collections of individuals, and they cannot act without their members’ contribution. Anything a group agent does depends on what its members do, under the group’s organizational structure. Insofar as those members are themselves fit to be held responsible, shouldn’t group responsibility somehow translate into member responsibility? In slightly more metaphysical terms, since all group-​level facts (facts about what a group agent does) presumably “supervene” on individual-​level facts (facts about what the members do), so the objection goes, it must be possible to “reduce” group-​level responsibility to individual-​level responsibility.30 The response is that this objection rests on a misconception. It overlooks the fact that, in real cases of group agency, unlike in cases of mere joint or collective action, the group itself is a locus of agency over and above its members. This means that the agent of the group’s actions is the group itself, understood in “corporate” terms, not its individual members, understood in “distributive” terms. While the group’s actions are causally and even ontologically dependent on the members’ contributions, it does not follow that those actions can also be agentially attributed to them. In a case of genuine group agency, the actions of the group agent are not correctly classified as actions of its members. For a structural analogy, consider the relationship between the actions of an individual person and the biological processes within the person’s organism. Even if we accept the standard view that the person’s actions are causally and ontologically dependent on the underlying bodily processes, it would be a mistake to treat any of the constituent body parts—​say, the visual cortex, the cerebellum, or the muscles—​as “agentially responsible” for those actions. The visual cortex, the cerebellum, and the muscles all contribute to the person’s performance of his or her actions, but these body parts do not qualify as agents of those actions. The locus of agency is the person as a whole, not any components of the underlying organism. Of course, in this example, the relevant body parts do not qualify as intentional agents themselves, and so it would never occur to us to ascribe any agential responsibility to them. But even if the function of one of these body parts—​say that of the cerebellum—​were performed by a hypothetical system that did qualify as an agent, it would still not follow that the locus of agency for the person’s overall actions was anything other than the person as a whole. Similarly, if we take the notion of group agency seriously, we must distinguish between the actions of a group 30  This objection is obviously motivated by a broader background assumption of methodological individualism. For a discussion of methodological individualism, see, e.g., Heath (2015).

144   Christian List agent and the actions of its members, and we must not confuse the former with the latter. Just as an individual person’s actions are “located” at the systemic level of the person as a whole, not the level of any body parts, so the actions of a group agent are “located” at the systemic level of the group as a whole, not the level of its individual members. A version of this point was already recognized by Virginia Held in a 1970 article, in which she discussed the crucial difference between organized groups and mere random collections: [I]‌f organized group G is morally responsible for the failure to do A it does not follow that member M of G is morally responsible for the failure to do A. If a random collection R can be represented as a set equivalent, say, to M & N & Q, then, if R is morally responsible, we would seem to be able to conclude that M is morally responsible & N is morally responsible & Q is morally responsible. On the other hand, if these same members formed an organized group, the group could not be adequately represented as equivalent simply to M & N & Q, because its depiction would have to include the decision method by which the members act as a group. And the distribution of moral responsibility over such a combination would not seem plausible.31

That said, depending on the details of the case, some individuals may bear individual responsibility for the roles they have played in relation to the group agent’s conduct. They may do so to the extent that they have knowingly and voluntarily played some of the following roles: • enacting roles, for example as managers, officials, or representatives of the group agent; • authorizing roles, for example as directors, board members, owners, share-​holders, or even regulators; or • organizational-​design roles, for example as founders, policy makers, or institutional designers.32 How exactly the occupants of these roles are responsible for what the group agent does is a contingent matter that may differ from case to case. But as I have already emphasized, there is no reason to assume that the sum total of their individual responsibility will always match the responsibility of the group agent as a whole. In short, recognizing individual responsibility in the context of a group agent is no substitute for recognizing group responsibility. It is worth acknowledging two further complications that arise when we try to “distribute” group responsibility to the underlying individuals. The first is this. Suppose we took the view that individuals are responsible for the difference they have knowingly and willingly made to the group’s actions: If they have made a big difference, they would

31 

32 

See Held (1970, p. 480). Cf. List and Pettit (2011, ch. 7).

Group Responsibility   145 bear significant responsibility, whereas if they have made only a small difference or no difference, they would bear very little responsibility or none. The problem with this view is that, even if individuals have knowingly and willingly contributed to the group’s actions, and have wholeheartedly endorsed them, they might still come out as bearing no responsibility at all. We can easily imagine cases in which the group collectively takes a momentous decision, and yet no one is individually pivotal. The easiest illustration is given by a democratically organized group that makes its decisions by majority rule. If a decision is reached by a margin of victory of more than one or two votes, then no individual voter is pivotal: No one can say that if he or she had voted differently, this would have changed the outcome (by flipping the majority). In such a case, no individual qualifies as a difference maker of the collective outcome. The lesson is that, even in cases in which it seems appropriate to assign responsibility to the group members on the basis of their knowing and willing support for the group’s decision, a simple difference-​making criterion would not vindicate this assignment of responsibility. Rather, a more subtle criterion would be needed. One example would be the so-​called “INUS” or “NESS” criterion. According to it, a minimal condition for an individual’s action to count as a significant contributor to a collective outcome would be that this individual’s contribution (which we here assume is made voluntarily and knowingly) is “an insufficient but necessary part of a condition which is itself unnecessary but sufficient for the result” or, more generally, “a necessary element of a [causally] sufficient set.”33 Of course, more could be said about such a criterion.34 The second complication is the following. Suppose we thought that at least in democratically organized groups, under background conditions of freedom and voluntariness, it should not be difficult to “distribute” group responsibility to the individual members. The reason is that, for any democratic group decision, we would expect that some individuals (normally a majority) will have voted in support of that decision. Those individuals could then qualify as bearing some responsibility, regardless of whether they were individually pivotal. In particular, we might take the view that a reasonable condition for bearing some individual responsibility for a collective decision is having voluntarily and knowingly voted in support of it. The problem with this view, however, is that it can still lead to responsibility gaps. A much-​cited example is that of a collegial court making a decision in a breach-​of-​contract case.35 Consider the following scenario. A three-​member court has to decide whether a defendant is liable for breach of contract. The relevant legal doctrine requires the court to make this decision on the basis of three background questions: First, did the defendant do the action in question; second, did the action go against the relevant contract; and third, was the contract valid in the first place? The defendant is liable if and only if the answer to all three questions is “yes.” If, conversely, the defendant did not do the action, or the action was permitted 33 

See Mackie (1965) and Wright (1985), respectively. See, e.g., Braham and van Hees (2012). 35  The example goes back to Kornhauser and Sager (1986). For more recent discussions, see List and Pettit (2011). 34 

146   Christian List Table 1 A Three-​Member Court Action done?

Prohibited by the contract? Contract valid? Liable?

Judge 1

Yes

Yes

No

No

Judge 2

Yes

No

Yes

No

Judge 3

No

Yes

Yes

No

Majority

Yes

Yes

Yes

No

by the contract, or the contract was invalid, then there is no liability. Now we can easily imagine a situation in which each member of the court gives a positive answer to two of the three questions, but not to the third, with different judges disagreeing about which question is to be answered in the negative. The pattern of individual judgments is shown in Table 1. In this case, the majority answer to all three questions is “yes,” and yet no individual judge thinks that the defendant is liable. If—​hypothetically—​the court were to use the so-​called “premise-​based procedure,” it would reach its overall verdict as follows. It would take majority votes on the three premises for liability (was the action done, was it against the contract, was the contract valid), and it would then derive its overall verdict on the case from its collective judgments on those premises. Under this arrangement, there would be no need to take a vote on the issue of liability itself; the verdict would be derived via a form of “collectivized reasoning,” from democratically agreed premises. Given the individual judgments shown in Table 1, the premise-​based procedure would support a “liable” verdict, despite the unanimous individual view that the defendant is not liable. This example illustrates that a collective decision may have been reached through a perfectly reasonable democratic procedure, such as the premise-​based procedure, and yet no individual may have personally supported the ultimate outcome—​here a “liable” verdict. If having voted for the outcome were a necessary condition for bearing some responsibility for it, then none of the individual judges would bear any responsibility in the present case.36 Of course, one might object that the premise-​based procedure is not the best decision procedure here, but the example generalizes. It can be shown that any democratically organized group that seeks to arrive at consistent collective judgments on several logically related issues cannot generally follow the majority judgment on every issue.37 Issue-​by-​issue majority voting would run the risk of producing an inconsistent pattern

36 

The analysis of this case would be more complicated if one took the view that, in cases such as that of the court, voting for an outcome is not a necessary condition for bearing some responsibility for it. One might argue, for instance, that strategically abstaining from voting on a controversial matter does not absolve one of responsibility. 37  For a review of relevant results and further references, see List and Pettit (2011, ch. 2).

Group Responsibility   147 of judgments, as illustrated by Table 1. If the group is to function as an instrumentally rational group agent, with consistent collective views, then it must sometimes (in some logically possible situations) deviate from the views of a majority of its members. And so, we cannot presume that every group decision will be supported by a majority of group members, to whom the relevant responsibility can then be assigned. Although the details are complicated, I hope the basic point is clear enough. For a further analysis of the present problem, I refer the reader to the literature on judgment aggregation.38 The bottom line is that there is no general scheme available by which group responsibility can always be distributed to the group’s members: Group responsibility is not generally reducible to individual responsibility.

6.  Can Groups Also Have Rights of Their Own? I have described and defended the view that certain collective entities can bear responsibility in their own right, over and above their individual members. In closing, I would like to ask whether accepting this view commits us to accepting that those groups should be given rights as well. If groups can be duty-​bearers, can they also be right-​holders? I will argue that, in one sense, the answer is quite unproblematically positive, but in another, it is clearly negative. To distinguish those senses, I must begin with two clarifications about the notion of a “right.” First of all, the term “right” can refer either to a legal right or to a moral one. Legal rights are those rights that agents have under the legal systems they are subject to. For instance, I have a legal right to vote in my country of citizenship, but I don’t have a legal right to vote in the country I visit as a tourist. Moral rights, by contrast, are rights that agents have from the perspective of morality. We arguably all have the moral right to certain basic liberties and respectful treatment, whether or not we are lucky enough to live in a society in which those rights are legally entrenched. The second clarification is this. When we say that an agent has a particular right, where this entails certain powers or permissions and perhaps duties on the part of others, then the agent might have that right either non-​derivatively or derivatively. A non-​derivative right is one the agent has by virtue of some underlying moral status. The right might be grounded, for instance, in the fact that the agent matters intrinsically: He or she is an object of moral concern. The natural rights that human beings are thought to have by virtue of their humanity, or by virtue of their status as moral persons, would be examples of non-​derivative rights. A derivative right, by contrast, is one the agent doesn’t have by virtue of any underlying moral status; rather, the right derives from something else. A derivative right might be grounded in its instrumental usefulness for

38 

For a review and further references, see again List and Pettit (2011).

148   Christian List promoting something else, which matters intrinsically. For example, a medical doctor might have certain derivative rights in his or her professional role, where those rights are justified by virtue of the fact that they are part of a normative system that serves the interests of society as a whole. Given these clarifications, I can now explain in what sense group agents can have rights, and in what sense they cannot. My claim is that organized collectives can have derivative rights (in principle, moral and legal), but they cannot have any non-​derivative ones. Let me begin with the positive part of this claim. Consider the functions that collective entities perform in the social world. The state provides all sorts of services; universities educate students and contribute to our knowledge of the world; medical institutions look after people’s health; and firms and financial organizations—​ notwithstanding their downsides—​are central to the economy. It is hard to deny that, given the way society is organized, those entities could not perform those functions unless they had certain rights, such as the right to own or rent property, the right to employ staff, the right to enter contracts and to engage in litigation when necessary, and so on. In recognition of these rights, the law treats the relevant collectives as legal persons and thereby considers them fit to be right-​holders and duty-​bearers. Of course, the rights in question are all derivative. They are not grounded in any special moral status of those collectives, but they are justified simply by virtue of the fact that, by and large, the normative system of which they are a part serves society’s interests. To be sure, there may be plenty of scope for improving that system. The philosophical point is just that, to the extent that giving certain rights to some collectives promotes the interests of the human members of society, group agents can unproblematically have those rights. That is the sense in which organized collectives can have derivative rights. Let me now turn to the negative part of my claim. I want to suggest that, unlike individual human beings, organized collectives cannot have any non-​derivative rights. The reason is that collective entities lack the relevant moral status: They do not matter intrinsically. Their significance is only instrumental. In this respect, group agents differ fundamentally from individual human beings, and arguably also from some nonhuman animals such as the great apes and other mammals, all of whom—​we may reasonably assume—​matter intrinsically. But while the differential categorization of group agents, as compared to individual humans and even animals, may seem intuitive, we must show that it is philosophically defensible.39 At first, one might think that if organized collectives not only meet the conditions for intentional agency, but also qualify as full-​blown moral agents, then it is hard to deny them the kinds of rights that individual moral agents have. If we accept collectivism with regard to responsibility, how can we defend individualism when it comes to rights? Wouldn’t a robust defense of group agency vindicate, for instance, the US Supreme Court’s decision in Citizens United v. 39  On the debate about corporate rights and the moral status of collectives, see also Pasternak (2017) and Silver (2018).

Group Responsibility   149 Federal Election Commission to extend certain free speech rights to corporations and other organizations, where those rights can then be exercised via corporate activities and monetary spending? To show that only individual human beings, and perhaps certain nonhuman animals, but not group agents can have non-​derivative rights, we must show that the latter lack the kind of moral status that would ground such rights. To do so, we must appeal to an appropriate background criterion as to what the basis of an entity’s moral status is and show that group agents fail to meet that criterion. There are a number of possible criteria on offer, which approach the present issue in different ways.40 One criterion for being an object of moral concern might be membership of the human species. This criterion is clearly satisfied by individual humans and violated by groups. But the criterion is somewhat ad hoc and has the unfortunate consequence of “throw[ing] animals under the bus,” as Will Kymlicka puts it.41 A less restrictive criterion for being an object of moral concern would be the possession of intentional agency. This criterion is satisfied not only by human beings, but also by many nonhuman animals as well as by group agents. Even robots and AI systems could easily satisfy it. The worry here is that the criterion is too permissive. On this criterion, we would have to extend our moral concern to entities ranging from British Petroleum to self-​driving cars. A related but more restrictive criterion would be the possession of moral, not just intentional, agency. But, as we have already seen, group agents can satisfy that criterion, whereas, for instance, the great apes, who arguably matter intrinsically, would lose out. Can we come up with a better criterion? I think a plausible necessary (though perhaps not sufficient) condition for being an object of moral concern is the possession of phenomenal consciousness or, more broadly, the potential for phenomenal consciousness. Having phenomenal consciousness means being an entity that is a locus of subjective experiences: There is something it is like to be that entity, in Thomas Nagel’s words.42 This criterion would support our assumption that human beings are objects of moral concern and can therefore have non-​derivative rights,43 and it would also allow us to make a similar case for animals. The claim that some nonhuman animals have subjective experiences, including experiences of pleasure and pain, is fairly uncontroversial. So, what about group agents? There are strong reasons to believe that group agents do not have anything like phenomenal consciousness or the potential for it. There is nothing it is like to be a group 40 

On the grounds of moral status, see Jaworska and Tannenbaum (2018). See Kymlicka (2017). 42  See Nagel (1974). I have previously discussed this criterion for having non-​derivative rights in List (2018a,b). On phenomenal consciousness, see also Chalmers (1996). 43  There are some complications here, which I cannot go into. One might ask whether babies or comatose patients would satisfy the present criterion. I think the answer is positive, provided we adopt a sufficiently broad interpretation of “potential” for phenomenal consciousness. 41 

150   Christian List agent.44 At most, there is something it is like to be a member of a group agent; but that is because the members are loci of subjective experiences as individuals; the group as a whole isn’t. In particular, group agents lack the kinds of properties which, according to contemporary neuroscience, are associated with phenomenal consciousness.45 Although I cannot go into the details here, phenomenal consciousness appears to occur only in information-​processing systems in which there are massive internal high-​ bandwidth feedback mechanisms, as exemplified by the human cortex. Arguably, this sort of cortex-​like informational integration is absent from organized collectives, even if we think of them as socially implemented intelligent systems.46 For there to be a phenomenally conscious group agent, we would have to imagine the kind of scenario proposed by Ned Block in his “China brain” thought experiment.47 In this scenario, each member of a very sizable population (say, that of China) is tasked with playing the role of one neuron in a biological brain, and the overall neural network is then established via internet connections. Conceivably, this kind of setup might replicate cortex-​like informational integration in a social structure, so that the resulting group agent might be a candidate for phenomenal consciousness. But this is evidently a science-​fiction scenario, and existing group agents, or plausible future ones, are not like this at all. In short, if the present empirical claims about the physical correlates of phenomenal consciousness are correct, and phenomenal consciousness (or the potential for it) is indeed a necessary condition for being an object of moral concern, then clearly group agents do not meet that condition, and consequently they cannot have any non-​ derivative rights. The bottom line, then, is that collective entities may unproblematically have certain derivative rights as well as legal personhood, but no non-​derivative rights or a full-​blown moral status. Furthermore, the criterion for adjudicating whether to give any derivative rights to collective entities is whether this serves the interests of those who do qualify as objects of moral concern, such as human beings and perhaps nonhuman animals. In the case of Citizens United, the only way in which extending free speech rights to corporations could be justified would be by showing that this does indeed serve the interests of human society—​a claim that is at least contentious. A mere appeal to the legal personhood of corporations would be insufficient as a basis for extending non-​derivative rights to groups. In conclusion, there is no contradiction in recognizing certain collective entities as moral agents who are fit to be held responsible while at the same time denying that they have the kinds of rights that individuals have. Collectivism about responsibility is perfectly compatible with individualism about rights.

44 

See List (2018a); cf. Schwitzgebel (2015). This claim is supported, for instance, by integrated information theory (Tononi and Koch 2015), but other theories of the physical correlates of consciousness (such as the neural synchronization theory of Crick and Koch 1990) would support it too. 46  See List (2018a), but see Schwitzgebel (2015). 47  See Block (1980). 45 

Group Responsibility   151

References Alvarez, M. 2009. Actions, Thought-​Experiments and the “Principle of Alternate Possibilities.” Australasian Journal of Philosophy 87(1): 61–​81. Björnsson, G., and K. Hess. 2017. Corporate Crocodile Tears? On the Reactive Attitudes of Corporate Agents. Philosophy and Phenomenological Research 94(2): 273–​298. Block, N. 1980. Troubles with Functionalism? In N. Block, ed., Readings in Philosophy of Psychology, Vol. 1: 268–​306. London: Methuen. Braham, M., and M. van Hees. 2011. Responsibility Voids. The Philosophical Quarterly 61(242): 6–​15. Braham, M., and M. van Hees. 2012. An Anatomy of Moral Responsibility. Mind 121(483): 601–​634. Braham, M., and M. van Hees. 2018. Voids or Fragmentation: Moral Responsibility for Collective Outcomes. The Economic Journal Volume 128(612): F95–​F113. Bratman, M. E. 1987. Intention, Plans, and Practical Reason. Cambridge, MA: Harvard University Press. Bratman, M. E. 1999. Faces of Intention: Selected Essays on Intention and Agency. Cambridge: Cambridge University Press. Bratman, M. E. 2014. Shared Agency: A Planning Theory of Acting Together. Oxford: Oxford University Press. Buell, S. W. 2018. The Responsibility Gap in Corporate Crime. Criminal Law and Philosophy 12(3): 471–​491. Chalmers, D. J. 1996. The Conscious Mind. New York: Oxford University Press. Collins, S. 2017. Filling Collective Duty Gaps. Journal of Philosophy 114(11): 573–​591. Collins, S. 2019. Collective Responsibility Gaps. Journal of Business Ethics. 154 (4):943–​954. Colvin, E. 1995. Corporate Personality and Criminal Liability. Criminal Law Forum 6: 3–​44. Copp, D. 2006. On the Agency of Certain Collective Entities: An Argument from “Normative Autonomy.” Midwest Studies in Philosophy 30(1): 194–​221. Crick, F., and C. Koch. 1990. Towards a Neurobiological Theory of Consciousness. Seminars in the Neurosciences 2: 263–​275. Dennett, D. 1987. The Intentional Stance. Cambridge, MA: MIT Press. Duijf, H. 2018a. Responsibility Voids and Cooperation. Philosophy of the Social Sciences 48(4): 434–​460. Duijf, H. 2018b. Let’s Do It! Collective Responsibility, Joint Action, and Participation. PhD thesis, University of Utrecht . Erskine, T. 2001. Assigning Responsibilities to Institutional Moral Agents: The Case of States and Quasi-​States. Ethics & International Affairs 15(2): 67–​85. Feinberg, J. 1968. Collective Responsibility. Journal of Philosophy 65(21): 674–​688. Fine, A. 1984. The Natural Ontological Attitude. In J. Leplin, ed., Scientific Realism, 83–​107. Berkeley: University of California Press. Frankfurt, H. 1969. Alternate Possibilities and Moral Responsibility. Journal of Philosophy 66(23): 829–​839. French, P. A. 1984. Collective and Corporate Responsibility. New York: Columbia University Press. Gilbert, M. 1989. On Social Facts. New York: Routledge. Heath, J. 2015. Methodological Individualism. In E. Zalta, ed., The Stanford Encyclopedia of Philosophy (Spring 2015 Edition) .

152   Christian List Held, V. 1970. Can a Random Collection of Individuals Be Morally Responsible? Journal of Philosophy 67(14): 471–​481. Himmelreich, J. 2015. Agency as Difference-​Making: Causal Foundations of Moral Responsibility. PhD thesis, London School of Economics . Jaworska, A., and J. Tannenbaum. 2018. The Grounds of Moral Status. In E. Zalta, ed., The Stanford Encyclopedia of Philosophy (Spring 2018 Edition) . Kornhauser, L. A., and L. G. Sager. 1986. Unpacking the Court. Yale Law Journal, 96(1): 82–​117. Kymlicka, W. 2018. Human Rights Without Human Supremacism. Canadian Journal of Philosophy. 48 (6):763–​779. List, C. 2018a. What Is It Like to Be a Group Agent? Noûs 52(2): 295–​319. List, C. 2018b. Group Agency and Artificial Intelligence. Working paper, LSE. List, C. 2019. Why Free Will is Real. Cambridge, MA: Harvard University Press. List, C., and P. Menzies. 2009. Non-​Reductive Physicalism and the Limits of the Exclusion Principle. Journal of Philosophy 106(9): 475–​502. List, C., and P. Pettit. 2006. Group Agency and Supervenience. Southern Journal of Philosophy 44(S1): 85–​105. List, C., and P. Pettit. 2011. Group Agency: The Possibility, Design, and Status of Corporate Agents. Oxford: Oxford University Press. Mackie, J. L. 1965. Causes and Conditions. American Philosophical Quarterly 2(4): 245–​264. Mas-​Colell, A., M. D. Whinston, and J. R. Green. 1995. Microeconomic Theory. Oxford: Oxford University Press. May, L. 1987. The Morality of Groups: Collective Responsibility, Group-​Based Harm, and Corporate Rights. Notre Dame, IN: University of Notre Dame Press. Nagel, T. 1974. What Is It Like to Be a Bat? Philosophical Review 83(4): 435–​450. Olson, M. 1965. The Logic of Collective Action: Public Goods and the Theory of Groups. Cambridge, MA: Harvard University Press. Ostrom, E. 1990. Governing the Commons: The Evolution of Institutions for Collective Action. Cambridge: Cambridge University Press. Pasternak, A. 2013. Limiting States’ Corporate Responsibility. Journal of Political Philosophy 21(4): 361–​381. Pasternak, A. 2017. From Corporate Moral Agency to Corporate Moral Rights. The Law & Ethics of Human Rights 11(1): 135–​159. Pettit, P. 2001. A Theory of Freedom: From the Psychology to the Politics of Agency. Cambridge and New York: Polity and Oxford University Press. Pettit, P. 2003. Groups with Minds of their Own. In F. Schmitt, ed., Socializing Metaphysics: 167–​193. New York: Rowan and Littlefield. Pettit, P. 2007. Responsibility Incorporated. Ethics 117(2): 171–​201. Portmore, D. W. 2018. Maximalism and Moral Harmony. Philosophy and Phenomenological Research 96(2): 318–​341. Quine, W. V. 1977. Ontological Relativity and Other Essays. New York: Columbia University Press. Quinton, A. 1976. The Presidential Address: Social Objects. Proceedings of the Aristotelian Society 76 (1):1–​28. Rovane, C. 1997. The Bounds of Agency: An Essay in Revisionary Metaphysics. Princeton, NJ: Princeton University Press.

Group Responsibility   153 Schwitzgebel, E. 2015. If Materialism Is True, the United States Is Probably Conscious. Philosophical Studies 172(7): 1697–​1721. Silver, K. 2018. Can a Corporation Be Worthy of Moral Consideration? Journal of Business Ethics. 159 (1):253–​265. Smiley, M. 2017. Collective Responsibility. In E. Zalta, ed., The Stanford Encyclopedia of Philosophy (Summer 2017 Edition) . Snidal, D. 1985. The Game Theory of International Politics. World Politics 38(1): 25–​57. Tollefsen, D. P. 2002. Collective Intentionality and the Social Sciences. Philosophy of the Social Sciences 32(1): 25–​50. Tollefsen, D. P. 2015. Groups as Agents. Cambridge: Polity Press. Tononi, G., and C. Koch. 2015. Consciousness: Here, There and Everywhere? Philosophical Transactions of the Royal Society B 370 (1668):20140167. Tuomela, R. 2007. The Philosophy of Sociality: The Shared Point of View. New York: Oxford University Press. Tuomela, R. 2013. Social Ontology: Collective Intentionality and Group Agents. Oxford: Oxford University Press. Waltz, K. 1979. Theory of International Politics. Long Grove, IL: Waveland Press. Wright, R. W. 1985. Causation in Tort Law. California Law Review 73(6): 1735–​1825.

PA RT I V

DE T E R M I N I SM A N D T H E A B I L I T Y TO D O OT H E RW I SE

CHAPTER 8

Moral Resp onsi bi l i t y, Alternat i v e P ossibiliti e s , a nd F r ankfu rt E x a mpl e s Derk Pereboom

1.  Introduction The intuition that moral responsibility for an action requires that the agent could have done otherwise is natural and powerful. If an agent is to be blameworthy for an action it seems crucial that she could have acted so as to avoid being blameworthy. But in an article published in 1969, Harry Frankfurt, using an intriguing type of example, argued that an agent’s moral responsibility for an action does not require that she could have done otherwise. Since then, there has been a heated and voluminous debate about this issue. The intuition that moral responsibility requires that the agent could have done otherwise is aptly expressed in what David Widerker has called the W-​Defense. About Jones, who breaks a promise and is allegedly blameworthy but could not have done otherwise, he writes: [S]‌ince you, Frankfurt, wish to hold [Jones] blameworthy for his decision to break his promise, tell me what, in your opinion, should he have done instead? Now, you cannot claim that he should not have decided to break the promise, since this was something that was not in his power to do. Hence, I do not see how you can hold Jones blameworthy for his decision to break the promise. (Widerker 2000: 191)

158   Derk Pereboom As Carlos Moya puts it, to be blameworthy for an action an agent must have had an exempting alternative possibility, one that, should she have availed herself of it, would have exempted her from legitimate blame (Moya 2006: 67; cf. Otsuka 1998; Pereboom 2001). To be praiseworthy for an action, it must be that the agent could have done something less admirable—​although this is more controversial. Susan Wolf (1980, 1990) and Dana Nelkin (2008, 2011) accept the alternative-​possibilities requirement for blameworthiness but not for praiseworthiness. On Wolf ’s view, if an agent does the right thing for the right reasons, and because she is so committed to doing what’s right she could not have acted less admirably, she is manifestly still praiseworthy. Traditionally, participants in the free will debate have generally affirmed the principle of alternative possibilities: (PAP) An action is free in the sense required for moral responsibility only if the agent could have done otherwise than she actually did.

Frankfurt’s challenge to PAP features an example in which an agent considers performing some action, but a someone else, a neuroscientist, say, is concerned that he will not perform the action after all. So if the agent were to manifest an indication that he will not or might not perform the action, the neuroscientist would intervene and induce him to perform that action. But as things actually go, the neuroscientist remains idle, because the agent performs the action on her own. Here is Carolina Sartorio’s version of Frankfurt’s original example (Sartorio 2016): A neuroscientist, Black, wants Jones to perform a certain action. Black is prepared to go to considerable lengths to get his way, but he prefers to avoid showing his hand unnecessarily. So he waits until Jones is about to make up his mind what to do, and he does nothing unless it is clear to him (Black is an excellent judge of such things) that Jones is going to decide to do something other than what he wants him to do. If it were to become clear that Jones is going to decide to do something else, Black would take effective steps to ensure that Jones decides to do what he wants him to do, by directly manipulating the relevant processes in Jones’s brain. As it turns out, Black never has to show his hand because Jones, for reasons of his own, decides to perform the very action Black wants him to perform.

The idea is that even though Jones could not have avoided acting as he did, it’s intuitive that he is morally responsible, blameworthy in particular, for the action: He did it on his own and Black never intervened. On Frankfurt’s diagnosis, such examples show that we need to distinguish between the factors that cause or causally explain an action and factors that contribute to making the action inevitable even though they do not cause or causally explain it (Frankfurt 1969; Fischer 1982, 1994). Black’s conditional intention to intervene renders Jones’s action inevitable, but this intention does not cause or causally explain the action; it is

Moral Responsibility and Alternative Possibilities    159 causally idle. The factors that cause and causally explain the action are instead the agent’s reasons and deliberation, as they are in an ordinary case in which an agent is morally responsible. Frankfurt contends that in the sort of example he introduced it is these causal factors, and not the factor that makes the action inevitable, that are relevant to the agent’s moral responsibility. Frankfurt examples yield two putative lessons about the nature of moral responsibility (McKenna and Pereboom 2016). The first is that because it’s intuitive that Jones is morally responsible, and the availability to him of alternative possibilities is ruled out, availability of alternative possibilities is not a necessary condition for moral responsibility. Second, what does account for moral responsibility are facts about the actual causal history of the action, in particular the deliberative process by which agents discern and weigh reasons for action and the sensitivity to reasons that this process involves (Frankfurt 1969; cf. Fischer 1982, 1994; McKenna 2005, 2008; Sartorio 2016). Michael McKenna formulates a response to the W-​Defense on the basis of Frankfurt’s conception, which he calls the L-​Reply (2005, 2008): “A person’s moral responsibility concerns what she does do and her basis for doing it, not what else she could have done.” (McKenna 2008: 785). In response to the challenge from Frankfurt examples, defenders of the alternative-​ possibilities requirement have advanced a number of strategies, and which one they deploy is sensitive to the specific nature of the example (McKenna and Pereboom 2016). A first is to contend that despite initial impressions, the example does feature a relevant alternative possibility after all; this has been called “the flicker of freedom strategy” (Fischer 1982, 1994). A second is to argue that despite initial impressions, the example requires causal determinism in the actual causal history of the action to secure the absence of alternatives, and this counts against the intuition that the agent is morally responsible (Kane 1996, Widerker 1995). A third response is to press the claim that the factors that make an action inevitable rule out moral responsibility even if they are not causally efficacious (Ginet 1996). In what follows, we will see how such strategies play out in the more specific types of objections that defenders of the alternative-​possibilities requirement have raised.

2.  Compatibilist and Incompatibilist Source Views In an important contribution to this debate, John Fischer (1982, 1994) has pointed out that Frankfurt examples do not challenge the incompatibilist claim that an agent’s moral responsibility for an action requires that it have an indeterministic causal history. Rather, it leaves it open that her moral responsibility requires that the action not result from a deterministic causal process that traces back to factors beyond her control—​back to

160   Derk Pereboom causal factors that the agent could not have produced, altered, or prevented (Pereboom 2001, 2014; Sartorio 2016: 201n). Notice that the Frankfurt example above does not specify that Jones’s action is causally determined in this way. If it did specify that his action is deterministically produced by factors beyond his control, then at least for some the intuition that he is morally responsible would vanish. This reflection suggests a different requirement for the sort of free will required for moral responsibility: (Source) An action is free in the sense required for moral responsibility only if its causal history, and in particular its causal source, is of an appropriate sort.

Frankfurt’s own compatibilist version of a source account is one on which the agent’s will to perform the action is endorsed by the agent’s second-​order desires: she must not only will the action, but she must want to will it (Frankfurt 1971). In this view, if the action has this sort of causal history, it will have its source in the agent in the way that moral responsibility requires. But the agent need not have access to alternative possibilities. Incompatibilists affirm that for an agent to be morally responsible for an action, its causal history must be indeterministic. However, incompatibilists who defend PAP typically endorse this requirement only because they maintain that such an indeterministic causal history required to secure access to alternative possibilities (Ginet 1996, 2007; Palmer 2011; Franklin 2011). Source incompatibilists, by contrast, contend that the role an indeterministic causal history has in explaining why an agent is responsible is independent of alternative possibilities. Rather, an indeterministic history—​ of the right sort—​allows the agent to be the source of her action in a way that secures the absence of causal determination by factors beyond her control (Stump 1990, 1996; Zagzebski 1991, 2000; Pereboom 1995, 2001, 2014; Hunt 2000, 2005; Shabo 2010). Source incompatibilists may in addition allow that alternative possibilities for action—​not necessarily of the robust sort—​are entailed by the actual causal history having these features (Della Rocca 1998; Pereboom 2001). Still, on any source incompatibilist view, the aspect of the action that has the important role in explaining why an agent is morally responsible is the nature of the actual causal history of the action, and not the alternative possibilities. Against source incompatibilism Fischer once argued that “there is simply no good reason to suppose that causal determinism in itself (and apart from considerations pertaining to alternative possibilities) vitiates our moral responsibility” (1994: 159). True, one widespread incompatibilist intuition is that if due to causal determination we could never have done otherwise, and therefore could never have refrained from any action we performed, we would not be morally responsible for them. But another widespread intuition is that if all of our actions were “in the cards” before we existed, in the sense that things happened before we were born that, due to a deterministic causal process, inevitably resulted in our actions, we would not be morally responsible for them. If

Moral Responsibility and Alternative Possibilities    161 all of our actions had this type of causal history, we would lack the kind of control over our actions moral responsibility requires.

3.  The Flicker of Freedom Defense Defenders of the alternative possibilities requirement have argued in various ways that it can withstand the argument from Frankfurt examples. We’ll address two prominent objections to this argument: the Flicker of Freedom Defense and the Dilemma Defense. A third important objection, the Timing Defense (Ginet 1996, 2002; Palmer 2011; Franklin 2011) we’ll set aside. It’s intricate, as are the responses to it (Hunt and Shabo 2013; Pereboom 2014; Capes 2016; McKenna 2018). The Flicker of Freedom Defense concedes the intuition that the agent in a Frankfurt example is morally responsible, but then claims that the example features an alternative possibility that explains the intuition. The core idea is this: Any Frankfurt example must feature some event that the neuroscientist, or her device, is set up to detect that could have but does not actually occur, such as the agent’s intending to do otherwise (e.g., van Inwagen 1983: 166–​180; Fischer 1994: 134–​147; and Naylor 1984).This factor—​a “flicker of freedom,” to use Fischer’s term—​can then serve as the alternative possibility required for moral responsibility. It’s not implausible that the agent’s intending to do otherwise should count as the relevant sort of alternative possibility, and thus account for the agent’s moral responsibility. Fischer replies that one can construct Frankfurt examples in which what a neuroscientist’s device detects cannot plausibly have this explanatory role. One might, for instance, imagine that Jones will decide to kill Smith only if Jones blushes beforehand, and that Black’s device will activate only if Jones does not blush by a certain time (Fischer 1982, 1994). Jones’s failure to blush by a certain time might be the alternative possibility that would trigger the intervention that causes him to kill Smith. Supposing that Jones acts without intervention, we might well have the intuition that he is blameworthy. The alternative possibility he has is that he could have failed to blush, but Fischer argues that this possibility is too “flimsy and exiguous” and is “insufficiently robust” to have a role in grounding moral responsibility. What is it for an alternative possibility to be robust, and why would the robustness of an alternative possibility be crucial? The intuition underlying the alternative-​ possibilities requirement is that if an agent is to be blameworthy for an action, he must have voluntary access to an exempting alternative. That is, as a result of voluntarily accessing that alternative possibility instead, he would thereby have avoided the blameworthiness he actually has for the action (Mele 1996; McKenna 1997; Wyma 1997; Otsuka 1998; Pereboom 2001). Accordingly, in the example above, Jones’s failing to blush is not robust, since by failing to blush Jones would not have avoided responsibility for killing Smith.

162   Derk Pereboom Robustness also has an epistemic dimension. Imagine that Joe decides to take an illegal deduction on his tax form, and that he is in fact blameworthy for so deciding. Suppose that just before beginning to work on his taxes the day he makes the decision, Joe went to the coffee shop next door with a friend and ordered a cup of coffee, and then a second. Unbeknown to him, the barrista, having overheard Joe discuss his plans for the day and making dismissive remarks about taxation, laces the second cup with a drug that induces compliance with the tax code for 24 hours (cf. Pereboom 2000, 2001, 2104; Moya 2006: 64).1 As it happens, Joe does not take a sip from the second cup, and leaves it at the coffee shop before returning home to work on this taxes. In this situation, Joe could have behaved voluntarily so as to preclude his decision to take the illegal deduction, as a result of which he would not have been blameworthy for making that decision. But in these circumstances, whether he could have voluntarily taken the sip from the second cup is irrelevant to explaining why or whether he is blameworthy for his decision, and this is because Joe does not understand, nor could he reasonably be expected to have understood, that taking the sip would have rendered him blameless. Thus if an agent is blameworthy because he has an alternative possibility, it must be that he understood, at least at some level, that or how it was available to him. Here is a substantial necessary condition that accommodates these reflections; Robustness: For an agent to have a robust alternative to her immoral action A, that is, an alternative relevant per se to explaining why she is blameworthy for performing A, it must be that

(i) she instead could have voluntarily acted or refrained from acting as a result of which she would be blameless, and (ii) for at least one such exempting acting or refraining, she understood, at some level, that she could so voluntarily act or refrain, and that if she voluntarily so acted or refrained she would then be, or would likely be, blameless. (Pereboom 2014: 13)

In accord with this criterion, the response to the flicker defense is that Frankfurt examples can be constructed which, although they feature an alternative possibility, that alternative possibility is not robust, and thus cannot serve to explain the agent’s moral responsibility for her action. The resulting source view opposes, specifically, the following version of PAP: (PAP-​Robust) An action is free in the sense required for moral responsibility only if the agent has access to a robust alternative to that action.

1 

The need for the timing provision in the coffee cup example is occasioned by Justin Capes’s (2016) criticism of Pereboom’s earlier (2001, 2014) version.

Moral Responsibility and Alternative Possibilities    163

4.  The Dilemma Defense The Dilemma Defense was initially suggested by Robert Kane and then developed in detail by David Widerker (Kane 1985: 51; Kane1996: 142–​44, 191–​92; Widerker 1995: 247–​261; cf. Ginet 1996). It’s an objection to Frankfurt cases raised specifically from the perspective of the libertarian. Here is how Widerker sets out the dilemma. For any Frankfurt example, if causal determinism is assumed to hold in the actual sequence that issues in the action, then a libertarian cannot be expected to have the intuition that the agent is morally responsible. Libertarians are, after all, incompatibilists about moral responsibility and causal determination. That’s the first horn of the dilemma. The second horn is that if indeterminism in the actual sequence is assumed, the scenario will not serve the Frankfurt-​defender’s purpose, for any such case will fall to a further dilemma. In Frankfurt examples the actual situation will feature a prior sign, such as Jones’s blush, that signals the fact that the intervention is not required. If the prior sign causally determined the action, or if it were associated with some factor that did, the intervener’s (or his device’s) predictive ability could be explained. However, then the libertarian, given her incompatibilism, should not be expected to have the intuition that the agent is morally responsible. But if the relation between the prior sign and the action is not causally deterministic, then it’s open that the agent could have done otherwise despite the occurrence of the prior sign. Then it’s also open that the intuition that the agent is morally responsible can be explained by access to an alternative possibility. Either way, PAP (or PAP-​Robust) emerges unscathed.

4.1  The Mele-​Robb Example and Hunt’s Blockage Strategy In response, a number of critics have proposed Frankfurt examples designed to evade this objection, including Eleonore Stump (1990, 1996); Al Mele and David Robb (1998); David Hunt (2000, 2005); Derk Pereboom (2000, 2001, 2014); Michael McKenna (2003); David Widerker (2006); and John Fischer (2010). In one kind of case, first proposed by David Hunt (2000), there are no prior signs to guide intervention, not even non-​robust flickers of freedom. One such example is set out by Mele and Robb (1998). It features Bob, who inhabits a world in which determinism is false, but in which some processes are deterministic: At t1, Black initiates a certain deterministic process P in Bob’s brain with the intention of thereby causing Bob to decide at t2 (an hour later, say) to steal Ann’s car. The process, which is screened off from Bob’s consciousness, will deterministically culminate in Bob’s deciding at t2 to steal Ann’s car unless he decides on his own to steal it or is incapable at t2 of making a decision (because, e.g., he is dead at t2). . . . The process is in no way sensitive to any “sign” of what Bob will decide. As it happens, at t2 Bob decides on his own to steal the car, on the basis of his own indeterministic

164   Derk Pereboom deliberation about whether to steal it, and his decision has no deterministic cause. But if he had not just then decided on his own to steal it, P would have deterministically issued, at t2, in his deciding to steal it. Rest assured that P in no way influences the indeterministic decision-​making process that actually issues in Bob’s decision.

Mele and Robb contend that Bob is morally responsible for his decision, and thus, given that Bob lacks a robust alternative, their case provides a successful response to the dilemma defense. One concern for this case is that it involves trumping preemption (Schaffer, 2000), and that this notion is incoherent. As the example is described, Bob’s deciding on his own to steal Ann’s car trumps and preempts the deterministic process Black initiates. What might the details of such a process be? Jonathan Schaffer proposes the following example of trumping preemption. Imagine that a group of soldiers is trained to obey a general rather than a major when these two officers issue orders at the same time (2000). Suppose that the general and the major shout, “March!” simultaneously. Here only the general’s order is causally efficacious, since it trumps and preempts the major’s order. But note that in this case the trumping preemption has a detailed explanation, so perhaps Mele and Robb owe us an explanation of how trumping preemption might work in their example (Pereboom 2000, 124–​128, 2001: 14–​180). Mele and Robb discuss a number of possible concerns for their case, one of which is whether we can make sense of what would happen at t2 if the deterministic process P and Bob’s indeterministic deliberative process were to diverge at t2, in particular if Bob’s deliberative process resulted in his deciding not to steal, while P, as the setup specifies, resulted in his deciding to steal. Here is how they describe the concern: The issue may be pictured, fancifully, as follows. Two different “decision nodes” in Bob’s brain are directly relevant. The “lighting up” of node N1 represents his deciding to steal the car, and the “lighting up” of node N2 represents his deciding not to steal the car. Under normal circumstances and in the absence of preemption, a process’s “hitting” a decision node in Bob “lights up” that node. If it were to be the case both that P hits N1 at t2 and that x does not hit N1 at t2, then P would light up N1. If both processes were to hit N1 at t2, Bob’s indeterministic deliberative process, x, would light up N1 and P would not. The present question is this. What would happen if, at t2, P were to hit N1 and x were to hit N2? That is, what would happen if the two processes were to “diverge” in this way? And why?

And here is the answer they provide: We extend Bob’s story as follows. Although if both processes were to hit N1 at t2, Bob’s indeterministic deliberative process, x, would preempt P and light up N1, it is also the case that if, at t2, P were to hit N1 and x were to hit N2, P would prevail. In the latter case, P would light up N1 and the indeterministic process would not light up N2. Of course, readers would like a story about why it is that although x would preempt P in the former situation, P would prevail over x in the latter. Here is one story.

Moral Responsibility, Alternative Possibilities, and Frankfurt    165 By t2, P has “neutralized” N2 (but without affecting what goes on in x). That is why, if x were to hit N2 at t2, N2 would not light up. More fully, by t2 P has neutralized all of the nodes in Bob for decisions that are contrary to a decision at t2 to steal Ann’s car (e.g., a decision at t2 not to steal anyone’s car and a decision at t2 never to steal anything). In convenient shorthand, by t2 P has neutralized N2 and all its “cognate decision nodes.” Bear in mind that all we need is a conceptually possible scenario, and this certainly looks like one. (1996: 104–​105)

The aspect of this story that might raise a concern is the provision that P neutralizes, by t2, N2 and all its cognate decision nodes. A libertarian might contend that P’s neutralizing procedure is equivalent to P’s causal determining Bob’s decision to steal the car; after all, the neutralization would seem to be a causal process that renders inevitable x’s hitting N1 instead. However, Mele and Robb do specify that P’s neutralizing activity does not affect what goes on in Bob’s indeterministic decision-​making process, and if so, it would seem that P would not causally determine the decision. Who is right, the libertarian or Mele and Robb? Let’s examine a strategy that initially invokes the neutralization idea, an approach known as “blockage,” proposed by David Hunt (2000). Consider two situations (Pereboom 2000, 125–​128, 2001: 15–​18): Situation A: Scarlet deliberately chooses to kill Mustard at t1, and there are no factors beyond her control that deterministically produce her choice. When Scarlet chooses to kill Mustard, she could have chosen not to kill him. There are no causal factors, such as intervention devices, that would prevent her from not making the choice to kill Mustard.

Here the libertarian defender of the alternative-​possibilities requirement would agree that Scarlet can be morally responsible for her choice. But against that requirement we can employ the following variant: Situation B: Scarlet’s choice to kill Mustard has precisely the same actual causal history as in A. But before she even started to think about killing Mustard, a neuroscientist had blocked all the neural pathways not used in Situation A, so that no neural pathway other than the one employed in that situation could be used. Let’s suppose that it is causally determined that she remain a living agent, and if she remains a living agent, some neural pathway has to be used. Thus every alternative for Scarlet is blocked except the one that realizes her choice to kill Mustard. But the blockage does not affect the actual causal history of Scarlet’s choice, because the blocked pathways would have remained dormant.

Should we also have the intuition that Scarlet is morally responsible for her choice in Scenario B, since it does not seem to feature any relevant divergence from Scenario A? This might be challenged upon more careful reflection on how Scarlet’s action is caused in Scenario B. A crucial question about a blockage case is one that Fischer asks, “Could

166   Derk Pereboom neural events bump up against, so to speak, the blockage?” (Fischer 1999: 119). If so, there might be alternative possibilities in Scenario B. However, if neural events can’t bump up against the blockage, then, as Kane suggests, there is a serious threat that the neural events are causally determined, that we now have “determinism pure and simple” (Kane 2000: 162). In response, the advocate of blockage cases might point out that in more standard Frankfurt examples the action is also inevitable, while the intuition that the agent is morally responsible for it is explained by the fact that it does not have an actual causal history by which it is made inevitable. What makes the action inevitable is instead some provision of the case that makes no difference to the action’s actual causal history. So it’s open that in standard Frankfurt cases the action’s being inevitable is compatible with its not being causally determined. How then is a blockage case different from the standard Frankfurt-​style cases? The blockage wouldn’t seem to be a feature of the actual causal history of the action either. Nevertheless, perhaps Kane’s charge of determinism can be defended. It might be that two-​situation cases such as the one just set out are misleading just because it is natural to assume that the actual causal history of the action is essentially the same in each. The sole difference between them is a feature that would seem not to alter the actual causal history of the action. But now consider a simple two-​situation case modeled on a reflection of Hunt’s (Fischer 1999: 119–​120). Imagine a universe correctly described by Epicurean physics (Pereboom 2001: 17). At the fundamental level this universe features only atoms and the frictionless void. The atoms fall in a determinate downward direction, except when they undergo uncaused swerves. Situation C: A spherical atom is falling downward through space, with a certain velocity and acceleration. Its actual causal history is indeterministic because at any time the atom can be subject to an uncaused swerve. Suppose that the atom can swerve in any direction other than upwards. In actual fact, from t1 to t2 it does not swerve.

The following counterfactual situation diverges from C only because it features a device that eliminates alternative possibilities and all differences thereby entailed: Situation D: This case is the identical to C, except that the atom is falling downward through a straight and vertically oriented tube whose interior surface is made of frictionless material, and whose interior is precisely wide enough to accommodate the atom. The atom would not have swerved during this time interval, and the trajectory, velocity, and acceleration of the atom from t1 to t2 are precisely what they are in C.

We might at first have the intuition that the causal history of the atom from t1 to t2 in these two situations is exactly the same. But this intuition is challenged by the fact that the restrictions present in D but not in C may change the causal history from one that is indeterministic to one that is deterministic. The tube prevents any alternative motion,

Moral Responsibility and Alternative Possibilities    167 and it would seem to rule out any indeterminism in the atom’s causal history between t1 to t2. If in this way the tube rules out indeterminism, it would appear to render the actual causal history deterministic. This difficulty may make it hard to assess moral responsibility in blockage cases with confidence. Returning to Mele’s and Robb’s example, it’s a crucial feature of the case that if Black’s deterministic process P and Bob’s indeterministic deliberative process x were to diverge at t2, P would neutralize N2 and all its cognate decision nodes. In this respect it is a blockage scenario, and Mele and Robb’s example inherits the challenge to these types of cases.

4.2  Pereboom’s Tax Evasion Case David Hunt and I have proposed a type of Frankfurt example, sometimes referred to as a “buffer” example, designed to respond to the Dilemma Defense (Pereboom 2000, 2001: 18–​22, 2003, 2009, 2012, and 2014: 14–​29; David Hunt defends a similar example in Hunt 2005, which he initially suggested but rejected in Hunt 2000). The distinguishing features of my example, Tax Evasion, are these: The trigger for the intervention of the neuroscientist’s device is a necessary condition for the agent’s accessing any robust alternative possibility (without the intervener’s device in place), while the cue is not a robust alternative possibility, and the non-​occurrence at any specific time of the cue for intervention in no sense causally determines the action the agent performs. Here is a version of the case (Pereboom 2014: 15): Tax Evasion (2): Joe is considering claiming a tax deduction for the registration fee that he paid when he bought a house. He knows that claiming this deduction is illegal, but that he probably won’t be caught, and that if he were, he could convincingly plead ignorance. Suppose he has a strong but not always overriding desire to advance his self-​interest regardless of its cost to others and even if it involves illegal activity. In addition, the only way that in this situation he could fail to choose to evade taxes is for moral reasons, of which he is aware. He could not, for example, fail to choose to evade taxes for no reason or simply on a whim. Moreover, it is causally necessary for his failing to choose to evade taxes in this situation that he attain a certain level of attentiveness to moral reasons. Joe can secure this level of attentiveness voluntarily. However, his attaining this level of attentiveness is not causally sufficient for his failing to choose to evade taxes. If he were to attain this level of attentiveness, he could, exercising his libertarian free will, either choose to evade taxes or refrain from so choosing (without the intervener’s device in place). However, to ensure that he will choose to evade taxes, a neuroscientist has, unbeknownst to Joe, implanted a device in his brain, which, were it to sense the requisite level of attentiveness, would electronically stimulate the right neural centers so as to inevitably result in his making this choice. As it happens, Joe does not attain this level of attentiveness to his moral reasons, and he chooses to evade taxes on his own, while the device remains idle. (Pereboom 2000, 2001, 2003, 2009; David Hunt suggested such a strategy

168   Derk Pereboom (2000), and later develops a similar example (2005); and Seth Shabo (2010) proposes valuable refinements).

Joe is intuitively blameworthy for deciding to evade taxes, even though he does not have access to a robust alternative possibility. Notice that cases proposed before the dilemma defense was formulated, such as Fischer’s blush example, also feature a necessary condition for doing otherwise. In Fischer’s case, that necessary condition is the blush not ever occurring. So the distinctive characteristic of Tax Evasion is not the presence of a necessary condition for doing otherwise, but a necessary condition for doing otherwise the absence of which at any specific time does not causally determine the agent to perform the action. This feature of the case ensures that at no specific time is the agent in the example causally determined to perform the action, and this facilitates the satisfaction of a requirement for moral responsibility the libertarian will not relinquish. In Tax Evasion, the necessary condition for Joe’s not deciding to evade taxes, that is, his having the specified level of attentiveness to the moral reasons, is the right sort, since its absence at any one particular time does not causally determine his deciding to evade taxes. At any one time at which the level of attentiveness is absent, Joe could still make it occur at a later time, and as a result he is not causally determined to decide to evade taxes by its absence at the previous time. (Although if there is a deadline, one needs to consider what happens if he waits to decide until then (Pereboom 2014: 26–​27)). Note that Fischer (2010) develops an argument that features a Frankfurt example similar to Tax Evasion but set in a deterministic context. If Fischer’s argument is persuasive, determinism in the actual sequence is not the obstacle to a successful Frankfurt example that the Dilemma Defense specifies it to be. Tax Evasion does feature alternative possibilities accessible to the agent—​ Joe’s attaining higher levels of attentiveness to moral reasons. This fact about the case prompts the objection that by voluntarily attaining the specified higher level of attentiveness, Joe would have voluntarily done something as a result of which he would have avoided the blameworthiness he actually incurs. Had he voluntarily attained the requisite level of attentiveness, the intervention would have occurred, whereupon Joe would have decided to evade taxes, but not in such a way that he would have been blameworthy for that decision. I respond by arguing that this alternative possibility is not robust. Joe has no understanding of the fact that by voluntarily attaining the requisite level of attentiveness he would not be (or would likely not be) blameworthy, and we would not reasonably expect him to have this understanding. Rather, he believes that attaining this level of attentiveness is compatible with his freely deciding to evade taxes (which would be true without the intervener’s device in place), and he has no reason to think otherwise. It can even be specified that Joe believes that if he did attain this level of attentiveness, he would still be highly likely to decide to evade taxes. But despite his lacking a robust alternative, Joe is intuitively blameworthy for his actual decision (Pereboom 2000, 2001, 2014). Against this, Carlos Moya (2011) objects that in an ordinary situation the alternative Joe has would not be robust; but this is a Frankfurt case, not an ordinary situation, and accessing the alternative he has, being more attentive to moral reasons, is the best he can

Moral Responsibility and Alternative Possibilities    169 do. My reply is that alternative still is not exempting, and Moya’s reason for changing the standard for robustness is suspect, since it is in effect motivated by PAP (Pereboom 2009, 2014: 17). Justin Capes (2016; cf. McKenna 2018) points out that whether Joe is attentive to moral reasons is not irrelevant to whether he is responsible for deciding to evade taxes. But even if Joe’s having this alternative is relevant to accounting for his blameworthiness, without the device in place this alternative is not exempting at all, and with the device in place it is not an exempting alternative that meets the epistemic condition on robustness, since Joe has no inkling that it’s exempting. One significant challenge to Tax Evasion is the claim that Joe’s responsibility for his decision is merely derivative of his responsibility for his earlier voluntary refraining from becoming more attentive to moral reasons, for which he does have a robust alternative (Widerker 2006; for the response recounted here, see Pereboom 2009, 2014). The example does allow that Joe has access to the alternative possibility of becoming more attentive to moral reasons—​indeed, this is the cue for intervention. A reply to this challenge begins by delineating what derivative moral responsibility amounts to. For a paradigm case, Biff decides to get drunk, understanding that when he is intoxicated he will no longer be able to avoid being abusive to his companions, and then after he becomes drunk he punches one of them. Here Biff satisfies generally accepted conditions on moral responsibility at the time he decides to get drunk, but not after he becomes drunk. If Biff is blameworthy for punching his companion, it is only derivatively so—​derivative on his being non-​derivatively morally responsible for deciding to get drunk, and his foreseeing that when he is drunk he is likely to become violent (Ginet 2000). Widerker (2006) proposes this challenge to Tax Evasion, contending that non-​ derivative moral responsibility is governed by the alternative-​possibilities requirement, while derivative responsibility need not be: [A]‌problem with Pereboom’s example is that, in it, the agent is derivatively blameworthy for the decision he made, because he has not done his reasonable best (or has not made a reasonable effort) to avoid making it. He should have been more attentive to the moral reasons than he in fact was—​something he could have done. And in that case, he would not be blameworthy for deciding to evade taxes, as then he would be forced by the neuroscientist so to decide. If this is correct, then Pereboom’s example is a case of derivative culpability, and hence is irrelevant to PAP, which . . . concerns itself only with direct or nonderivative culpability. (Widerker 2006: 173; Ginet anticipates this objection in Ginet 1996)

I respond by arguing that there is a sense in which Widerker’s challenge is an unsatisfying response, since it explicitly cites PAP in support of its verdict about Joe’s responsibility. Widerker’s thought appears to be that Joe is non-​derivatively morally responsible only for not deciding to be more attentive to the moral reasons because only relative to this decision does he have a robust alternative, and hence any responsibility he has for deciding to evade taxes must be derivative of this decision. One might more generally

170   Derk Pereboom apply the following alternative-​possibilities schema to any case of this sort: An agent is non-​derivatively morally responsible for acting or refraining at a particular time only if a robust alternative possibility relative to so acting or refraining is accessible to her then, and all other moral responsibility is derivative of such non-​derivative responsibility. However, an objection to this schema is that it will miss the force of any potential counterexample, and will thus risk failing to engage an important objection. It is in a sense question-​begging (Pereboom 2009, 2014). In judging whether the schema does have this defect, the drawbacks for imposing it on situations like Joe’s need to be assessed. One might at first suppose that there won’t be any such drawbacks because intuitions about whether agents are morally responsible do not distinguish between non-​derivative and derivative responsibility. It may be intuitive that Joe is morally responsible for his decision, and not intuitive that he is non-​derivatively by contrast with merely derivatively morally responsible. But in reply, a paradigm for derivative responsibility is the drunkenness example cited earlier. In this case, when Biff is drunk, he does not satisfy some of the uncontroversial general conditions on non-​derivative moral responsibility, yet they are all satisfied at the time at which he decides to get drunk. However, Biff ’s circumstances are relevantly different from Joe’s. Biff has knowingly put himself in a position in which some of the uncontroversial general conditions for non-​derivative moral responsibility will not be satisfied at relevant later times. This not the case for Joe. One might object that if an agent can’t attain sufficient attentiveness to moral reasons, one such uncontroversial condition is not satisfied. However, at any time when Joe is not sufficiently attentive to moral reasons, he understands that it is open to him to become sufficiently more attentive at a later time. Thus Joe’s situation differs in a crucial respect from the paradigm example of derivative responsibility. In my view, this shows that the application of the alternative-​possibilities schema to Joe in the way proposed in Widerker’s objection is inappropriate (Pereboom 2009, 2014). McKenna (2018) replies by arguing that this sort of Frankfurt example is a mixed case. Joe understands, or is at least cognitively sensitive to the fact that he must be attentive to moral reasons in order to avoid deciding to evade taxes. So part of his blameworthiness is grounded in the failure to decide to be attentive, for which he has a robust alternative. One might respond that for any decision D that we make, there is a significant background of previous decisions that bear on our moral responsibility for D, so it’s usually the case that our responsibility for a decision is explained partly by our responsibility for earlier decisions. The key point is that the plausibility of PAP-​Robust is grounded in the intuition that decisions for which we’re blameworthy require access to exempting alternatives, and Joe’s responsibility for his failure to be attentive is not exempting relative to his responsibility for deciding to evade taxes, and it’s the notion of an exempting alternative that grounds PAP and PAP-​Robust. McKenna (2018) proposes an amendment to the example that responds to the problem he raises: Set up the example so that whether Joe achieves the specified attentiveness to moral reasons is not determined, but is also not voluntary. So whether the attentiveness occurs is due to an indeterministic process breaking a certain way, and in

Moral Responsibility and Alternative Possibilities    171 actual fact it does not. Thus in the actual situation, Joe does not become aware, but he might have due to an indeterministic process that does not involve a choice he has between alternatives. McKenna does raise the objection that libertarians might not think that Joe is sufficiently free given this setup. My sense is that some may find the original case more persuasive, while others the one that features McKenna’s suggested revision, and that it’s good to have both versions in the Frankfurt defender’s toolkit. (In fact, in the earliest versions of Tax Cut, the trigger for intervention is a moral reason occurring to Joe with a certain force, and I say that “A moral reason can occur to him with that force either involuntarity or as a result of his voluntary activity” (Pereboom 2000: 128, 2001: 19)).

5.  Is a General Ability to Do Otherwise Enough? Virtually everyone, compatibilists and incompatibilists, would agree that an agent’s causal determination is compatible with having the ability to do otherwise given some characterization, that is, one that does not involve holding the laws or the past fixed. For example, you and I are in Rome, you’re from the USA but speak Italian well, but you order a Negroni in English. Even if you’re causally determined by the past and the laws so to act, there is a sense in which it’s true to say that you could have ordered in Italian. If the past and the laws are held fixed, then your ordering in Italian in ruled out, but not if we loosen what’s held fixed so as to accommodate the truth of “you could have ordered in Italian.” There is a reason not to hold the past and the laws fixed in fixing the truth conditions of this sentence. Intuitively, in this situation you retain the general ability to speak and order in Italian, despite being causally determined by factors to act. Susan Wolf (1990) proposes that what is crucially necessary is a general power to act and to have acted differently. A number of compatibilists have pointed out that an agent in an appropriately constructed Frankfurt example retains a general ability or power to refrain from performing the action at issue (Wolf 1990; McKenna 1997; Vihvelin 2004; Fara 2008; Nelkin 2011: 66–​76). In Tax Evasion, Joe retains the general ability to refrain from deciding as he does. On Wolf ’s more specific proposal, which Nelkin develops and endorses, Joe retains what Nelkin calls an interference-​free ability to refrain from performing the action. That is, the agent possesses the skills, talents, and so on, required for refraining, and nothing actually interferes with or prevents the exercise of this power to refrain (Wolf 1990: 110; Nelkin 2011: 66–​68). The libertarian defender of an alternative-​possibilities condition is apt to raise the following objection to this proposal. Due to the intervention setup, it is not up to Joe to successfully exercise his interference-​free ability to do otherwise at the time of his actual decision to evade taxes. Moral responsibility requires more than just the general ability to do otherwise: One’s circumstances must also permit the exercise of the capacity.

172   Derk Pereboom Whether this is required for moral responsibility is what’s at issue between the defender of an alternative possibilities condition and her opponents (McKenna 1997; Kane 2002; Clarke 2009; Whittle 2010). The counterclaim can be stated as follows. Despite possessing an interference-​free ability to refrain from making this decision, Joe nevertheless lacks access to an exempting alternative. If when he decides to evade taxes it’s not up to him to successfully exercise his interference-​free ability to refrain from so deciding, it’s not up to him to avoid the blameworthiness he actually incurs.

6.  Frankfurt Cases for Omissions Some of those who have been convinced by Frankfurt examples to reject the Principle of Alternative Possibilities for actions nevertheless contend that one can be directly responsible for an omission only if one has access to an alternative. The control condition on moral responsibility would then feature an asymmetry: Direct responsibility for an action does not require an ability not to perform that action, whereas direct responsibility for not performing a certain action requires an ability to perform that action. This is a complicated debate, especially because, as Randolph Clarke (2021, ­chapter 5) notes, some, such as Clarke himself (1994, 2011) reject certain asymmetries and accept others.2 Against Fischer and Ravizza (1998, 254–​255) Carolina Sartorio argues that the following pair of Frankfurt examples indicate an asymmetry for moral responsibility for the outcomes of actions and moral responsibility for the outcomes of omissions (2005, 2017): Active Frank: A child is swimming in a pond. Frank wants him to die, so he jumps in and pushes the child’s head under water until he drowns. An evil neuroscientist, who also wanted the child to die, had been monitoring Frank’s brain activity. Had Frank wavered in his decision to kill the child, the neuroscientist would have intervened by manipulating Frank’s brain in such a way that he would have made exactly the same decision. Omissive Frank: Frank notices that the child is starting to drown. Since he wants the child to drown, he decides not to jump in. This time there were no sharks in the water, but there was a neuroscientist in the background. Had Frank wavered in his decision not to jump in, the neuroscientist would have manipulated Frank’s brain in such a way that he would have made the same decision.

2  Fischer (1985–​86) and Fischer and Ravizza (1991) argued for the asymmetry but they (1998) subsequently rejected it, as have Byrd (2007), Clarke (1994, 2011), Frankfurt (1994), and McIntyre (1994). Fischer (2017) and Sartorio (2005) each argue for a different asymmetry (from the one characterized in the text here) between responsibility for actions and responsibility for omissions (as does Clarke [1994, 2011]). Note, further, that other writers (e.g., Ginet [2003], Widerker [1995]) reject the asymmetry because they reject the first half.

Moral Responsibility and Alternative Possibilities    173 Here is what Sartorio (2017) says about those who reject the asymmetry: The philosophers who reject the asymmetry between responsibility for/​by actions and responsibility for/​by omissions would suggest that Omissive Frank is just as responsible as Active Frank: he is responsible for his decision not to save the child, for his failure to save the child, and for the child’s death. This is so even though he couldn’t have made a different decision and thus he couldn’t have saved the child. Hence, they would suggest, Frankfurt-​style omission cases show that responsibility for/​by omissions doesn’t require the ability to do otherwise in the same way that Frankfurt-​style action cases show that responsibility for/​by actions doesn’t require that ability.

Sartorio defends the asymmetry by arguing that although Omissive Frank is responsible for his decision not to save the child, a positive mental action, he is not blameworthy for his omission to save the child or for the child’s death. Frank’s moral responsibility for his decisions doesn’t carry over to his not saving the child, or to the child’s death, because it doesn’t bear the right kind of relation to those things. In Sartorio’s view, “the child doesn’t die as a result of what Frank in fact decided to do, but as a result of what he didn’t decide to do (i.e. because he didn’t decide to jump into the water to save him).” Moreover, “it’s not at all clear that he is responsible for not having made the relevant decision (an omission, or an absence); all that’s reasonably clear is that he is responsible for having made certain decisions” (Sartorio 2005, 2017; cf. Clarke 2014: 138–​143) and Swenson (2015)). Omissive Frank, as well as the cases that Fischer and Ravizza (1998) adduce, all concern moral responsibility for consequences of a decision not to act. Responsibility for consequences is a complex matter. In addition, decisions not to act are not themselves omissions; rather they are basic actions. At this point we ask whether there is a successful Frankfurt case in which it’s intuitive that the agent is directly responsible for the omission of a basic action, such as a decision, by contrast with responsibility for consequences of a basic action. Clarke (1994) has proposed a Frankfurt case for an omission in which the omission at issue is forgetting a promise: Sam promises to babysit little Freddy. But Sam forgets. No one makes Sam forget; it just slips his mind. Consequently, he fails to show up to babysit little Freddy. Unbeknownst to Sam, a mad scientist is monitoring his thoughts. Had Sam been going to remember his promise, the scientist would have intervened and prevented him from remembering it. The scientist would not have intervened in any other way. As it happened, the scientist did not intervene at all; there was no need to.

Clarke (1994) affirms the intuition of responsibility for such a case, but later (Clarke 2011) suggests that it might be that it is merely derivative: “Perhaps the way to go is to require that the agent be responsible for some action(s) that caused, in some unproblematic way, his thinking of what he was in fact thinking of at the pertinent time.”

174   Derk Pereboom One concern for this example is that direct responsibility for forgetting (supposing this is possible, by contrast with responsibility for forgetting always being derivative) isn’t a clear example of direct responsibility for omitting to decide or omitting to intend. Clarke (2011) frames a forgetting case as an instance of direct responsibility for ceasing to intend, which does involve omitting to continue to intend, but not omitting to decide or to intend. Clarke’s case also involves the flicker “going to remember his promise” which, without further specification, threatens to be robust. I’ve proposed an example in which the agent would be directly responsible for omitting to decide (Pereboom 2015). Note first that in Sartorio’s Omissive Frank it may be that the only way Frank can omit to make the decision to rescue the child is by making the decision not to rescue the child. Arguably this is only because of the gravity of this decision. It seems psychologically impossible to be aware of the danger the child is in and not make a decision either way about saving it. However, in circumstances in which the harm at issue is not severe, an agent might omit to make a decision to help without making a decision not to help. So consider: Stranded Motorist: Frank is driving along a lightly-​traveled highway thinking about Frankfurt cases. He notices a motorist struggling to replace a flat tire. After momentarily being distracted by the thought that the motorist would benefit from his help, Frank resumes thinking about Frankfurt cases without ever deciding not to stop to help. He just stops thinking about the motorist’s plight and resumes thinking about Frankfurt cases. The motorist continues to struggle with the flat tire.

It’s intuitive that Frank is directly morally responsible for not making the decision to stop to help. We can now produce a Frankfurt example by adding in the neuroscientist: A neuroscientist is monitoring Frank’s psychological states. In Frank’s case, deciding to stop to help requires seriously considering doing so in advance, and suppose he can in fact do this. But had Frank seriously considered deciding to stop to help, the neuroscientist would have intervened and caused him not to decide either way. Frank does not seriously consider doing so, and so the neuroscientist does not intervene. (Pereboom 2015)

The flicker of freedom in this case is not a robust alternative, since Frank might seriously consider stopping to help and still not actually decide to stop to help. The available alternative would thus not be exempting. This is arguably also not an instance of derivative responsibility. By in contrast with Biff in the drunkenness paradigm, Frank hasn’t knowingly put himself in a position in which he will fail to satisfy some uncontroversial condition on responsibility at the relevant subsequent time. If all of this is correct, then there would no general asymmetry for an alternative-​possibilities requirement regarding basic actions and omissions of basic actions.

Moral Responsibility and Alternative Possibilities    175

7.  Final Words Frankfurt’s challenge has had a massive and continuing influence on the free will debate. Many participants in the debate, both compatibilist and incompatibilist, are convinced by his argument. But a significant and vocal minority disagrees, and for these theorists the intuition that blameworthiness requires access to an exempting alternative remains strong.

References Byrd, Jeremy (2007). Moral responsibility and omissions. Philosophical Quarterly 57 (226):56–​67. Capes, Justin. (2016). “Blameworthiness and Buffered Alternatives,” American Philosophical Quarterly 53, pp. 269–​280. Clarke, Randolph. (1994). “Ability and Responsibility for Omissions,” Philosophical Studies 73, pp. 195–​208. Clarke, Randolph (2009). Dispositions, Abilities to Act, and Free Will: The New Dispositionalism. Mind 118 (470):323–​351. Clarke, Randolph. (2011). “Omissions, Responsibility, and Symmetry,” Philosophy and Phenomenological Research 82, pp. 594–​624. Clarke, Randolph (2014). Omissions: Agency, Metaphysics, and Responsibility. Oxford University Press. Della Rocca, Michael. (1998). “Frankfurt, Fischer and Flickers,” Noûs 32, pp. 99–​105. Fara, Michael. (2008). “Masked Abilities and Compatibilism,” Mind 117, pp. 843–​865. Fischer, John Martin (1986). Responsibility and Failure. Proceedings of the Aristotelian Society 86, 251–​270. Fischer, John Martin, and Mark Ravizza (1998). Responsibility and Control: A Theory of Moral Responsibility, Cambridge: Cambridge University Press. Fischer, John Martin. (1982). “Responsibility and Control,” Journal of Philosophy 79, pp. 24–​40. Fischer, John Martin & Ravizza, Mark (1991). Responsibility and inevitability. Ethics 101 (2):258–​278. Fischer, John Martin. (1994). The Metaphysics of Free Will, Oxford: Blackwell Publishers. Fischer, John Martin. (1999). “Recent Work on Moral Responsibility,” Ethics 110, pp. 93–​139. Fischer, John Martin. (2010). “The Frankfurt Cases: The Moral of the Stories,” Philosophical Review 119, pp. 315–​336. Fischer, John (2017). Responsibility, Autonomy, and the Zygote Argument. The Journal of Ethics 21 (3):223–​237 Frankfurt, Harry. (1969). “Alternate Possibilities and Moral Responsibility,” Journal of Philosophy 66, pp. 829–​839. Frankfurt, Harry. (1971). “Freedom of the Will and the Concept of a Person,” Journal of Philosophy 68, pp. 5–​20. Frankfurt, Harry (1994). An alleged asymmetry between actions and omissions. Ethics 104 (3):620–​623.

176   Derk Pereboom Frankfurt, Harry. (2002). “Reply to John Martin Fischer,” Contours of Agency: Essays on Themes from Harry Frankfurt, S. Buss and L. Overton, eds., Cambridge, MA: MIT Press. Franklin, Christopher. (2011). “Neo-​Frankfurtians and Buffer Cases: The New Challenge to the Principle of Alternative Possibilities,” Philosophical Studies 152, pp. 189–​207. Ginet, Carl. (1996). “In Defense of the Principle of Alternative Possibilities: Why I Don’t Find Frankfurt’s Arguments Convincing,” Philosophical Perspectives 10, pp. 403–​417. Ginet, Carl (2000). The Epistemic Requirements for Moral Responsibility. Noûs 34 (s14):267–​277. Ginet, Carl. (2002). “Review of Living Without Free Will,” Journal of Ethics 6, pp. 305–​309. Ginet, Carl (2007). An Action Can be Both Uncaused and Up to the Agent. In Lumer (ed.), Intentionality, Deliberation, and Autonomy. Ashgate. pp. 243–​255. Ginet, Carl (2003). Libertarianism. In Michael J. Loux & Dean W. Zimmerman (eds.), The Oxford Handbook of Metaphysics. Oxford: Oxford University Press. pp. 587–​612. Hunt, David, and Seth Shabo. (2013). “Frankfurt Cases and the (In)significance of Timing: A Defense of the Buffering Strategy, Philosophical Studies 164, pp. 599–​622. Hunt, David. (2000). “Moral Responsibility and Unavoidable Action,” Philosophical Studies 97, pp. 195–​227. Hunt, David. (2005). “Moral Responsibility and Buffered Alternatives,” Midwest Studies in Philosophy 29, pp. 126–​145. Kane, Robert. (1985). Free Will and Values, Albany: SUNY Press. Kane, Robert. (1996). The Significance of Free Will, New York: Oxford University Press. Kane, Robert. (2000). “The Dual Regress of Free Will and the Role of Alternative Possibilities,” Philosophical Perspectives 14, pp. 57–​80. Kane, Robert (ed.) (2002). The Oxford Handbook of Free Will. Oxford University Press. McKenna, Michael (2003). “Robustness, Control, and the Demand for Morally Significant Alternatives,” in Freedom, Responsibility, and Agency: Essays on the Importance of Alternative Possibilities, Michael McKenna and David Widerker, eds., Aldershot UK: Ashgate, 2003, pp. 201–​216. McKenna, Michael. (1997). “Alternative Possibilities and the Failure of the Counterexample Strategy,” Journal of Social Philosophy 28, pp. 71–​85. McKenna, Michael. (2005). “Where Frankfurt and Strawson Meet,” Midwest Studies in Philosophy 29, pp. 163–​180. McKenna, Michael. (2008). “Frankfurt’s Argument Qgainst the Principle of Alternative Possibilities: Looking Beyond the Examples,” Noûs 42, pp. 770–​793. McKenna, Michael, and Derk Pereboom (2016). Free Will: A Contemporary Introduction, London: Routledge. McKenna, Michael. (2018). Philosophical Studies 175, pp. 3117–​3129, https://​doi.org/​10.1007/​ s11098-​017-​0997-​z Mele, Alfred, and David Robb. (1998). “Rescuing Frankfurt-​Style Cases,” Philosophical Review 107, pp. 97–​112. Moya, Carlos. (2006). Moral Responsibility: The Ways of Skepticism, Oxford: Oxford University Press. pp. 195–​210. Moya, Carlos. (2011). “On the Very Idea of a Robust Alternative,” Critica 43, pp. 3–​26. Naylor, Marjory. (1984). “Frankfurt on the Principle of Alternate Possibilities,” Philosophical Studies 46, pp. 249–​258. Nelkin, Dana. (2008) “Responsibility and Rational Abilities: Defending an Asymmetrical View,” in Pacific Philosophical Quarterly 89, pp. 497–​515.

Moral Responsibility and Alternative Possibilities    177 Nelkin, Dana. (2011). Making Sense of Moral Responsibility, Oxford: Oxford University Press. Otsuka, Michael (1998). Incompatibilism and the avoidability of blame. Ethics 108 (4):685–​701. Palmer, David (2011). Pereboom on the Frankfurt cases. Philosophical Studies 153 (2):261–​272. Pereboom, Derk. (1995). “Determinism Al Dente,” Noûs 29, pp. 21–​45. Pereboom, Derk. (2000). “Alternative Possibilities and Causal Histories,” Philosophical Perspectives 14, pp. 119–​137. Pereboom, Derk. (2001). Living Without Free Will, Cambridge: Cambridge University Press. Pereboom, Derk. (2003). “Source Incompatibilism and Alternative Possibilities,” in Freedom, Responsibility, and Agency: Essays on the Importance of Alternative Possibilities, M. McKenna and D. Widerker, eds., Aldershot, UK: Ashgate Press, pp. 185–​199. Pereboom, Derk. (2009). “Further Thoughts About a Frankfurt-​Style Argument,” Philosophical Explorations 12, pp. 109–​118 Pereboom, Derk. (2012). “Frankfurt Examples, Derivative Responsibility, and the Timing Objection,” Philosophical Issues 22, pp. 298–​315. Pereboom, Derk. (2014). Free Will, Agency, and Meaning in Life, Oxford: Oxford University Press. Pereboom, Derk. (2015). “Omissions and the Different Senses of Responsibility,” in Agency and Moral Responsibility, Andrei Buckareff, Carlos Moya, and Sergi Rosell, eds., New York: Palgrave-​Macmillan, pp. 179–​191. Sartorio, Carolina (2005). A new asymmetry between actions and omissions. Noûs 39 (3):460–​482. Sartorio, Carolina. (2016). Causation and Free Will. Oxford: Oxford University Press. Sartorio, Carolina. (2017). “The Puzzle of Frankfurt-​Style Omission Cases,” The Ethics and Law of Omissions, Dana Nelkin and Samuel Rickless, eds., New York: Oxford University Press. Schaffer, Jonathan (2000). Trumping preemption. Journal of Philosophy 97 (4):165–​181. Shabo, Seth. (2010). “Uncompromising Source Incompatibilism,” Philosophy and Phenomenological Research 80, pp. 349–​383. Shabo, Seth. (2011). “Agency Without Avoidability: Defusing a New Threat to Frankfurt’s Counterexample Strategy,” Canadian Journal of Philosophy, 41, pp. 505–​522. Stump, Eleonore. (1990). “Intellect, Will, and the Principle of Alternate Possibilities,” in Christian Theism and the Problems of Philosophy, Michael Beaty, ed., Notre Dame IN, University of Notre Dame Press, pp. 254–​285. Stump, Eleonore. (1996). “Libertarian Freedom and the Principle of Alternative Possibilities,” in Faith, Freedom, and Rationality, Jeff Jordan and Daniel Howard Snyder, eds., Lanham, MD, Rowman and Littlefield, pp. 73–​88. Swenson, Philip. (2015). “A Challenge for Frankfurt-​Style Compatibilists,” Philosophical Studies 172, pp. 1279–​1285. van Inwagen, Peter. (1983). An Essay on Free Will, Oxford: Oxford University Press. Vihvelin, Kadri. (2004). “Free Will Demystified: A Dispositional Account,” Philosophical Topics 32, pp. 427–​450. Whittle, Ann (2010). Dispositional Abilities. Philosophers' Imprint 10. Wyma, Keith D. (1997). Moral Responsibility and Leeway for Action. American Philosophical Quarterly 34 (1):57–​70. Widerker, David. (1995). “Libertarianism and Frankfurt’s Attack on the Principle of Alternative Possibilities,” The Philosophical Review 104, pp. 247–​261. Widerker, David. (2000). “Frankfurt’s Attack on Alternative Possibilities: A Further Look,” Philosophical Perspectives 14, pp. 181–​201.

178   Derk Pereboom Widerker, David. (2006). “Libertarianism and the Philosophical Significance of Frankfurt Scenarios,” Journal of Philosophy 103, pp. 163–​187. Widerker, David. 1995. “Libertarianism and Frankfurt’s Attack on the Principle of Alternative Possibilities,” Philosophical Review 104, pp. 247–​261. Wolf, Susan. (1980). “Asymmetrical Freedom,” Journal of Philosophy 77, pp. 151–​166. Wolf, Susan. (1990). Freedom Within Reason, Oxford: Oxford University Press. Zagzebski, Linda. (1991). The Dilemma of Freedom and Foreknowledge, New York: Oxford University Press. Zagzebski, Linda. (2000). “Does Libertarian Freedom Require Alternate Possibilities?,” Philosophical Perspectives 14, pp. 231–​248.

CHAPTER 9

M an ipu l ation A rg ume nts Against C om pat i bi l i sm Derk Pereboom and Michael McKenna

1.  Introduction A widespread concern for our being morally responsible for our actions is the prospect that they are naturalistically causally determined, that is, rendered inevitable by virtue of the remote past in accord with the laws of nature, factors beyond our control. However, many compatibilists about naturalistic causal determination and moral responsibility are not strongly moved by this concern. Most people who enter into this controversy begin with the assumption that human agents are morally responsible in the basic desert sense when they knowingly do wrong. That is, they deserve to be blamed or punished just because they’ve knowingly done wrong. The justification for blaming or punishing thus does not invoke further reasons—​most saliently, it does not appeal to consequentialist or contractualist reasons (Pereboom 2001, 2014; cf. Feinberg 1970; Scanlon 2013). Some of these are natural compatibilists, and for them the prospect that whenever we act we are causally determined by factors beyond our control would not change this assumption. Others are natural agnostics about compatibilism and incompatibilism, and for them the moral responsibility assumption would be challenged but not defeated by the prospect of causal determination. Incompatibilists contend that these reactions fail to face up to the implications of the causal determination of our actions by factors beyond our control, and they aspire to an effective strategy for promoting this perspective. One prominent strategy for supporting the incompatibilist’s position draws on an analogy between naturalistic causal determination and intentional causal determination by other agents. Manipulation arguments for the incompatibility of naturalistic causal determination and moral responsibility begin with an imaginary case in which

180    Derk Pereboom and Michael McKenna an agent is intentionally deterministically manipulated by another agent to perform an action. The story specifies that the conditions compatibilists would take to be sufficient for moral responsibility are satisfied. The case is meant to elicit the intuition that, due to the nature of the manipulation, the agent is not morally responsible for what she does. It is then argued that this manipulated agent differs in no responsibility-​relevant respect from agents who are naturalistically determined so to act. And because there is no responsibility-​relevant difference between the intentionally manipulated agents and those who are naturalistically determined, the naturalistically determined agents aren’t morally responsible for the action either. The conclusions drawn include that any normal, non-​manipulated agent who is naturalistically determined to perform an action is not morally responsible for it even if she satisfies all the prominent compatibilist conditions on moral responsibility; the compatibilist conditions are insufficient for moral responsibility; and naturalistic causal determination in general is incompatible with moral responsibility.1

2.  Varieties of Manipulation Argument An early formulation of a manipulation argument is Richard Taylor’s (1974). His argument targets a compatibilist about determinism and free action who follows Thomas Hobbes (1654), David Hume (1739/​1978, 1748/​2000), and A. J. Ayer (1954) in characterizing acting freely as acting without impediment, that is, acting as one desires to act. Taylor’s formulation in terms of free action warrants a comment on the relation between (1) acting freely and (2) moral responsibility for acting. Numbers (1) and (2) can be understood so that it would be open whether acting freely is a necessary condition on moral responsibility, or else so that this is not open. Those who believe that there are successful examples of the sort Frankfurt (1969) devised maintain that one notion of acting freely, acting with the ability to do otherwise, is not necessary for moral responsibility. On a contrasting widespread current convention, acting freely or acting with free will is conceived as acting with whatever control is necessary for moral responsibility. Your coauthors have adopted this convention, for example, in defining compatibilism and incompatibilism and in describing the target of manipulation arguments (e.g., McKenna and Pereboom 2016). The target of Taylor’s argument can be understood in this way. But in this chapter we dispense with the reference to free action and free will, and more simply characterize the target of manipulation arguments as affirming the compatibility of moral responsibility (understood in the basic desert sense) and causal determination in general. 1 

This article covers much of the same ground as Chapter 7 of McKenna and Pereboom (2016) and is indebted to that chapter.

Manipulation Arguments Against Compatibilism    181 Here is Taylor’s (1974) argument: But if the fulfillment of these conditions renders my behavior free—​that is to say, if my behavior satisfies the conditions of free action set forth in the theory of soft determinism—​then my behavior will be no less free if we assume further conditions that are perfectly consistent with those already satisfied. We may suppose further, accordingly, that while my behavior is entirely in accordance with my own volitions, and thus “free” in terms of the conception of freedom we are examining, my volitions themselves are caused. To make this graphic, we can suppose that an ingenious physiologist can induce in me any volition he pleases by pushing various buttons on an instrument to which, let us suppose, I am attached by numerous wires. All of the volitions I have in that situation are, accordingly, precisely the ones he gives me. By pushing one button, he evokes in me the volition to raise my hand; and my hand, being unimpeded, rises in response to that volition. By pushing another, he induces the volition in me to kick, and my foot, being unimpeded, kicks in response to that volition. We can even suppose that the physiologist puts a rifle in my hands, aims it at a passerby, and then, by pushing the proper button, evokes in me the volition to squeeze my finger against the trigger, whereupon the passerby falls dead of a bullet wound. This is the description of a man who is acting in accord with his inner volitions, a man whose body is unimpeded and unconstrained in its motions, these motions being the effects of his inner states. It is hardly the description of a free and responsible agent. It is the perfect description of a puppet. (Taylor 1974: 45)

The compatibilism Taylor was aiming to discredit falls short of the more robust compatibilisms more recently defended, but his strategy is suggestive of how these more recent proposals can be contested. In more contemporary manipulation arguments (e.g., Pereboom 1995, 2001; Mele 2006), an agent performs an immoral action that conforms to all of the prominent compatibilist conditions historically advocated. It’s specified that the agent satisfies those proposed by Hobbes (1654) and Hume (1739/​ 1978, 1748/​2000): Her action is unimpeded, that is, she acts as she wants to act, so that if she had wanted to act otherwise she would have; while the desire that motivates her to act is nevertheless not irresistible for her, and thus she does not act from internal compulsion (Ayer 1954). The agent satisfies the compatibilist condition proposed by Harry Frankfurt (1971): Her effective desire (i.e., her will) to act conforms appropriately to her second-​order desires for which effective desires she will have. She also conforms to the reasons-​responsiveness condition advocated by John Fischer and Mark Ravizza (1998) (cf. Fischer 1982, 1994: the agent’s desires can be modified by, and some of them arise from, her rational consideration of the reasons she has, and if she believed that the bad consequences for herself that would result from her action would be much more severe than she actually thinks them likely to be, she would have refrained from acting for that reason); rationality conditions are advocated by Wolf 1990; Haji 1998; McKenna 2000, 2012; Nelkin 2008, 2011; Sartorio 2016, among others. The agent in addition satisfies the related condition advanced by Jay Wallace (1994): She has the general ability to grasp, apply, and regulate her actions by moral reasons. For instance, when egoistic reasons

182    Derk Pereboom and Michael McKenna that count against acting morally are weak, she will typically regulate her behavior by moral reasons instead. She also has the capacity to reflectively revise and develop her moral character and commitment over time, and for her actions to be governed by these moral commitments, a condition on moral responsibility Al Mele (2006) and Ish Haji (1998) defend. Treating “X” as a variable for some way of intentionally and deterministically manipulating an agent, here is a general schema for a manipulation argument (SMA): 1. Agent S who satisfies the compatibilist conditions on moral responsibility and is manipulated in manner X to perform action A is not morally responsible for performing A (premise, based on intuition). 2. Agent S who satisfies the compatibilist conditions on moral responsibility and is manipulated in manner X to perform action A differs in no respect relevant to moral responsibility from any normal, non-​manipulated agent who satisfies the compatibilist conditions on moral responsibility and is naturalistically determined to perform action A (no-​difference premise). 3. Therefore, any normal, non-​manipulated agent who satisfies the compatibilist conditions on moral responsibility and is naturalistically determined to perform action A is not morally responsible for performing A.2 How might compatibilists respond to SMA? They standardly resist either Premise 1 or Premise 2. To reject Premise 1 is to adopt a hard-​line reply, since it involves the uncompromising resistance to the intuition, essential to the incompatibilist’s case, that the intentionally and deterministically manipulated agent is not morally responsible. Michael McKenna (2008) defends such a hard-​line reply. To reject Premise 2 is to adopt a soft-​ line reply, one that does not resist the intuition that the manipulated agent is not morally responsible, and instead aims to show that an agent who is merely naturalistically determined differs from the manipulated agent in a way relevant to moral responsibility (the terminology is due to McKenna, 2003: 216–​217; 2008a: 143–​144). McKenna, who champions a hard-​line reply, grants that compatibilists might win battles by adopting a soft-​line reply to particular instances of manipulation arguments, which involves proposing a new compatibilist condition not satisfied by the manipulated agent. As Mele (2008) puts it, “even if the [manipulation case] is a counterexample to [the sufficiency for moral responsibility of a set of compatibilist conditions], it does not follow that incompatibilism is true. Perhaps the story is not a counterexample to some superior candidate for being a set of conceptually sufficient conditions for free action and moral responsibility that is consistent with the truth of determinism and has not yet been proposed by compatibilists.” But McKenna argues that a problem for this strategy is that it is open to the incompatibilist to respond with a simple adjustment to her case 2  Pereboom (2001, 2014) appeals to a best explanation in prosecuting his argument, and Mele (2006: 141–​144, 2008) takes issue with Pereboom on this point. Pereboom replies to Mele in (Pereboom 2007 and 2014).

Manipulation Arguments Against Compatibilism    183 of manipulation, one that accommodates the newly proposed compatibilist condition. Thus, in adopting a soft-​line reply, one merely forestalls a revised manipulation argument for which compatibilists, at least eventually, will not have the option of a soft-​ line reply. Mele (2008) distinguishes three different forms of manipulation arguments. The first, which he calls a straight manipulation argument, has the structure of SMA above. The second form, a manipulation argument to the best explanation, replaces the second premise in SMA with a best-​explanation premise such as, “the best explanation that S is not morally responsible when manipulated in manner X to perform A is that S is causally determined to do so.” Pereboom’s argument (1995, 2001, 2014) has this form. The third form, an original design argument of the sort Mele himself has developed (2006), works not from a case of an already existing agent who is manipulated, but from an agent who is originally designed to perform some act, A. Here is his (2008) version: [A goddess] Diana creates a zygote Z in Mary. She combines Z’s atoms as she does because she wants a certain event E to occur thirty years later. From her knowledge of the state of the universe just prior to her creating Z and the laws of nature of her deterministic universe, she deduces that a zygote with precisely Z’s constitution located in Mary will develop into an ideally self-​controlled agent who, in thirty years, will judge, on the basis of rational deliberation, that it is best to A and will A on the basis of that judgment, thereby bringing about E. If this agent, Ernie, has any unsheddable values at the time, they play no role in motivating his A-​ing. Thirty years later, Ernie is a mentally healthy, ideally self-​controlled person who regularly exercises his powers of self-​control and has no relevant compelled or coercively produced attitudes. Furthermore, his beliefs are conducive to informed deliberation about all matters that concern him, and he is a reliable deliberator. So he satisfies a version of my proposed compatibilist sufficient conditions for having freely A. Compare Ernie with Bernie, who also satisfies my compatibilist sufficient conditions for free action. The zygote that developed into Bernie came to be in the normal way. A major challenge for any compatibilist who claims that Ernie A-​s unfreely whereas Bernie A-​s freely is to explain how the difference in the causes of the two zygotes has this consequence. Why should that historical difference matter, given the properties the two agents share?

Pereboom (1995, 2001, 2014) proposes that a manipulation argument acquires more force by virtue of setting out several such cases, the first of which features the most radical sort of manipulation consistent with the proposed compatibilist conditions and with intuitive conditions on agency, each progressively more like a fourth, which the compatibilist might envision to be ordinary and realistic, in which the action is causally determined in a natural way. An additional challenge for the compatibilist is to point out a difference between any two adjacent cases that would show why the agent might be morally responsible in the later example but not in the earlier one. Pereboom argues that this can’t be done: In transitions between cases there is no responsibility-​relevant difference between the modes of manipulation, so that if we are to treat the first case as one in which the manipulated agent is not free and responsible, that will carry over to

184    Derk Pereboom and Michael McKenna the next in the series, and so forth, with the fourth and last merely being a case in which determinism is true. The agent’s non-​responsibility thus generalizes from the first of the manipulation examples to the ordinary case. In his setup (2001, 2014), in each of the four cases Plum decides to kill White for the sake of some personal advantage, and he succeeds in doing so. The four cases exhibit different ways in which Plum’s murder of White might be causally determined by factors beyond his control. In a first type of counterexample (Case 1) to the prominent compatibilist conditions, neuroscientists manipulate Plum in a way that directly affects him at the neural level, but so that his mental states and actions feature the psychological regularities and counterfactual dependencies that are compatible with ordinary agency: Case 1: A team of neuroscientists has the ability to manipulate Plum’s neural states at any time by radio-​like technology. In this particular case, they do so by pressing a button just before he begins to reason about his situation, which they know will produce in him a neural state that realizes a strongly egoistic reasoning process, which the neuroscientists know will deterministically result in his decision to kill White. Plum would not have killed White had the neuroscientists not intervened, since his reasoning would then not have been sufficiently egoistic to produce this decision. But at the same time, Plum’s effective first-​order desire to kill White conforms to his second-​order desires. In addition, the process of deliberation from which the decision results is reasons-​responsive; in particular, this type of process would have resulted in Plum’s refraining from deciding to kill White in certain situations in which his reasons were different. His reasoning is consistent with his character because it is frequently egoistic and sometimes strongly so. Still, it is not in general exclusively egoistic, because he sometimes successfully regulates his behavior by moral reasons, especially when the egoistic reasons are relatively weak. Plum is also not constrained to act as he does, for he does not act because of an irresistible desire—​the neuroscientists do not induce a desire of this sort. (Pereboom 2014)

In Case 1, Plum's action satisfies the compatibilist conditions; but intuitively, he is not morally responsible for his decision. It is causally determined by what the neuroscientists do, which is beyond his control, and this makes it intuitive that he is not morally responsible for it. Consequently, it would seem that these compatibilist conditions are not sufficient for moral responsibility—​even if all are taken together. Case 2: Plum is just like an ordinary human being, except that a team of neuroscientists programmed him at the beginning of his life. The neural realization of his reasoning process and of his decision is exactly the same as it is in Case 1 (although their causal histories are different).

Although Plum satisfies all the prominent compatibilist conditions, intuitively he is not morally responsible for his decision. So Case 2 also shows that these compatibilist conditions, either individually or in conjunction, are not sufficient for moral responsibility. Moreover, it would seem unprincipled to claim that here, by contrast with Case

Manipulation Arguments Against Compatibilism    185 1, Plum is morally responsible because the length of time between the programming and his decision is now great enough. Whether the programming occurs a few seconds before or 40 years prior to the action seems irrelevant to the question of his moral responsibility. Causal determination by what the neuroscientists do, which is beyond his control, plausibly explains Plum’s not being morally responsible in the first case, and I think we are forced to say that he is not morally responsible in the second case for the same reason. Case 3: Plum is an ordinary human being, except that the training practices of his community causally determined the nature of his deliberative reasoning processes so that they are frequently but not exclusively rationally egoistic (the resulting nature of his deliberative reasoning processes are exactly as they are in Cases 1 and 2).

For the compatibilist to argue successfully that Plum is morally responsible in Case 3, he must adduce a feature of these circumstances that would explain why he is morally responsible here but not in Case 2. It seems there is no such feature. Causal determination by what the controlling agents do, which is beyond Plum’s control, most plausibly explains the absence of moral responsibility in Case 2, and we should conclude that he is not morally responsible in Case 3 for the same reason. Case 4: Everything that happens in our universe is causally determined by virtue of its past states together with the laws of nature. Plum is an ordinary human being, raised in normal circumstances, and again his reasoning processes are frequently but not exclusively egoistic, and sometimes strongly so (as in Cases 1–​3). His decision to kill White issues from his strongly egoistic but reasons-​responsive process of deliberation, and he has the specified first and second-​order desires. The neural realization of Plum’s reasoning process and decision is exactly as it is in Cases 1–​3; he has the general ability to grasp, apply, and regulate his actions by moral reasons, and it is not because of an irresistible desire that he decides to kill.

Given that we are constrained to deny moral responsibility in Case 3, could Plum be responsible in this ordinary deterministic situation? It appears that there are no differences between Case 3 and Case 4 that would justify the claim that Plum is not responsible in Case 3 but is in Case 4. In both of these cases Plum satisfies the compatibilist conditions on responsibility. In each the neural realization of his reasoning process and decision is the same, although the causes differ. From this Pereboom argues that we can conclude that causal determination by other agents was not essential to what was driving the intuition of non-​responsibility in the earlier cases. Because it’s intuitive that Plum is not morally responsible in Case 1, and there are no differences between Cases 1 and 2, 2 and 3, and 3 and 4 that can ground his non-​responsibility in the former of each pair but not in the latter, we are led to the conclusion that he is not responsible in Case 4. Pereboom conceives this as an argument to the best explanation. In his view, the salient common factor that best explains the judgment that Plum is not responsible in any of the cases, a judgment that on reflection we

186    Derk Pereboom and Michael McKenna are led to make, is that in each case he is causally determined by factors beyond his control to decide as he does.

3.  The Dialectic of Manipulation Arguments An important feature of the dialectic of manipulation arguments is illuminated by Patrick Todd’s reply (2013) to an objection by Stephen Kearns (2012), an objection that explicitly targets Mele’s original design argument, although it also applies to other versions. Kearns’s objection is that if in a manipulation case it’s the manipulation that is supposed to make it intuitive that the agent is not responsible, the reason for the non-​ responsibility verdict does not transfer to the ordinary case, since the ordinary case involves no manipulation. But if manipulation does not have the role of making it intuitive that the agent is not responsible, then it should be possible to start the argument with a deterministic case that does not feature manipulation, and this would render the manipulation examples unnecessary. In response, Todd points out that . . . the proponent of the argument contends—​and clearly must contend—​that the manipulation is irrelevant as concerns what makes the agent unfree. She instead says that the manipulation can help us see that something does make the agent unfree. In other words, she first presents the scenario (say) to an agnostic, and asks whether the agnostic thinks that the agent is free (or responsible) in that scenario. And suppose the agnostic says “no.” She then points out that whatever would make the agent unfree in that scenario would also make the agent unfree in a qualitatively identical scenario, except in which blind natural causes have taken the place of an intentional agent. (Todd 2013: 202)

In accord with Todd’s assessment, here is an account of the dialectic. Incompatibilists believe that compatibilists fail adequately to face up to the implications of causal determination. The way manipulation arguments aim to remedy this putative shortcoming is by first presenting a deterministic manipulation case with the hope that it will be more successful at eliciting a non-​responsibility intuition than causal determination alone does. The next step is to argue that non-​responsibility is preserved even when the manipulation is subtracted, on the ground there is no responsibility-​relevant difference between the deterministic case that features manipulation and one that doesn’t. The salient common element is causal determination by factors beyond the agent’s control, and this feature will therefore be sufficient for non-​responsibility. A deterministic manipulation case, rather than being unnecessary, would thus serve to reveal that ordinary causally determined agents lack the sort of free will at issue. A compatibilist might deny having any intuition that an agent in a manipulation case is not morally responsible. If someone has this unconflicted intuition even upon

Manipulation Arguments Against Compatibilism    187 reflection, then this manipulation argument will not have the power to persuade him. But as Haji says, “if anything is clear, the literature reveals that targeted compatibilists have, generally, not taken the Four-​Case Argument to be toothless. . . .” (Haji 2009: 127). He also points out that there is no consensus among compatibilists about the conditions that delineate when a manipulated agent is responsible, and that this lack of agreement results in the manipulation argument having a certain allure (Haji 2009: 126–​127). A response that commands agreement among compatibilists hasn’t emerged, which perhaps indicates that compatibilists have intuitions about manipulation cases that prove difficult to systematize in a compatibilist theory. Todd (2011) makes the case that proponents of manipulation arguments have assumed too heavy a burden: They do not need to make it plausible that manipulated agents are not morally responsible, but only that their responsibility is mitigated. This is because compatibilists will have as difficult a time accounting for mitigated responsibility in a manipulation case as they would have explaining non-​responsibility. So if the compatibilist were to agree that Plum’s blameworthiness in Case 1 is mitigated, what would, on the compatibilist view, account for this? If it is intentional deterministic manipulation, then because there is no relevant difference between intentional deterministic manipulation and natural determination, the compatibilist’s position would be compromised. However, both Justin Capes (2013) and Hannah Tierney (2013) have argued that compatibilism is consistent with affirming that determinism can mitigate responsibility, since it may allow for forms of moral responsibility incompatible with manipulation while preserving the claim that causal determination is compatible with the key sort of moral responsibility. Andrew Khoury’s (2014) reply to Todd points out that implicit in mitigation is a concession that the manipulated agent is responsible to some degree (for reservations about Khoury’s strategy, see Tierney 2014).

4.  Three Soft-​L ine Replies 4.1  Relevant Difference: Causal Determination by Other Agents One prominent soft-​line reply to the manipulation argument is that a feature of an ordinary naturalistic deterministic case (e.g., Case 4 in Pereboom’s argument) that amounts to a responsibility-​relevant difference from manipulation cases is that in the ordinary case the causal determination is not brought about by other agents (Lycan 1987). The key soft-​line claim is that what is generating the non-​responsibility intuition in manipulation cases is not causal determination per se but causal determination by other agents. Adam Feltz (2013), as well as Dylan Murray and Tania Lombrozo (2017), have tested this suggestion in empirical studies. Feltz found judgments of diminished moral

188    Derk Pereboom and Michael McKenna responsibility in cases of causal determination by other agents relative to naturalistic determination. But only if the manipulation by other agents was intentional and direct, as in Case 1, did his subjects, on average, fall below the midpoint between “strongly agree” and “strongly disagree.” On Murray and Lombrozo’s interpretation, their results show that intentional control by other agents robustly generates intuitions of absence of responsibility and free will, while causal determination per se does not do so. They conclude that because causal determination per se does not robustly generate non-​ responsibility intuitions, drawing a parallel to manipulation by agents fails to support incompatibilism. However, the incompatibilist in fact predicts that immediate assessments about responsibility will generally differ between Case 1 and Case 4, but maintains that at this point a second phase of the argument becomes pertinent: a request to explain the differences in intuition in a principled way. Should intentional direct manipulation by other agents make a difference relative to natural causal determination in assessing moral responsibility in the basic desert sense? It would be valuable to survey respondents taking these considerations into account. Subjects might be challenged to explain differential judgments across the cases, and then tested to see whether their judgments about the individual cases change as a result. The incompatibilist might also object that despite the reactions of the subjects, the difference between intentional manipulation by other agents and naturalistic determination might still be irrelevant to moral responsibility in the basic desert sense. One might test this hypothesis by having subjects imagine further cases that are exactly the same as Case 1 or Case 2, except that the states at issue are instead produced by a spontaneously generated machine—​a machine specified to have no intelligent designer (Pereboom 2001) or a force field (Mele 2006). However, it’s hard to separate the idea of a sophisticated machine from intelligent designers of that machine, even if it’s specified that the machine is spontaneously generated. The mechanism by which a force field manipulates may be too unclear, and it might well suggest bypassing to at least some subjects. In response to these concerns, Gunnar Björnsson constructed a scenario where all the prominent compatibilist conditions on moral responsibility are satisfied but in which a cause that isn’t an agent—​an infection—​slowly renders the agent in the example increasingly egoistic without bypassing or undermining his agential capacities. Björnsson predicted that if subjects were prompted to view the agent’s behavior as dependent on this non-​agential cause, attributions of responsibility would be undermined to roughly the same extent as in cases of an intentional manipulator. This turned out to be true: In a study involving 416 subjects, the infection undermined attributions of free will and moral responsibility to the same degree as in cases of intentional manipulation (Björnsson and Pereboom 2016). Björnsson’s study suggests that incompatibilists might be able to employ the generalization strategy used in manipulation arguments without introducing intentional manipulation, and thus avoiding the experimentally driven objections developed by Feltz, Murray, and Lombrozo.

Manipulation Arguments Against Compatibilism    189

4.2  Relevant Difference: No Agency in Case 1 A different soft-​line reply to Pereboom’s argument has been proposed by a number of soft-​line critics (Fischer 2004; Mele 2005; Baker 2006; Demetriou 2010) of Pereboom’s argument. In his Case 1, Plum was created by a team of neuroscientists who can manipulate him directly through radio-​like technology, but he is as much like an ordinary human being as is possible, given his history. The scientists locally manipulate him from moment to moment, by pushing buttons on their device. In this particular situation, by doing so they cause him to undertake an egoistic process of reasoning which deterministically leads to his killing of White (Pereboom 2001: 112–​113). The concern raised by the critics is that in this first case Plum’s mental states lack the causal cohesion required for him to be a genuine agent. The specific worry raised by Demetriou (2010) is that when the scientists manipulate Plum at some particular instant, they suppress the causal efficacy of his prior mental states, thereby making it the case that Plum’s mental states are causally isolated from one another (Demetriou 2010: 608; cf. Matheson 2016: 1972). While Plum in Case 1 and Plum in Case 4 have the same mental states, these states differ radically with regard to how they are causally connected. This difference yields a difference in agency and, potentially, in moral responsibility. In response, Pereboom (2014) embellished the description of Case 1 so that the neuroscientists manipulate him only sporadically. In this particular instance they (the team of neuroscientists) do so by pressing a particular button just before he begins to reason about his situation, which they know will produce in him a neural state that realizes a strongly egoistic reasoning process, which the neuroscientists know will deterministically result in his decision to kill White (as suggested by Shabo 2010: 376). Pereboom suggests that in ordinary life, external influences, such as finding out that the home team lost, have a similar sort of effect, and in such situations, agency is clearly preserved. In embellished Case 1, while agency is preserved, Plum is intuitively not morally responsible since had the neuroscientists not intervened, he would not have killed White—​or at least so Pereboom contends (Pereboom 2014: 76). McKenna criticized this embellished case on the ground that it is much easier to accept that Plum is morally responsible here than it is for the version on which he is constantly being manipulated. This stands to reason, but Pereboom believes that the embellished case is strong enough. Imagine that Plum is on trial by jury, and that it was revealed that the neuroscientists intervened as they did, and that he would not have killed White had it not been for their intervention. What jury would convict him? McKenna (2008, 2014) contends that a better version of Case 1 features regular local manipulation while retaining the causal integration of Plum’s states. Plum in Case 1 so acts as a result of numerous “micro” interventions that “steer” Plum in certain directions rather than others while leaving Plum’s identity and agency intact.3 Benjamin Matheson (2016) develops a 3 

These interventions need be no more invasive than those due to changes in hormone levels or fluctuations in blood sugar.

190    Derk Pereboom and Michael McKenna detailed account of how this might work, specifically with Demetriou’s critique in mind. A key feature of his account is this: At times we are motivated to perform an action, but not sufficiently strongly enough to do so without someone else’s badgering. I may want to reorganize the basement, but I won’t do it unless my wife presses me to do it. In this example, my earlier mental states are not suppressed, but rather their motivational efficacy is enhanced, and my agency is clearly retained. So similarly, the neuroscientists, rather than suppressing the efficacy of all of Plum’s relevant mental states, strengthen some of them and diminish the efficacy of others, and he retains his agency. These proposals for defending Pereboom’s manipulation Case 1—​ Pereboom’s, McKenna’s, and Matheson’s—​lend support to McKenna’s contention about soft-​line replies: They are unstable because a slight modification by her opponent to the disputed example can thwart the reply. Such a modification argument simply builds in the compatibilist condition that in an earlier iteration the soft-​line critic found to be lacking. This suggests that for compatibilists to effectively respond to well-​executed versions of manipulation arguments, they should be prepared to defend a hard-​line reply. In this case, it would seem that they must be prepared to resist Pereboom’s plausible claim that in Case 1 Plum does not act freely and is not morally responsible.

4.3  Relevant Difference: The Manipulated Agent Is Not the Causal Source of the Action Interventionist theories of causation claim that one variable is the cause of another just in case an intervention on the first would lead to a change on the second.4 Several compatibilists, including Oisin Deery and Eddy Nahmias (2017) and Marius Usher (2020) have used this theory of causation to design responses to manipulation arguments. Here we focus on Deery and Nahmias’s version. The manipulation argument they criticize features two contrasting cases devised by Al Mele (2013): First, imagine Danny. One evening in 1986, Danny’s parents made love, hoping to conceive a child. They got lucky. A zygote was formed (at time t1), and nine months later Danny was born. Thirty years later, Danny is walking down a deserted street and he finds a wallet with the owner’s ID in it and $500. Danny takes himself to have good reasons for keeping the money, but also for returning the wallet. He deliberates for a while, and in the end he decides to keep the money, and he does so (at time t30). Assume that this occurs in a deterministic universe—​that is, a universe in which, for each event E, the laws of nature and some set of events that occurred prior to E are such that these events cause E to occur with probability 1. If determinism is true, then some 4  This view has been developed and defended by Judea Pearl (2000, 2009) in computer science and James Woodward (2003, 2006) in philosophy.

Manipulation Arguments Against Compatibilism    191 set of events prior to Danny’s act of stealing the wallet at t30 are (together with the laws) such that they cause his deliberating and acting in that way, at that time, with probability. (Deery and Nahmias 2017: 1257)

Contrast this with the following variation: . . . a powerful Goddess, Diana, has the power to know what will happen in the future and to act in ways that ensure that specific events occur in the distant future. Diana has these abilities in part because she exists in a deterministic universe and is able to get enough information about events occurring in it (e.g., at t1) to deduce exactly what she needs to do at that time to ensure that a particular event occurs thirty years later. In this case, Diana assembles atoms in a specific way at t1 so as to create a zygote that develops into a child, grows up, finds a wallet thirty years later, and at t30 decides to keep the money it contains. For some reason, Diana wants to ensure that this event occurs at t30, and she possesses the power to alter events at t1 precisely so that she ensures that it does occur. As it turns out, the life of this intentionally created person (whom we will call Manny since he is Man-​ipulated) follows the exact same course as the life of deterministic Danny (as described above). (Deery and Nahmias 2017: 1257)

Advocates of manipulation arguments contend that Manny is not morally responsible for stealing the money—​they have the non-​responsibility intuition in the manipulation case. They also stress that Manny and Danny each satisfy the compatibilist sufficient conditions for free will and moral responsibility, and contend that as a result there is no responsibility-​relevant difference between Manny and Danny. On an interventionist approach to causation, both Manny’s and Danny’s compatibilist agential structure, that is, their intentions and powers to steal, conceived as compatible with causal determinism, count as actual causes of their stealing the money. Changes in the value of the variables representing these intentions and powers result in changes to the value of the variables representing their stealing the money. But Deery and Nahmias argue that there is a relevant difference between how the intentions and powers cause their actions in the two cases. While Danny’s intention and powers are the causal source of his action, Manny’s are not, and Manny can’t be free and morally responsible with regard to an action if he, by virtue of his intention and powers, isn’t its causal source (Deery and Nahmias 2017: 1267) Deery and Nahmias contend that for a variable to represent the causal source of an event, that variable must bear the strongest causal invariance relation to that event among all the variables that are causally related to it, and, a causal invariance relation, R1, that obtains between two causal variables, X and Y, is stronger than another such relation, R2, obtaining between Y and another of its prior causal variables—​for instance, W—​if:

192    Derk Pereboom and Michael McKenna (1) holding fixed the relevant background conditions, C, R1 predicts the value of Y under a wider range of interventions on X than R2 does under interventions on W; and (2) R1 predicts the value of Y across a wider range of relevant changes to the values of C than R2 does. (Deery and Nahmias 2017: 1262–​1263) Danny’s intention and powers are the causal source of his stealing the money because changes in the value of the variable representing Danny’s intention and powers to steal issue in the same action across many changes in the background conditions, and there is greater invariance for this variable than for any competing variable. However—​and this is the crucial point—​Deery and Nahmias contend that there is a stronger causal invariance relation between the variable representing Diana’s intention and powers to manipulate and the one representing Manny’s stealing than between the variables representing Manny’s intention and powers and his stealing. Diana intends for Manny to steal the money at t30 and she is able to ensure that he does so. There is only a limited variety of changes in the background circumstances that would break the causal connection between the variables representing Diana’s intention and Manny stealing the money. By contrast, there are comparatively more changes in the background circumstances that would break the link between the variables representing Manny’s intention and powers and his stealing the money. Consequently, Diana’s intention and powers relate more invariantly to Manny’s stealing the money than do Manny’s intention and powers, and for this reason Diana is the causal source of the action, and not Manny. In response, one factor that Deery and Nahmias do not take into account is that in constructing a manipulation case, both the nature of the manipulation and the nature of the manipulated agent can be made to vary, with the result that the two can be made to match in degrees and kinds of invariance. Suppose Diana is quite powerful, but not omniscient or omnipotent, but due to her impending death, she has only one shot at manipulating Manny. She does what’s specified in Deery and Nahmias description, but due to lack of information, she believes she has only a small chance of succeeding. But what she does succeeds in is causally determining Manny to steal. Let’s imagine that the other possibilities for manipulation she was considering would have failed. In this case she can’t ensure that Manny steals, but our intuition that Manny isn’t responsible is still in place. However, in this example, it’s false that there are very few changes in background circumstances that could break the causal connection between Diana’s intention and Manny stealing the money. Moreover, Manny’s abilities can be specified so that the degrees of invariance relative to his stealing the money match hers precisely. In this scenario, it’s false, on the interventionist approach, that Manny is not the causal source of his stealing. The following more vivid case might make the point even clearer. Diana, as it happens, has considerable powers to affect things from learning details about the laws of nature and the state of the world at a time. But she has no knowledge of these things at all, and for the most part she cannot gain access to such knowledge. Instead, she works for Zeus

Manipulation Arguments Against Compatibilism    193 as a clerk doing mostly menial tasks. For whatever reason, Zeus has a very odd system of managing all his knowledge of the state of the world and how any possible future will unfold. He keeps every such sequence of possible ways a total history of the world will unfold recorded in a set of sheets of paper. Each such sequence is printed out (in astoundingly tiny print) on single pages. He also carries them all about in a big manila folder, maybe to show off to everyone how much he knows. One day, Zeus is walking down the hallway near her office, and one piece of paper happens to fall from his folder without his noticing it. Diana picks it up, and it is a sequence that, if actualized, will result in Manny’s misdeed. Out of nothing but perverse interest, like the desire to see the fall of an arrangement of dominos, Diana decides it would be amusing to witness Manny perform this moral wrong as an upshot of this one way the world might unfold. Based on this one-​off bit of knowledge that she pilfered from Zeus, she intervenes in the unfolding of the world to get the desired result. Manny steals the money at t30. Here it would seem that Manny is the cause of his stealing by interventionist standards, and that the intuition that he is not responsible would be widespread. Here is a resulting lesson for constructing manipulation arguments (Pereboom 2014): set up the cases so as to secure as close as possible a causal match between the manipulation cases and the natural deterministic case, while preserving the intuition of non-​responsibility in the manipulation cases. Applying this rule, as in the cases as just described, the resulting manipulation matches natural determination better than the Mele cases as Deery and Nahmias interpret them.5 In fairness to Deery and Nahmias, it bears mentioning that weakening a manipulator’s invariance as in the above cases may also weaken intuitions of non-​ responsibility. Indeed, by Deery and Nahmias’s standards, a compatibilist might protest that in these cases Manny remains the causal source of his wrongdoing, not Diana. So, the compatibilist might contend that Manny’s moral responsibility is not compromised. However, a proponent of the manipulation argument will disagree. She can grant that the intuition of non-​responsibility in cases of this sort might be weaker than it is in a case wherein the manipulator is strongly invariant and so is the causal source of the manipulated agent’s targeted wrongdoing. Nevertheless, she can argue, it is strong enough to elicit a judgment of non-​responsibility in an audience of neutral inquirers. Still, in our estimation, it counts in favor of Deery and Nahmias’s defense of compatibilism that they have put pressure on the incompatibilist to revise her argument in a way that results in a somewhat less compelling example of manipulation as it figures in the first premise of SMA.

5  A different sort of objection to this strategy is raised by Hannah Tierney and David Glick (2020). They criticize Deery and Nahmias’s analysis of sourcehood by drawing a distinction between two forms of causal invariance that can come into conflict on their account. They conclude that any attempt to resolve this conflict will either result in counterintuitive attributions of moral responsibility or will undermine Deery and Nahmias’s response to manipulation arguments.

194    Derk Pereboom and Michael McKenna

5.  A Hard-​L ine Reply What if the compatibilist claims that it’s intuitive that Plum is morally responsible in Case 4, and because there are no responsibility-​relevant differences between adjacent cases, the verdict of responsibility carries backward through to Case 1? Then the compatibilist and the incompatibilist would be at a standoff, with the manipulation argument yielding progress for the incompatibilist. However, for the compatibilist to claim that it’s intuitive that Plum is in fact morally responsible in Case 4 would be overly ambitious, since that’s what’s fundamentally at issue in the debate. McKenna’s (2008) hard-​line reply proposes instead that an agnostic stance about Case 4 is initially reasonable, and that this agnostic stance transfers to Case 1 unimpeded due to the absence of responsibility-​relevant differences between adjacent cases. Credible doubt is accordingly cast on the claim that Plum is not morally responsible in Case 1. In this way, McKenna addresses the multiple-​case argument with a less ambitious but more compelling hard-​line strategy. In developing this strategy, McKenna (2008a, 2014) first invites us to consider pertinent agential and moral properties of Plum as displayed in Case 4, in which Plum’s action is naturalistically determined. 6 There, Plum is just as much like an ordinary moral agent as any one of us. He is capable of feeling remorse, or instead ambivalence about killing White, or even pride in that accomplishment. We may in addition imagine that Plum has a history of moral development from childhood into adulthood like that of any normal person. We can suppose that Plum features the emotional complexity that Strawsonians highlight—​he is capable of moral resentment and gratitude, indignation and approval, guilt and pride, and he is sensitive to the significance of these emotions in his relations with others. Plum is, in consequence, a full member of the moral community, one whose standing enables the personal relationships characteristic of fulfilled human lives. At this point, McKenna argues that if all of this is true of Plum in Case 4, it can equally as well be true of Plum in Case 1. In his view, the point of attending to these details is to help elicit intuitions that are friendly to compatibilists—​to elicit the sense that Plum is in Case 1 a complex agent just like any normal person we might encounter. Sofia Jeppsson (2019) reinforces this part of the strategy by arguing that in assessing manipulated Plum’s moral responsibility, taking on his agential and moral perspective is the natural and the morally obligatory one for ordinary, everyday interaction, and that an incompatibilist who wants to claim that we ought to make moral responsibility judgments from a detached causal perspective therefore owes us an argument for this claim. Next, McKenna has us consider the initial claims the compatibilist is entitled to make regarding Plum in Case 4. She is not entitled to insist that the incompatibilist grant that here Plum is morally responsible. Rather, she can legitimately expect that a neutral 6 

This strategy, as recounted here, is most completely spelled out in McKenna (2014).

Manipulation Arguments Against Compatibilism    195 inquirer, who is undecided whether determinism rules out moral responsibility, initially allow that it’s open that Plum is morally responsible in Case 4 (and that it’s open that he isn’t). McKenna then points out that incompatibilists may not assume at the outset that Plum in Case 4 is not morally responsible, nor should they expect that the open-​minded neutral inquirer’s intuition about Plum in Case 4 is that he is not morally responsible. Again, that is what is in dispute, and assuming this at the outset would be question-​ begging. Thus, the stance toward Case 4 that an incompatibilist should take is the same as the one that a compatibilist should take. Both must initially allow that in Case 4 it is open whether Plum is morally responsible for killing White (and that it’s open that he’s not). With this in mind, McKenna attempts to turn Pereboom’s strategy against him. Pereboom,7 assuming an intuition of non-​responsibility in Case 1, marches through to his Case 4, drawing an incompatibilist conclusion. He grants that Pereboom does gain some intuitive advantage in favor of incompatibilism by beginning with the non-​ responsibility intuition in Case 1 and having this verdict transmit through to Case 4. That has to be weighed on the scales when evaluating the overall force of the argument. But McKenna proposes that the compatibilist now move in the opposite direction, and that the result also be weighed on the scales. It’s open that Plum in Case 4 is morally responsible; there are no responsibility-​relevant differences between adjacent cases; we can thus conclude that it’s open that Plum is morally responsible in Case 1. Finally, McKenna aims to mitigate the potentially jarring effect of claiming that it’s open that agents in manipulation cases are morally responsible by drawing our attention to actual, fairly common events in human life similar in key respects to manipulation cases. In such actual cases, McKenna contends, our intuitions do not favor an incompatibilist diagnosis. Nomy Arpaly suggests that there are many kinds of circumstances in which agents undergo dramatic changes that affect how they act, such as a religious conversion, the birth of a child, a traumatic accident, or the death of a loved one (Arpaly 2006: 112–​113). If such changes leave the agent sane and rational, allowing satisfaction of the compatibilist conditions on moral responsibility, few would be inclined to judge that they undermine moral responsibility. McKenna suggests that reflection on the similarity of such actual circumstances to manipulation adds credence to his compatibilist strategy. Pereboom (2008, 2014) responds that from the point of view of a neutral party, one not initially committed to compatibilism or to incompatibilism, it is initially epistemically rational not to believe that the agent in Case 4 is morally responsible, and not to believe that he isn’t, but to be open to clarifying considerations that would make one or the other of these beliefs rational. This neutral inquiring stance differs from the stance of the confirmed agnostic, who maintains that it is not clearly the case that the ordinary determined agent is morally responsible, and that it is not clearly the case that he isn’t, while it’s rational to consider inquiry into the issue closed, and thus not open to further clarifying

7 

See also Sekatskaya (2019).

196    Derk Pereboom and Michael McKenna considerations. If the confirmed agnostic’s stance were rational, McKenna’s conclusion could be generated. But Pereboom argues that on the neutral inquiring stance, a manipulation case can function as a clarifying consideration that makes rational the belief that the ordinary determined agent is not morally responsible. In marching backward through the cases from Case 4, the intuition that it’s open that Plum is morally responsible might well not survive the clarifying considerations that the manipulation cases provide. As Pereboom sees it, intuitions in Cases 1–​3 are friendly to incompatibilism, and these cases provide the right sort of analogy to Case 4 to help clarify the rationality of the claim it’s open that Plum is morally responsible. This is so, even granting that attention to the agential and moral properties of Plum in Case 1, carrying over, as McKenna recommends, from Case 4, do lend some weight to a compatibilist diagnosis.8 The dispute between Pereboom and McKenna—​that is, the dispute between us, your coauthors—​appears to hinge primarily upon these two interrelated questions: In light of cases like Case 1 and Case 2, has the neutral inquiring audience been given sufficient reason to move from open-​minded agnosticism toward an incompatibilist verdict? Or is McKenna’s hard-​line strategy sufficient to establish that it remains rational to maintain that it is open whether determined agents are morally responsible?

6.  Manipulation and Praiseworthiness Manipulation arguments typically focus on immoral actions, on agents who would be blameworthy rather than praiseworthy. Dana Nelkin (2011), who together with Susan Wolf (1990) endorses an asymmetry thesis that access to alternative possibilities is required for blameworthiness but not for praiseworthiness, asks us to consider a manipulated agent, like Plum in Case 1, who does the right thing for the right reasons: The neuroscientists have created Plum so that he fully understands and recognizes good reasons for acting. He is moved by people in distress and desires to help them so as to relieve their suffering. Suppose that one day he finds himself in a situation in which he can help a child only at great risk to himself. He thinks about the relevant considerations and decides that all things considered, helping is the right thing to do and resolves to do it. (Nelkin 2011: 56–​57)

Nelkin reports that for her the intuitive force of this example seems to count in favor of the appropriateness of praise. Pereboom (2014, 102–​103) agrees, but suggests that praise can at times be an expression of approbation or delight about the goodness of the actions

8 

See Daniel Haas (2013) and Sekatskaya (2019) for defenses of McKenna on this point, and Pereboom (2014) for a reply to Haas.

Manipulation Arguments Against Compatibilism    197 or accomplishments of another, which would be appropriate even if it is not basically deserved. Nevertheless, intuitions about this case would appear to provide support for the asymmetry thesis that Wolf and Nelkin defend.

7.  Final Words One last question is whether the sorts of real-​life situations Arpaly and McKenna adduce should carry greater weight than the artificial manipulation cases in assessing the effect of external causes on moral responsibility. McKenna (2008a, 2014) contends that we should assign greater value to real-​life cases; Pereboom (2008, 2014) believes we should assign greater value to the artificial cases. The Spinozistic concern he invokes at this point is that in ordinary life such judgments will have been shaped by ignorance of hidden deterministic causes (Spinoza 1677/​1985), which are brought to the fore in manipulation cases. Manipulation cases may be artificial, but the artificiality is required to make the deterministic causation salient. Dana Nelkin concurs, and suggests that “one might argue that their unrealistic quality helps ensure that we are focused on the stipulated features, and that we aren’t implicitly but unconsciously relying on background assumptions that we bring to ordinary life. In this way, the intuitions are arguably more reliable than the real-​life ones” (Nelkin 2013: 125). At some time in the near future, remote local manipulation, as in featured in Case 1, may become actual. At a Material Research Society Conference in 2019, a Cornell team presented a paper that describes “the fabrication and characterization of wireless, implantable devices for the electrical stimulation of neural tissue.” The background research focuses on a neuronal disorder in mice that leads to obesity, in which neurons active in indicating that enough food has been ingested aren’t adequately stimulated. The implanted device, activated by remote control, provides a fix by causing the requisite neurons to fire.9 When such devices are implanted in human beings, and are used to induce morally salient behavior, we’ll have the opportunity to test our intuitions in actual, and not merely fictional cases.

References Arpaly, Nomy. (2006). Merit, Meaning, and Human Bondage, Princeton, NJ: Princeton University Press. Ayer, Alfred J. (1954). “Freedom and Necessity,” in A. J. Philosophical Essays, A. J. Ayer, ed., London: Macmillan, pp. 271–​284. Baker, Lynne R. (2006). “Moral Responsibility Without Libertarianism,” Noûs 40, pp. 307–​330.

9 

https://​mrsspring2019.zerista.com/​event/​member/​557804

198    Derk Pereboom and Michael McKenna Björnsson, Gunnar, and Derk Pereboom. (2016). “Traditional and Experimental Approaches to Free Will,” in The Blackwell Companion to Experimental Philosophy, Wesley Buckwalter and Justin Sytsma, eds., Oxford: Blackwell Publishers, pp. 142–​157. Capes, Justin. 2013. “Mitigating Soft Compatibilism,” Philosophy and Phenomenological Research 87: pp. 640–​663. Deery, Oisín, and Eddy Nahmias (2017). “Defeating Manipulation Arguments: Interventionist Causation and Compatibilist Sourcehood,” Philosophical Studies 174(5), pp. 1255–​1276. Demetriou, Kristin. (2010). “The Soft-​Line Solution to Pereboom’s Four-​Case Argument,” Australasian Journal of Philosophy 88, pp. 595–​617. Feinberg, Joel. (1970). “Justice and Personal Desert,” in Doing and Deserving, Princeton, NJ: Princeton University Press. Feltz, Adam (2013). Pereboom and premises: Asking the right questions in the experimental philosophy of free will. Consciousness and Cognition 22 (1), 53–​63. Fischer, John Martin. (1982). “Responsibility and Control,” Journal of Philosophy 79, pp. 24–​40. Fischer, John Martin. (1994). The Metaphysics of Free Will, Oxford: Blackwell Publishers. Fischer, John Martin. (2004). “Responsibility and Manipulation,” The Journal of Ethics 8, pp. 145–​177. Fischer, John Martin, and Mark Ravizza. (1998). Responsibility and Control: A Theory of Moral Responsibility, Cambridge: Cambridge University Press. Frankfurt, Harry. (1969). “Alternate Possibilities and Moral Responsibility,” Journal of Philosophy 66, pp. 829–​839. Frankfurt, Harry. (1971). “Freedom of the Will and the Concept of a Person,” Journal of Philosophy 68, pp. 5–​20. Haas, Daniel. (2013). “In Defense of Hard-​Line Replies to the Multiple-​Case Manipulation Argument,” Philosophical Studies 163, pp. 797–​811. Haji, Ishtiyaque. (1998). Moral Accountability, New York: Oxford University Press. Haji, Ishtiyaque (2009). Demotivating semi-​compatibilism. Ideas Y Valores 58 (141), 125–​140. Hobbes, Thomas. (1654). Of Libertie and Necessity: A treatise, wherein all controversie concerning predestination, election, free-​will, grace, merits, reprobation, &c., is fully decided and cleared, in answer to a treatise written by the Bishop of London-​Derry, on the same subject, London, printed by W. B. for F. Eaglesfield. Hume, David. (1739/​1978). A Treatise of Human Nature, Oxford: Oxford University Press. Hume, David. (1748/​2000). An Enquiry Concerning Human Understanding, Oxford: Oxford University Press. Jeppsson, Sofia. (2019). “The Agential Perspective: A Hard-​ Line Reply to the Four-​ Case Manipulation Argument,” Philosophical Studies, 1–​17. https://​doi.org/​10.1007/​ s11098-​019-​01292-​2 Kane, Robert. (1996). The Significance of Free Will, New York: Oxford University Press. Kearns, Stephen. (2012). “Aborting the Zygote Argument,” Philosophical Studies 160(3), pp. 379–​389. Khoury, Andrew (2014). “Manipulation and Mitigation,” Philosophical Studies 168(1), pp. 283–​294. Lycan, William. G. (1987). Consciousness, Cambridge, MA: MIT Press University Press. Matheson, Benjamin (2016). In defence of the Four-​Case Argument. Philosophical Studies 173 (7), 1963–​1982. McKenna, M. (2000). “Assessing Reasons-​Responsive Compatibilism,” International Journal of Philosophical Studies 8, pp. 89–​114.

Manipulation Arguments Against Compatibilism    199 McKenna, Michael (2003). Robustness, control, and the demand for morally significant alternatives: Frankfurt examples with oodles and oodles of alternatives. In David Widerker & Michael McKenna (eds.), Moral Responsibility and Alternative Possibilities: Essays on the Importance of Alternative Possibilities. Ashgate. pp. 201–​217. McKenna, Michael. (2008). “A Hard-​ Line Reply to Pereboom’s Four-​ Case Argument,” Philosophy and Phenomenological Research 77, pp. 142–​159. McKenna, Michael. (2012). Conversation and Responsibility, New York: Oxford University Press. McKenna, Michael. (2014). “Resisting the Manipulation Argument: A Hard-​Liner Takes It on the Chin,” Philosophy and Phenomenological Research 89: 467–​484. McKenna, Michael, and Derk Pereboom. (2016). Free Will: A Contemporary Introduction, London: Routledge, 2016. Mele, A. R. (2005). A critique of Pereboom's 'four-​case argument' for incompatibilism. Analysis 65 (1), 75–​80. Mele, Alfred. (2006) Free Will and Luck, New York: Oxford University Press. Mele, Alfred. (2008). “Manipulation, Compatibilism, and Moral Responsibility,” The Journal of Ethics 12, pp. 263–​286. Mele, Alfred R. (2013). Manipulation, Moral Responsibility, and Bullet Biting. The Journal of Ethics 17 (3), 167–​184. Murray, Dylan & Lombrozo, Tania (2017). Effects of Manipulation on Attributions of Causation, Free Will, and Moral Responsibility. Cognitive Science 41 (2), 447–​481. Nelkin, Dana. (2008) “Responsibility and Rational Abilities: Defending an Asymmetrical View,” Pacific Philosophical Quarterly 89, pp. 497–​515. Nelkin, Dana. (2011). Making Sense of Freedom and Responsibility, Oxford: Oxford University Press. Nelkin, Dana. (2013). “Reply to Critics,” Philosophical Studies 163, pp. 123–​131. Pearl, Judea. (2000, 2009). Causality. Cambridge: Cambridge University Press; first edition 2000, second edition 2009. Pereboom, Derk. (1995). “Determinism Al Dente,” Noûs 29, pp. 21–​45. Pereboom, Derk (2001). Living Without Free Will. Cambridge University Press. Pereboom, Derk. (2007). “On Mele’s Free Will and Luck,” Philosophical Explorations 10, pp. 163–​172. Pereboom, Derk. (2008). “A Hard-​Line Reply to the Multiple-​Case Manipulation Argument,” Philosophy and Phenomenological Research 77, pp. 160–​170. Pereboom, Derk. (2014). Free Will, Agency, and Meaning in Life, Oxford: Oxford University Press. Sartorio, Carolina. (2016). Causation and Free Will. New York: Oxford University Press. Scanlon, T. M. (2013). “Giving Desert Its Due,” Philosophical Explorations 16, pp. 101–​116. Sekatskaya, Maria. (2019). “Double Defence Against Multiple Case Manipulation Arguments,” Philosophia 47(4), pp. 1283–​1295. Shabo, Seth. (2010). “Uncompromising Source Incompatibilism,” Philosophy and Phenomenological Research 80, pp. 349–​383. Spinoza, Benedict. (1677/​1985). Ethics, in The Collected Works of Spinoza, Edwin Curley, ed. and tr., Volume 1, Princeton NJ: Princeton University Press. Strawson, Peter F. (1962). “Freedom and Resentment,” Proceedings of the British Academy 48, pp. 187–​211. Taylor, Richard. (1974). Metaphysics, 4th ed., Englewood Cliffs, NJ: Prentice-​Hall.

200    Derk Pereboom and Michael McKenna Tierney, Hannah. (2013). “A ManeuverAaround the Modified Manipulation Argument,” Philosophical Studies 165, pp. 753–​633. Tierney, Hannah. (2014). “Taking It Head-​on: How to Best Handle the Modified Manipulation Argument,” Journal of Value Inquiry 48, pp. 663–​675. Tierney, Hannah, and David Glick. (2020). “Desperately Seeking Sourcehood,” Philosophical Studies 177(4), pp. 953–​970. Todd, Patrick. (2011). “A New Approach to Manipulation Arguments,” Philosophical Studies 152, pp. 127–​133. Todd, Patrick. (2013). “Defending (a Modified Version of the) Zygote Argument,” Philosophical Studies 164, pp. 189–​203. Usher, Marius. (2020). “Agency, Teleological Control and Reliable Causation,” Philosophy and Phenomenological Research 100(2), pp. 302–​324. Wallace, R. Jay. (1994). Responsibility and the Moral Sentiments, Cambridge, MA: Harvard University Press. Wolf, Susan. (1990). Freedom Within Reason, Oxford: Oxford University Press. Woodward, James. (2003). Making Things Happen: A Theory of Causal Explanation, New York: Oxford University Press. Woodward, James. (2006). “Sensitive and Insensitive Causation,” The Philosophical Review 115, pp. 1–​50.

PA RT V

SK E P T IC I SM

CHAPTER 10

Illu sioni sm Saul Smilansky

Illusionism (about free will and moral responsibility; henceforth the disclaimer will be dropped) is broadly the view that illusion is central and, in many ways, positive in the context of the free will problem. I begin by explaining what Illusionism is, the forms it may take, and the reasons for holding it. Then I explore Illusionism’s relations to other views on the free will problem, and the question of the role of illusion concerning free will and moral responsibility as compared to other contexts. I am primarily interested in the sort of free will that is a basis for moral responsibility and the concomitant notions such as desert, blameworthiness, and praiseworthiness. Finally, I discuss objections to Illusionism, concluding with reflections on why the position is not more widely held, and its future prospects. Illusionism emerges from the fundamentals of the human condition both morally and amorally insofar as they concern free will. In some ways, it seems to be a matter of common sense and careful risk aversion. And yet Illusionism is also a radical position, in many ways counterintuitive, philosophically and morally disturbing: perhaps even dangerous. It is also a complex, varied, and contextually flexible position. All these factors combine to create a need for taking care in understanding the view and in dealing with it. One of the special features of the view is that it is self-​reflexive; it invites reflection on whether it itself should be made widely known. Illusionism combines three elements: (a) A descriptive and normative claim that illusion concerning free will and moral responsibility is central in human life, including morality, justice, self-​understanding, interpersonal relations, appreciation (including self-​respect and respect for others), our understanding of history, and the meaning of life. (b) A claim that the role of illusion is all things considered positive. And (c) views on how the benefits of illusion as well as its risks and drawbacks are to be dealt with. These claims and views will be expanded upon in this chapter. What is an illusion of the sort we are considering? In the strong sense, we can think of illusions as false beliefs, which then generate epistemically dubious emotional reactions and social practices. One primary example is the belief in a robust libertarian free will

204   Saul Smilansky (LFW). But there is no reason for Illusionism to limit itself to something as strong as positive conscious belief. It suffices that people have rather vague tacit belief-​like states about, say, the actual ability they had in the past to have acted differently from how, in fact, they did act, and there is no need for explicit, let alone philosophically explicated, libertarian beliefs. Similarly, Illusionism need not be strongly committed to the motivated element in illusions. We do not always have to determine whether people are in fact actively deluding themselves, or would do so only when challenged. Indeed, illusion in such contexts (in contrast to, e.g., perceptual illusions) mostly seems to need to be accompanied by an “eraser” or “double-​illusion,” the illusion that we are not under illusion, if it is to be viable (Smilansky 2000: 219). When talking about illusion we are thus speaking of false belief-​like states, that may or may not be occurrent, and may or may not be actively motivated. The claims of Illusionism are ambitious. They are not merely that some illusions in the free will context are pleasant, and make life easier to deal with. The claims are that free will illusions are central, life-​enabling, and morality-​enabling features of the human condition: free will-​related false beliefs, and the related reactions and practices, are central in human self-​understanding, respect for persons, a moral order, and a decent society. Illusion is an enabling condition for many of the most important elements of our view of ourselves, for our ability to maintain moral seriousness, a sense of value, and indeed perhaps our very sense of self.1 Free will illusions, in other words, are not sugarcoating, but a large part of the very food that sustains us. For these reasons, we ought to maintain illusions. Illusion is so basic as to be, in part, constitutive of ourselves, value, and morality, while it is also in some ways fragile, and we ought to take care of our largely positive illusions, for fear of the consequences. Admittedly, free will-​related beliefs can be harmful. For example, as hard determinists and others have emphasized, robust libertarian beliefs can be held in ways which generate undue vindictiveness in punishment, or an overbearing and unnecessary degree of guilt feelings (see, e.g., Galen Strawson 1986/​2010; Pereboom 2001, 2014; Sommers 2007; Waller 2011; Caruso 2016).2 Yet Illusionism holds that these issues are relatively minor, when compared to the overwhelming positive importance of illusions concerning free will and moral responsibility.

1  We will be assuming that while the absence of free will would rule out desert, blameworthiness, and praiseworthiness, it would not rule out moral obligation, rightness, and wrongness. This assumption has been challenged (notably Haji 2002), yet it remains standard in the debate. One is morally obligated to stop and save a toddler who is about to run into a busy street, although (for hard determinists) one is not blameworthy if one does not do so. 2  By “hard determinism” I refer to the position that denies free will and moral responsibility, rejecting both libertarian and compatibilist ideas of free will. Determinism per se is not required, as long as any indeterminism (e.g., in quantum mechanics) does not generate free will. Some use the terms “free will skepticism,” “hard incompatibilism,” or “free will denialism” for the view, but I have chosen to remain here with the more familiar traditional terminology.

Illusionism   205

1.  Roads to Illusion Illusionism is not a common-​sense view, but the place to start understanding it is common sense. We can begin with what we can call the Core Conception of beliefs concerning free will; moral responsibility; and the concomitant beliefs, reactions, and practices (Smilansky 2000: ch. 2). This is quite general, and often quite vague, as well as an idealized version of prevailing views, which are often not so pure. According to this Core Conception, free will or control is the basis for moral responsibility, which is in turn a condition for deserving many sorts of evaluations, reactions, and treatments. Self-​attitudes, interpersonal relations, social interactions, institutions, and practices, should all, in manifold ways and within constraints and limitations but nevertheless very broadly, respond to and track responsibility-​relevant control. This is moral responsibility of the desert-​involving kind, not only a forward-​looking one. Hence, for example, pro tanto we ought to create social orders where personal choice and responsibility can be widely exercised and rewarded, accept moral responsibility and blame ourselves for our negative free actions, be grateful to those who freely help us, appreciate people (including ourselves) who make significant sacrifices or take risks for good moral reasons, and take great care and even risks in order not to blame or punish the innocent who lacked control over their actions. There is an intimate connection between the Core Conception concerning free will and moral responsibility, and respect for persons (Smilansky 2000: sec. 2.1, 2005; see also Bradley 1927: essay 1; Hart 1970; Morris 1976; Berlin 1980). To respect persons is to take seriously their choices and actions in themselves, not just for consequentialist reasons, to see them as autonomous and responsible agents who are appreciated and treated in accordance with their intentions, efforts, and doings. As John Gardner put it, “moral agency and moral responsibility represent a significant part of what it is to be a human being” (Gardner 2007: 218). We ought then to establish and cherish a Community of Responsibility (Smilansky 2000: 83), which follows the Core Conception values (as allowed for by libertarian or compatibilist beliefs). If one follows hard determinism or some other form of free will and moral responsibility denialism, then the Core Conception becomes largely impossible to apply, yet retains its negative salience, as we shall see ahead. The most effective belief for defending this Core Conception is robust libertarianism: the idea that although human beings are certainly affected and to some extent constrained by their heredity and environment, operating through their character, desires, and beliefs, they nevertheless are able to transcend these influences. Hence when looking back one can see oneself as having been in large measure “where the buck stops,” as a self-​determining cause who was not completely determined to choose and act as one has. One could, strictly, have decided to have done otherwise, exactly as one was, and this option was actually within one’s control. Robust LFW makes one morally responsible for one’s choices and actions in a similarly robust sense, deserving of a whole

206   Saul Smilansky set of moral and amoral evaluations, reactions, and treatments. This needs to be distinguished from “non-​robust” libertarianism, such as that put forward by Robert Kane (1996). The latter famously lacks the control element, and hence cannot give us much beyond what compatibilism can provide. Henceforth, when referring to libertarianism, I will be speaking of the robust forms of the view. Even if a libertarian view were correct, perhaps some room for illusion might still remain, yet in the free will context the need for illusion would then be quite minimal. I assume in this paper that libertarian free will does not exist, either because determinism is true, or because the sort of indeterminism that is true cannot generate LFW, or because, independently of determinism or indeterminism, the very notion of LFW is incoherent. Illusionism can be usefully seen as a way to deal with the implications of the absence of LFW. Another option is compatibilism, the view (to summarize) that although we do not have LFW, nevertheless more limited, local forms of control which are compatible with the absence of LFW are available, and suffice for moral responsibility and the related notions. Yet by retaining the centrality of control but setting the bar lower than the libertarians do, compatibilists open themselves to attack both philosophically and pragmatically. Think about criminals whom compatibilism considers fully free and morally responsible, who have done terrible things, and have ended up in long-​term punishment. Although they have done the awful things knowingly and willingly, their becoming such people was not ultimately up to them. When they are punished (and even if all other common conditions for punishment are met) they are, in an important sense, victims of the circumstances that have constituted them, making them what they are, with their particular set of motivations and sensibilities. In other words, they are being punished for constitutive luck (Nagel 1986, 2012; Smilansky 2000: 45–​47, 2003; Levy 2011), for matters ultimately beyond their control. And that is unfair and unjust. A similar story can be told in other matters, such as with self-​respect and the appreciation of others, and we confront here dangers such as the Danger of Worthlessness (Smilansky 2000: 153) and the Danger of Dissociation (Smilansky 2000: 209), or the risks to gratitude (more to follow). The philosophical limitations of compatibilism lend themselves directly to the recognition of the important role of illusion. Bonafide compatibilism (rather than, say, a utilitarian sort of consequentialism) seems to depend upon stopping our inquiries before they reach dangerous points for compatibilism. And that sort of philosophical holding back is laden with illusion. As Bernard Williams put it, “To the extent that the institution of blame works coherently, it does so because it attempts less than morality would like it to do . . . [it] takes the agent together with his character, and does not raise questions about his freedom to have chosen some other character” (1985, 194). According to R. Jay Wallace, “it is reasonable to hold agents morally accountable when they possess the power of reflective self-​control; and when such accountable agents violate the obligations to which we hold them, they deserve to be blamed for what they have done” (Wallace 1994: 226). He anticipates a countermove: It might be claimed that “the fairness of holding people responsible depends on their having the specific ability to

Illusionism   207 exercise their general powers of reflective self-​control, and thereby to avoid the harms of moral sanction” (Wallace 1994: 226) (my emphasis). Such specific opportunity to avoid is ruled out by determinism even under the compatibilist interpretation, hence creating grave unfairness. But Wallace merely dismisses the worry by saying that this intuition is not the one generating the sense of unfairness in the case of the standard compatibilist excuses (Wallace 1994: 207). Yet even if he is correct, it means only that we have more than one pertinent moral concern. One is compatibilist and is about the absence of, or limitations in, compatibilist control in particular cases, while the other concerns the ultimate-​level implications of the absence of LFW. Wallace does nothing to deny the sense of unfairness of the second sort of concern, which compatibilism cannot overcome (Smilansky 2003, 2008). Our decisions, even as ideal compatibilist agents, are an unfolding of the given; they reflect the way we were formed, and we had no opportunity to have been formed differently. If in the end it is only our bad constitutive luck, then in a deep sense it is not morally our fault. Ultimately, no one can be truly blame-​worthy. A compatibilist may respond in another way: “[M]‌y challenge to Smilansky and other advocates of the Ultimacy Argument is to produce some positive account of the element of free action that ultimacy affords and that a compatibilist cannot provide. Until then, I remain skeptical that any argumentative force should be granted to the U[ltimacy] condition as an aid in an argument for incompatibilism” (McKenna 2008: 198). In response, one need not defend libertarianism, but one does need to show why its absence is significant. And that can be done. Given that we incarcerate criminals for long periods of time and severely harm them, it matters that any compatibilist justification cannot overcome the idea that those criminals are ultimately victims of the factors that constituted them, ultimately beyond their control. It matters that they are basically trapped into a course which leads to their harming and being harmed. By contrast, had they LFW, they could transcend and avoid constitutive luck. Their punishment would then be much more just. So it matters a great deal that we cannot have LFW. And similarly for the basis for deserving the appreciation of others, self-​respect, and so on; all would be much deeper if we had more than the pedestrian sort of compatibilist control (see Smilansky 2000: ch.3; Smilansky 2003). This philosophical weakness of compatibilism translates into practical difficulties and dangers. The ethical importance of the Core Conception paradigm of free will and responsibility as a basis for desert, blame, and punishment should be taken very seriously, but the ultimate hard determinist perspective threatens to present it as a farce, a mere game without foundation; or a masquerade for consequentialist manipulation. Similarly with the crucial idea of a personal sense of value and appreciation that can be gained through our free actions: This is unlikely to be adequately maintained by individuals in their self-​estimates, nor warmly and consistently projected by others and by society. It is a widespread topic for conversation among philosophers teaching the free will problem, how hard it is to convince many of our students to take compatibilism seriously; not to see it instinctively, with Kant, as a “wretched subterfuge” (Kant 1788/​2015: 78). The abandonment of libertarian beliefs frequently leads to a free fall into the denial of all

208   Saul Smilansky senses of free will and moral responsibility. A widespread loss of moral and personal confidence can be expected, which compatibilism, even if true, could not fully contain. The risks of understanding and internalization are great. The widespread tacit, vague libertarian beliefs are the best safeguard of the partly justified compatibilist form of life; they are the “hothouse” that shelters the fragile compatibilist plants. And to the extent that these false but useful libertarian assumptions are threatened, and the compatibilist beliefs have to stand on their own, the sort of biases and limitations in recognizing the threats to compatibilism that we noted can come to the rescue. Turning now to hard determinism, we see different and in a sense even more fundamental difficulties, which also point toward the importance of illusion. Hard determinists lack the positive Core Conception resources of both libertarianism and compatibilism, the ways of evaluating, justifying, reacting to, and motivating people on the basis of their free will and moral responsibility-​based desert. This is not to deny that even without free will and moral responsibility we have some basis for appreciation, as we value babies although they lack free will, or the beauty of nature. Yet this is clearly much more limited, once the free will-​based forms of evaluation and appreciation are given up. Hard determinism is morally flattening. Everyone, whatever they do, is equally “innocent,” no one may justly deserve blame or praise, gratitude, or reward. One major difficulty concerns justice and punishment. For hard determinists, no one can be justly punished, and hence cannot be effectively threatened with punishment. At most hard determinists can opt for “funishment,” incarceration joined with very generous living conditions (in order to compensate for the undeserved injustice of incarceration), which is then self-​defeating for hard determinism (Smilansky 2011; for criticism of my “funishment” argument, see Levy 2012; Waller 2015: 197–​200; Pereboom 2017a). Joining forces here with a utilitarian-​like consequentialism (or other such ways of justifying punishment) means betrayal of the negative but true insights of a morally deep hard determinist position, concerning universal innocence and the concomitant injustice of punishment.3 It also involves hard determinism with the grave dangers of such forward-​looking consequentialism, which cannot be elaborated here (see Smilansky 2017b; 2019). Hard determinists have emphasized the drawbacks of traditional views which follow the free will paradigm, but these views are often also safeguards of equity, decency, and human rights.4 One example concerns punishment 3  It might be argued that if there is no free will in any sense, we must abandon the Core Conception, the very notion of desert ceases to make any sense, and with it any sense of fairness or injustice based on it. And then perhaps the pragmatic justification of punishment (or of other attitudes and practices) can proceed undisturbed. But this argument is unconvincing. Desert is a way of justifying divergence from a moral baseline. If desert becomes impossible, this does not mean that the moral baseline has somehow disappeared, and “anything goes” (cf. Smilansky 1996a, 1996b). If the moral baseline is that everyone ought to be treated as innocent unless proven guilty (through his or her free actions), then, if no one can become guilty, the moral baseline of innocence remains. Not to respect it would be unjust. It is thus a mistake to believe that since hard determinism rules out control-​based desert, there is no justice (or injustice) in a hard determinist world. 4  Villhauer (2013) offers a Rawlsian attempt to safeguard the deontological constraints in a free will and moral responsibility-​denialist world; which I argue also does not succeed (Smilansky 2019).

Illusionism   209 of the innocent. Criticism of utilitarianism here has traditionally focused on extreme examples such as scapegoating. But as I have shown long ago, the real danger concerns a systematic, moderate lowering of standards for prosecution and conviction (Smilansky 1990; 2000: 28–​30 with relevance to free will; see also the development of this argument in a critique of the hard determinist’s dominant quarantine model by Lemos 2016). Beyond the issue of punishment, hard determinists have had to recruit substitutes for the familiar beliefs, reactions, and practices based upon the belief in free will and moral responsibility. For example: “Instead of blaming people, the determinist might appeal to the practice of moral admonishment and encouragement. One might, for example, explain to an offender that what he did was wrong, and then encourage him to refrain from performing similar actions in the future” (Pereboom 2001: 325). One can also protest his intransience (Pereboom 2017b). But for a genuine hard determinist, this cannot go very far. If the admonished offender does not respond adequately, even after we protest, he cannot, after all, be blamed. Since inducing a sense of responsibility-​based blame is forbidden, people are unlikely to grow up with a dependable set of dispositions to behave responsibly—​they will not even be acquainted with the notion, except as that held by other, mistaken people, who have been misled into believing in personal responsibility and desert.5 We confront here The Present Danger of the Future Retrospective Excuse (Smilansky 2000: 153). We cannot sensibly tell people that they should behave in a certain way, that it is morally crucial that they do, but then, if they do not, say that this is (in every case) excusable, or at least exempts them from blameworthiness. As Tony Honore put it “to treat people as responsible promotes individual and social well-​being . . . It helps to preserve social order by encouraging good and discouraging bad behaviour. At the same time, it makes possible a sense of personal character and identity that is valuable for its own sake” (Honore 1999: 125). Morality has a crucial interest in restricting the influence of the ultimate-​level, hard determinist perspective. People as a rule ought not to be fully aware of the ultimate inevitability of what they have done. The knowledge that such an escape from moral responsibility will be available in the future is likely to affect the present view, and hence cannot be safely fully admitted even in its retrospective form. Thoughts about automatic exemption are very likely to cause great harm, hindering the inner pressure to resist temptation and reform oneself, and conveying that, whatever one does, one will be safe from the (just) blame and punishment of others, as well as justified guilt and compunction.6 Hard determinism lacks adequate resources to deal with such dangers. The substitutes hard determinists can offer, whether consequentialist or not, are either very weak, or involve the hard determinist in injustice and the victimization and harassment

5 

It might be claimed that traditional “shame” cultures provide an alternative. Considering this would require us to engage in complex empirical debates, and consider the moral problematics. But even without doing this, it can be dismissed, using William James’s term, as a “live option” for modern, mass societies; surely there is no way back. 6  This is an instance of what I have elsewhere called “Teflon Immorality” (Smilansky 2013a).

210   Saul Smilansky of the blameless and undeserving. The ideas of free will and moral responsibility are crucial. The difficulties go much beyond issues of justice and moral motivation. Consider the case of a mother who, for decades and at great sacrifice, overcoming poverty and hardship, conscientiously devoted herself to her children. Call her the Devoted Mother. On the ultimate hard determinist level, real appreciation and a real sense of her individual attainment make no sense. All her hard choices, her continuous and overwhelming efforts, her suffering and her triumphs, become embedded in what must be seen as only an inevitable chain of events beyond her control. What room, then, is there for a special sense that here one did “the best that one could”? What room for the pride of overcoming, an achievement which is never even thought about as being simply given? What room for deep appreciation and gratitude, an appreciation and a gratitude that is so different from merely being happy that something favorable happened? Bruce Waller does not deny the importance of gratitude, but simply thinks that we can get it all without assuming moral responsibility. As he says, responding to my doubts—​ “Nonetheless, it seems to me that understanding the deeper causes of character and behavior does nothing to compromise either admiration or gratitude. All parties to this discussion are aware of the profound biological causes at work in producing a mother’s deep love and devotion to her children; but that causal knowledge in no way reduces or demeans the wonderful nature of that love, nor the gratitude one feels for the sacrifices and love and acts of kindness bestowed by a loving mother” (Waller 2017: March 13). This seems to me highly implausible. I agree that there is a biological origin and some built-​in biological basis for the efforts and sacrifices of mothers, but there are, after all, huge actual and undeniable differences among mothers, in their commitment, devotion, efforts, and sacrifices. Viewing such efforts and sacrifices without assuming choice-​based and desert-​generating moral responsibility makes things very different. We can see this by reflecting on the Ungrateful Daughter of such an unusually effortful and sacrificing mother as our Devoted Mother. Normally we would take the Ungrateful Daughter to be at fault for not giving the mother what she deserves; to be unfeeling, unresponsive, and unjust. And we can sympathize with the deep disappointment of such a mother, whose lifelong devotion, effort, and contribution is not appreciated, who is not as it were really seen as a choosing, acting person, and who does not get the gratitude she deserves. But what can the hard determinist say? Not very much, it would seem. The mother operated as she was molded, an unfolding of the given. Under hard determinist views, some mothers are neglectful and abusive, others are loving and devoted, even heroic; neither type is morally responsible for their performance, and neither is deserving of appreciation in the true, credit-​giving sense. The central form of gratitude that is based upon appreciation of past and indeed present agency and effort becomes dubious. There is an equality of value, as it were, among all mothers—​the admirable one is not to be particularly appreciated and thanked. And if gratitude is not deserved, then ingratitude can hardly be such a grave fault as we normally feel that it is. As elsewhere, with gratitude hard determinists are living with meager resources, and offering us a near starvation

Illusionism   211 diet. The connection between appreciation and gratitude and certain central forms of love further shows the psychological and ethical poverty of hard determinism. Beyond the pragmatic dangers (e.g., to motivation), human life risks losing a deep sense of value and of meaning that are intimately connected to the idea of free and responsible agency, and of what one acquires by the way one exercises them. On the ultimate level, determinism is “the great eraser,” disconnecting human life from central aspects of a backward-​looking sense of desert for one’s goodwill, efforts, and contributions; crucial sources for generating (self)-​respect, a sense of value, and appreciation. Even if (or even to the extent that) free will is an illusion, it creates and enables a “form of life” that is not only actual but humanly central, and for which we have no substitute. I cannot discuss in detail further difficulties, but the denial of free will and moral responsibility will have many repercussions that, as with gratitude and ingratitude, can be seen as highly negative and disturbing. Consider situations of the greatest awfulness, such as a sexually predatory and abusive relative, or agents of mass historical cruelty and extermination. If all is thought to be, in the end, inevitable, how can it make sense for a malefactor to accept responsibility and feel compunction for doing things that, however awful, he now thinks that he could not have ultimately avoided doing? But how can we function morally, and be morally serious, without such expectations and feelings? Similarly disturbing results emerge if we now switch to the point of view of the victims. Think about non-​repentant wrongdoers who are now dead. On hard determinist lines, their living victims ought not to feel resentment toward their tormentors, and should automatically forgive them!7 So should everyone else, including the relatives of any murdered victims. Since these are varied, systematic, and grave problems for hard determinism, the benefits of illusion emerge. Each example shows another angle; further ones can of course be given. Together, they illustrate the simple fact that the Core Conception of free will and responsibility matters enormously for human life and respect for persons; yet hard determinism rejects this whole picture, offering but pale substitutes. We need what the hard determinist would consider as illusory beliefs, whether of the libertarian or of the compatibilist form, in order for society, human relations, personal life, and self-​respect to function adequately. Moreover, we need them if a sense of moral depth, fundamentals of justice, and the attainment of value are to continue to make much of their sense. There are direct roads leading from hard determinism to the recognition of the positive role of illusion, once the optimism of much of contemporary hard determinism is seen to be unsound.8

7 

Although they might still regard wrongdoing as a reason to weaken or dissolve a relationship (see Pereboom 2001: 201; cf. Scanlon 2008). 8  For criticism of Illusionism on these matters, see, for example, Sommers (2007): 144–​148, 169–​ 170; Pereboom (2014): 194–​199; Waller 2018: 20–​23. There are helpful discussions of various relevant questions, which I cannot consider here for reasons of space. Concerning friendship and the reactive attitudes, for example, see on the pessimistic side Shabo (2012); and on the optimistic (denialist) side Milam (2016).

212   Saul Smilansky People who believe that free will matters greatly follow the Core Conception and believe that agency and control are inherently morally salient, in a way that does not depend on whether taking agency seriously is (say) utility maximizing. The concern is about whether people deserve blame or praise, not only whether blaming or praising them maximizes good consequences. And this is central in respecting persons, as compared to managing and manipulating them. When concern for free will is deficient in a society, our humanity is pro tanto threatened. There is a danger that people will treat one another in large measure as though they are mere carriers of features or “symptoms” that are to be dealt with, rather than treating one another as agents capable of reasoned choice and responsibility, who should be responded to and treated according to their choices and actions. Appreciation, gratitude, self-​respect, acceptance of responsibility, are all desert-​ based notions: backward-​looking responses to our past efforts and achievements rather than mere forward-​looking manipulative means. It is notable, and worrisome, that in more recent years a strong inclination toward a broadly utilitarian-​like consequentialist, forward-​looking element can be discerned within the debate: Consider, for example, Manuel Vargas (2013) from the compatibilist side and Pereboom (2014) from the hard determinist side (cf. Smilansky 2017a).9 These are complex and subtle positions which cannot be taken up here in detail, but the present point here is that both share the strong forward-​looking inclination, with the accompanying dangers. This also helps us to see illusion as a safer, more humane alternative, helping to safeguard us. Illusionism also has a consequentialist component insofar as it warns of the dangers of living without free will-​based illusions. But the consequentialist element comes on the meta-​normative level, that is, Illusionism seeks in large part to defend non-​consequentialist concerns, for example, with desert, moral depth, self-​and-​other appreciation and respect, and the sanctity of moral innocence. I cannot of course consider all positions and variations that are alternatives to Illusionism, and their connection to the latter, but one further example must be noted, that of P. F. Strawson in “Freedom and Resentment” (Strawson 1962/​2003). Interpreting Strawson is a matter of debate (see, e.g., McKenna & Russell, eds., 2008), but for our purposes it suffices to take up two points that Strawson makes. The first is the Humean claim that philosophy and indeed reason are weak in the free will context, so that our beliefs and practices will in fact be closely intertwined with the “reactive attitudes,” attitudes such as resentment, indignation, and gratitude, which simply assume that most people nearly all of the time are proper targets for the attribution of responsibility for their choices and actions. This, Strawson claims, makes determinism or indeed any other free will-​related worry rather moot, for human beings seem by and large hardwired to hold themselves and others as morally responsible. 9 

As distinguished from those philosophers who have no inherent interest in free will because they start from utilitarianism, such as Sidgwick (1907/​1963) and Smart (1961). But note that utilitarian-​like views will be sympathetic to Illusionism insofar as the good is maximized through illusion, which as we saw there are excellent reasons to believe will often be the case. Moreover such views will have no inherent constraints about endorsing illusion.

Illusionism   213 However, Strawson’s “no need to worry” claim is far too strong, and must be viewed as complacent (as I show in detail, Smilansky 2001; see also Sommers 2012). History and comparative studies of current societies indeed show no indication of societies where people are not held responsible in any sense. This makes the mountain that “happy hard determinists” (optimistic free will and moral responsibility denialists) have to climb that much taller. Yet at the same time, these studies also clearly show that in no way can we rest assured that elementary moral distinctions and constraints will be respected. Respect for persons and the idea of a decent society require that we take the Core Conception seriously, but this is not automatic. Humans have been and are all too ready to blame and punish collectively, disregard agency and innocence, not care about ability (in any sense) to have avoided the given sanction, and the like. Strawson’s “reactive-​naturalism” cannot be saved without an overwhelming element of Illusionism. Another aspect of P. F. Strawson’s thought is the emphasis on our emotions, which also appears with other philosophers, such as Galen Strawson (1986/​2010); Ted Honderich (1988); Richard Double (1991), and Paul Russell (1995). The temptation is to bypass the cognitive difficulties, and find shelter in self-​sustaining emotions. The difficulties of this non-​cognitivist direction can be seen, for example, from the worries about gratitude or compunction, as noted before. Once we leave behind P. F. Strawson’s assurance of built-​in stability, such an approach cannot make what matters safe from free will-​related doubts. As Bernard Williams reminded us, “Blame that is perceived as unjust often fails to have the desired result, and merely generates resentment” (Williams 1995: 15; see also Wolf 1981). Illusionism not only respects the person as subject, our common phenomenology of the openness of choice, of our sometimes “doing one’s best,” and the like, but creates a safe place for it through defending the related beliefs. Prevalent views on free will are seen, then, to be far too optimistic. Each proponent puts the optimism at a different stage and has, for example, libertarian, compatibilist, or hard determinist forms of indefensible optimism, which we then see as complacent (Smilansky 2010; for an exception see Russell 2013). Once these forms of overoptimism are recognized, illusion is not far off. Whether as a pragmatic “buffer” or as an inherent or existential necessity, illusion is seen to be central and indispensable. Moreover, since Illusionism is seen to follow from different positions, its importance is enhanced even if we are doubtful as to which position is the most plausible. My own view is that the most persuasive position is neither a full-​blown compatibilism nor a full-​blown hard determinism or denialism, but a mixed view, which combines the limited, but valid, insights of both. This Fundamental Dualism (Smilansky 2000: ch. 6) or compatibility pluralism has been grudgingly beginning to find some support in the debate in recent years (and certainly more so than Illusionism). Compatibility pluralism can in part be seen as context dependent, whereby in certain contexts compatibilism is more salient, and in others the ultimate hard determinist perspective is more helpful. But beyond that, often both perspectives matter together. I argue that in order to respect persons, we must create and maintain a Community of Responsibility that will follow the Core Conception, and in general track compatibilist distinctions in terms of

214   Saul Smilansky the presence (or occasional absence) of responsibility-​relevant control. This is not only a matter of forward-​looking consequences and psychological needs, but an inherent ethical obligation. Overall, compatibilism is “the best game in town.” But even when ideal compatibilist conditions are satisfied, moral justification, evaluation, appreciation, and meaning are inherently limited and shallow. And that too matters a great deal. That is the human condition—​our being creatures who typically have a large measure of local compatibilist control, who ought to be treated as responsible agents, who are allowed to live out the consequences of our choices—​but we are at the same time determined beings, operating as we were molded, and this often generates severe injustice and great limitations in value and meaning. This dual perspective on the compatibility question also generates its own requirements for illusion, in order to maintain the uneasy balance of the opposing views (see the Dissonance Problem, Smilansky 2000: sec. 7.2).10 Finally, while one ought to be wary of easily translating empirical psychological studies about the role of free will beliefs into philosophical conclusions, and there are both interpretative issues and some studies going in the broadly opposite direction, there has been in recent years a very large empirical body of literature supporting my skepticism about the benefits of the prospect of living without belief in free will and moral responsibility. This clearly seems to support my conservatism, and blows wind onto the sails of Illusionism. A recent survey will serve to show the breadth of the conclusions: It has been repeatedly shown that attenuating the belief in free will leads to various negative outcomes, such as undermined sense of agency (being in control of one’s actions) (Lynn et al. 2014), self-​alienation (Seto and Hicks 2016) . . . decrease in perceived meaningfulness in life (Crescioni et al. 2015), corruption of helpfulness and increase in aggression (Baumeister et al. 2009), weakening of the feeling of gratitude (MacKenzie et al. 2014), and reduction of cooperative behavior (Protzko et al. 2016). The belief in free will is also associated with . . . stronger sense of self-​efficacy, life satisfaction, meaning in life, gratitude, greater commitment in relationships (Crescioni et al. 2015), learning from emotional experiences (Stillman and Baumeister 2010), and beliefs in morality (Bergner and Ramon 2013). (Baliavsky 2020)

10 

Further reasons to see Illusionism as persuasive, including examples and indirect supporting evidence, can be found in Smilansky 2000: 178ff. Critical attempts to check the case for my Illusionism on free will on the basis of folk psychology are Nadelhoffer and Feltz (2007), and Nadelhoffer and Mateeva (2009). The general results of experimental philosophical surveys of common attitudes on the free will problem are consistently complex, pluralistic, contradictory, and inconclusive (see, e.g., Knobe and Nichols 2017). Arguably this lends support both to my pluralism on the compatibility question and to Illusionism, but I will not make the case here. Suffice for our purposes is that there does not plausibly seem to be firm grounding to resist Illusionism on this basis.

Illusionism   215

2.  Is Illusionism on Free Will Special? Is the case for Illusionism on free will special, or indeed unique? We cannot enter here on a general discussion of the importance of avoiding illusion in life. Within the bounds of this chapter, two things can be done. The first is to consider free will as a context for Illusionism in comparison to other contexts. Here I shall claim that the case for Illusionism on free will and moral responsibility is particularly strong. The second task is to nevertheless see Illusionism as firmly based in common ways in which human beings deal with themselves and the world. Reflecting upon Illusionism in comparative and historical terms may reasonably seem to lead to skepticism about it. The belief in God in particular was widely thought among the Victorians, for example, to be a condition for social order and personal satisfaction, so that the abandonment of such belief was viewed with horror. This pessimism has been disproven. Today it would be hard to deny that some of the most successful societies ever, in terms of humane and just social interaction as well as personal happiness, such as those of certain Western European nations like the Scandinavian countries, have very low levels of significant religious belief. This is not to deny that some religious beliefs can play and have played important positive, morally civilizing roles, irrespective of their truth. Even if one considers those very same Scandinavian societies, going back a thousand years one finds Vikings who thrived on murder of the innocent, rape, and pillage. The influence of the Judeo-​Christian tradition, operating through the Christianization of the Viking tribes, has been instrumental in forming the considerate, humane, peace-​loving, egalitarian, democratic and broadly just societies in modern day Scandinavia. And yet today religious beliefs cannot be thought to be necessary, at least in certain well-​ordered societies. Once the historical ladder has been climbed, it can be dispensed with. Nevertheless, this stock example of the opponent of the idea of the “morally necessary illusion” does not harm the case for Illusionism concerning free will and moral responsibility. For, while certain religious beliefs can have a civilizing effect, they are hardly fundamental to the human condition, in the way in which we have seen free will-​related beliefs to be. The idea of God is not crucial for our notions of justice and moral worth. This has been known as far back as the time of Socrates and Plato, as can be seen in the Euthyphro. Free will aside, people can gain moral worth irrespective of the existence of a God who will reward or sanction them. Indeed, arguably, lack of belief in divine response makes one more deserving for one’s good will and deeds, which are then further purified of self-​interest. With free will, by contrast, it is the very notion of moral worth that is threatened (Smilansky 2000: 245–​246).11

11  Paradoxically, belief in hard determinism has the perverse moral advantage of potentially making one especially morally worthy, because one does not think that one’s moral worth is at stake in the free will-​related ways. This is enhanced if one chooses morally, believes in hard determinism, but hard

216   Saul Smilansky While the case of free will and moral responsibility has a special claim to be so central and fundamental as to generate Illusionism, we nevertheless need to see the illusion-​ friendly background that Illusionism can build upon. There is abundant and overwhelming evidence, particularly from social psychology, that people widely delude themselves in important ways, and that such bias is broadly beneficial. Most people greatly overestimate matters of self-​concern, such as the importance of their job, their work performance, their children, or the way in which others view them. In fact, the only category of people who get that last sort of issue right are depressives! The issue of when these prevalent illusions are positive or negative is complex and often context dependent (see, e.g., Elster 1983; Goleman 1985; Taylor 1989; Taylor and Brown 1994; Makridakis and Moleskis 2015; Schütz and Baumeister 2017). There is no place to delve more deeply into these matters here. Their importance for us, however, is obvious. We see that humanity wallows in illusion, the effects of which frequently seem to be largely beneficial. Illusionism on free will and moral responsibility is not, then, some alien and exotic possibility, an artificial construct of grouchy philosophers, but grows upon a central, typical, and patently successful human form of engagement with the world.

3.  The Prospects of Illusionism Why has Illusionism not been more widely espoused? Illusionism on free will and moral responsibility has been quite rare. Erasmus believed in LFW, but noted the dangers of free will-​denying views such as those of Wycliffe and Luther. People, as things stand, tend toward evil, and “There is no sense in pouring oil upon the fire” (Erasmus 1981: 12). “Definitely, it seems to me, it is not only unsuitable, but truly pernicious to carry on such disputations when everybody can listen” (Erasmus 1981: 12). Nietzsche holds that the concept of moral freedom is a lie (1979: 96), but sometimes notes its usefulness. Vaihinger (1924) speaks about free will as a useful open fiction, but this is part of his general position not only about free will, which is wildly implausible (see Smilansky 2000: 146–​148). William James’s famous “Will to believe” position (James 1896/​1956) is hospitable to and was motivated by the desire for belief in free will beyond the evidence (see Smilansky 1992). Occasionally legal scholars have expressed illusionistic views. For example, Sanford Kadish says that “the ancient notion of free will may well in substantial measure be a myth. But even a convinced determinist should reject a governmental regime which is founded on anything less in its system of authoritative disposition of citizens” (Kadish 1968: 287). Galen Strawson (1986/​2010) explored the phenomenology of freedom and the idea that libertarian beliefs (which he considers incoherent) are

determinism is at least partly false and there is some free will-​related moral worth. See Smilansky (1994) and a response in Double (2004); Smilansky (2000 sec. 10.1).

Illusionism   217 inextricably embedded in our common psychology. Smilansky (2000, 2012) is the major philosophical proponent of Illusionism on free will. To the extent that people are easily misled or successfully delude themselves about free will and moral responsibility, so the lack of recognition of the prevalence of illusion is only to be expected. The position has the very resources to explain its limited popularity. Yet more can be said in the direction of providing an “error theory” (Mackie 1977) here. As P. F. Strawson explained, “The fact that we find ourselves in our desires and preferences and do not, in general, find them as alien presences within ourselves; the experience of deliberation which heightens and strengthens our sense of self; and the constantly repeated experience of agency—​all these contribute to, perhaps constitute, the sense of freedom” (Strawson 1992: 135). Another major factor is the prevailing optimism we already noted (Smilansky 2010). There seems to be a deep general inclination toward optimism concerning free will, which cannot, I claim, be philosophically defended, but which then prevents many philosophers working in the field from realizing the need for illusory support. Finally, what is the future of Illusionism? This question is ambiguous between (i) the actual role of illusion and (ii) the recognition of this role. We will take these two questions in turn. To some extent, we saw free will-​related illusion to be inherent in the human condition. But in other ways, the situation is much more fluid. For example, in recent years there has been a surge in claims that “free will is an illusion,” primarily by scientists (e.g., Libet et al. 1983; Wegner 2002). There are excellent reasons to be skeptical about these scientific claims (see, e.g., Mele 2014; Shepherd 2017), but they have been influential. In a situation (such as prevailed until not long ago) where public awareness of the free will problem is very limited, Illusionism is best served by the mere continuation of the lack of awareness of the problem. Yet if skepticism or denialism about free will and moral responsibility would continue to rise to a threatening level, the game will change. There is, for example, no pragmatic need for compatibilism if nearly everyone is a tacit libertarian, but if denialism begins to prevail, compatibilism becomes attractive. As before, I believe that compatibilism is partly true and important, but also much weaker philosophically than its adherents believe. Yet, pragmatically, a growth in belief in compatibilism, without full awareness of its limitations, is likely to be welcome under such new and dangerous conditions. The idea of “positive illusions” is highly unpleasant and confronts opposing values and concerns such as knowledge, intellectual honesty, veracity, integrity, authenticity, openness, and public transparency (see Smilansky 2000: ch. 11). I am not claiming that we need to induce illusory beliefs concerning free will. I would not wish to be engaged in a project of deception. I do not perceive myself as deceiving by omission. My position here is rather passive and conservative. Illusory beliefs are in place, and the role they play is largely positive. Humanity is fortunately mistaken and deceived on the free will issue, and this seems to be a condition of civilized morality and a personal sense of value. But in the light of the importance of the free will illusions, the strength of the case for Illusionism is itself depressing, and its logic can lead, in different contingencies, to

218   Saul Smilansky highly disturbing conclusions. Indeed Illusionism on free will seems to be a prime example of the idea of a “Crazy ethics” (Smilansky 2013b). As to (ii), the future acceptance of Illusionism as a philosophical position, it is hard to tell. Philosophically I hope that this chapter has helped to show the merits of Illusionism. Others, who disagree, have argued and can continue to argue against it. Within philosophical space, that is what we should hope for (on Illusionism and the role of philosophy, see Smilansky 2000: sec. 11.4). In the public sphere, the rise of the recognition of the importance of illusion in the free will context cannot be simply welcomed, for reasons that I have explicated. Moreover, as already noted, illusion typically functions best when hidden; the idea that we can induce in ourselves or live with ideas that we fully realize to be illusory is highly implausible. To some extent we can consider the idea of Unillusioned Moral Individuals (UMI’s) concerning free will (see Smilansky 2000: 246). But mostly illusion will continue to function more or less well, even for those in the know, because of the power of the illusion, and the psychological difficulty of truly internalizing the truth.12 If I am correct, and the free will problem is toxic stuff that by and large one would do best to avoid, then a Humean-​like escape from the philosophy may be safest, if one begins to become ensnared. Illusionism naturally invites skepticism, indeed incredulity, and perhaps even hostility. Yet the philosophical case for it is strong; assuming the absence of robust libertarian free will. The largely positive, crucial role of illusion follows from the importance of the Core Conception values at stake, and the grave weaknesses of compatibilism, hard determinism, and the other alternatives for dealing with them. There is a pragmatic worry that the truth will cause overreaction in many, who will cease to follow the partly justified compatibilist distinctions which are required for respect for persons, and fall into crass utilitarian-​like consequentialist attitudes or forms of nihilism. And there are even deeper worries, resulting from the inherent role of illusion in creating our moral and personal selves, in the face of the awfulness of the truth.13

References Beliavsky, Vlad (2020). Freedom, Responsibility and Therapy. London: Palgrave Macmillan. Berlin, Isaiah (1980). “From Hope and Fear Set Free,” in Concepts and Categories. Oxford: Oxford University Press. Bradley, F.H. (1927). Ethical Studies. Oxford: Clarendon Press.

12 

For very interesting reflections on these matters see G. Strawson (2003). I am very grateful to the editors for the invitation to participate in this volume. I presented an early version of the paper to an international workshop on my work in Lund, and to a related meeting in Gothenburg, in October 2017, both part of the Swedish GLRP; and am grateful to participants for their comments. I am also very grateful to Gregg Caruso, Amihud Gilead, Robert Hartman, Arnon Keren, Iddo Landau, Sam Lebens, Ariel Meirav, Dana Nelkin, Guy Pinku, Derk Pereboom, Tonni Ronnow-​ Rasmussen, Tamler Sommers, Daniel Statman, Galen Strawson, Andras Szigeti and Rivka Weinberg, for helpful comments on drafts of my paper. 13 

Illusionism   219 Caruso, G.D. (2016). “Free will skepticism and criminal behavior: A public health-​quarantine model,” Southwest Philosophy Review 32: 25–​48. Double, Richard (1991). The Non-​reality of Free Will. New York: Oxford University Press. Double, Richard (2004). “The ethical advantages of free will subjectivism,” Philosophy and Phenomenological Research 69: 411–​422. Elster, Jon (1983). Sour Grapes. Cambridge: Cambridge University Press. Gardner, John (2007). Offences and Defences. Oxford: Oxford University Press. Goleman, Daniel (1985). Vital Lies, Simple Truths. New York: Simon & Schuster. Haji, Ishtiyaque (2002). Deontic Morality and Control. Cambridge: Cambridge University Press. Hart, H.L.A. (1970). Punishment and Responsibility. Oxford: Clarendon Press. Honderich, Ted (1988). A Theory of Determinism. Oxford: Clarendon Press. Honore, Tony (1999). Responsibility and Fault. Oxford: Hart Publishing. James, William (1896/​1956). “The Will to Believe,” in The Will to Believe and Other Essays in Popular Philosophy. New York: Dover Publications. Kadish, Sanford H. (1968). “The decline of innocence,” Cambridge Law Journal 26: 273–​290. Kane, Robert (1996). The Significance of Free Will. New York: Oxford University Press. Kant, Immanuel (1788/​2015). Critique of Practical Reason. Mary Gregor, trans. Revised Edition. Cambridge: Cambridge University Press. Knobe, Joshua and Shaun Nichols (2017). “Experimental philosophy,” Standord Encyclopedia of Philosophy (online). Lemos, John (2016). “Moral concerns about responsibility denial and the quarantine of violent criminals,” Law and Philosophy 35: 461–​483. Levy, Neil (2011). Hard Luck. Oxford: Oxford University Press. Levy, Neil (2012). “Skepticism and sanction,” Law and Philosophy 31: 477–​493. Libet, Benjamin, Gleason, Curtis A., Wright, Elwood W. & Pearl, Dennis K. (1983). Time of conscious intention to act in relation to onset of cerebral activity (readiness-​potential). The unconscious initiation of a freely voluntary act. Brain 106 (3):623–​664. Mackie, J.L. (1977). Ethics. Harmondsworth: Penguin. Makridakis, Spyros and Andreas Moleskis (2015). “The costs and benefits of positive illusions,” Frontiers in Psychology 6, 1–​11. doi: 10.3389/​fpsyg.2015.00859. McKenna, Michael (2008). “Ultimacy and sweet Jane,” in Nick Trakakis and Daniel Cohen, eds., Essays on Free Will and Moral Responsibility. Cambridge: Scholars Publishing. McKenna, Michael and Paul Russell, eds. (2008). Free Will and Reactive Attitudes: Perspectives on P.F. Strawson’s “Freedom and Resentment.” Farnham, UK: Ashgate. Mele, Alfred (2014). Free: Why Science Hasn’t Disproved Free Will. New York: Oxford University Press. Milam, Per-​Erik (2016). “Reactive-​attitudes and personal relationships,” Canadian Journal of Philosophy 46: 102–​122. Morris, Herbert (1976). “Persons and punishment,” in On Guilt and Innocence. Berkeley and Los Angeles: University of California Press. Nadelhoffer, Thomas and Adam Feltz (2007). “Folk intuitions, slippery slopes, and necessary fictions: An essay on Saul Smilansky’s free will Illusionism,” Midwest Studies in Philosophy 31: 202–​213. Nadelhoffer, Thomas and Tatyana Mateeva (2009). “Positive illusions, perceived control and the free will debate,” Mind & Language 24: 495–​522. Nagel, Thomas (1986). The View from Nowhere. Oxford: Oxford University Press.

220   Saul Smilansky Nagel, Thomas (2012). “Moral luck,” in Mortal Questions. Cambridge: Cambridge University Press. Nietzsche, Friedrich (1979). “On truth and lies in a normal sense,” in Philosophy and Truth: Selections from Nietzsche’s Notebooks of the Early 1870’s. Atlantic Highlands, NJ: Humanities Press. Pereboom, Derk (2001). Living Without Free Will. Cambridge: Cambridge University Press. Pereboom, Derk (2014). Free Will, Agency and Meaning in Life. New York: Oxford University Press. Pereboom, Derk (2017a). “A defense of free will skepticism: Replies to commentaries by Victor Tadros, Saul Smilansky, Michael McKenna, and Alfred R. Mele on Free Will, Agency, and Meaning in Life,” Criminal Law and Philosophy 11: 617–​636. Pereboom, Derk (2017b). “Responsibility, regret and protest,” Oxford Studies in Agency and Responsibility 4: 121–​140. Russell, Paul (1995). Freedom and Moral Sentiments. New York: Oxford University Press. Russell, Paul (2013). “Compatibilist-​fatalism: Finitude, pessimism, and the limits of free will,” in The Philosophy of Free Will, Paul Russell and Oisin Deery, eds. Oxford: Oxford University Press. Scanlon, T.M. (2008). Moral Dimensions: Permissibility, Meaning, Blame. Cambridge, MA: Harvard University Press. Schütz, Astrid and Roy F. Baumeister (2017). “Positive illusions and the happy mind,” in Michael D. Robinson and Michael Eid, eds., The Happy Mind: Cognitive Contributions to Well-​Being. Cham, Switzerland: Springer. Shabo, Seth (2012). “Where love and resentment meet: Strawson’s intrapersonal defense of compatibilism,” Philosophical Review 121: 95–​124. Shepherd, Joshua (2017). “Neuroscientific threats to free will,” in Meghan Griffith, Neil Levy, and Kevin Timpe, eds., The Routledge Companion to Free Will. New York: Routledge. Sidgwick, Henry (1907/​1963). The Methods of Ethics (7th ed.). London: Macmillan. Smart, J.J.C. (1961). “Free will, praise and blame,” Mind 70: 291–​306. Smilansky, Saul (1990). “Utilitarianism and the ‘punishment’ of the innocent: The general problem,” Analysis 50: 256–​261. Smilansky, Saul (1992). “Did James deceive himself about free will?,” Transactions of the C.S. Peirce Society 28: 767–​779. Smilansky, Saul (1994). “The ethical advantages of hard determinism,” Philosophy and Phenomenological Research 54: 355–​363. Smilansky, Saul (1996a). “Responsibility and desert: Defending the connection,” Mind 105: 157–​163. Smilansky, Saul (1996b). “The connection between responsibility and desert: The crucial distinction,” Mind 105: 385–​386. Smilansky, Saul (2000). Free Will and Illusion. Oxford: Oxford University Press. Smilansky, Saul (2001). “Free will: From nature to illusion,” Proceedings of the Aristotelian Society 101: 71–​95. Smilansky, Saul (2003). “Compatibilism: The argument from shallowness,” Philosophical Studies 115: 257–​282. Smilansky, Saul (2005). “Free will and respect for persons,” Midwest Studies in Philosophy 29: 248–​261. Smilansky, Saul (2008). “Free will and fairness,” in Nick Trakakis and Daniel Cohen, eds., Essays on Free Will and Moral Responsibility. Cambridge: Scholars Publishing.

Illusionism   221 Smilansky, Saul (2010). “Free will: some bad news,” in Action, Ethics and Responsibility, Joseph Keim Campbell, Michael O’Rourke, and Harry S. Silverstein, eds., Cambridge, MA: MIT Press. Smilansky, Saul (2011). “Hard determinism and punishment: a practical reductio,” Law and Philosophy 30: 353–​367. Smilansky, Saul (2012). “Free will and moral responsibility: The Trap, the appreciation of agency, and the bubble,” Journal of Ethics 16: 211–​239. Smilansky, Saul (2013a). “Why moral paradoxes matter: ‘Teflon Immorality’ and the perversity of life,” Philosophical Studies 165: 229–​243. Smilansky, Saul (2013b). “Free will as a case of ‘Crazy ethics,’” Exploring the Illusion of Free Will and Moral Responsibility, Gregg D. Caruso, ed., Lanham, MD: Lexington Books. Smilansky, Saul (2017a). “The free will problem: Nonstandard views,” in The Routledge Companion to Free Will. New York: Routledge. Smilansky, Saul (2017b). “Pereboom on punishment—​funishment, innocence, motivation, and other difficulties,” Criminal Law and Philosophy 11: 591–​603. Smilansky, Saul (2019). “Free will denial and deontological constraints,” in Elizabeth Shaw, Derk Pereboom, and Gregg D. Caruso, eds. Free Will Skepticism in Law and Society. New York: Cambridge University Press. Sommers, Tamler (2007). “The objective attitude,” Philosophical Quarterly 57: 321–​341. Sommers, Tamler (2012). Relative Justice. Princeton, NJ: Princeton University Press. Strawson, Galen (1986/​2010). Freedom and Belief. Oxford: Oxford University Press. Strawson, Galen (2003). “You cannot make yourself the way you are.” A Conversation with Tamler Sommers. Believer (online, issue 1). Strawson, P.F. (1962/​2003). “Freedom and resentment,” in Gary Watson (ed.), Free Will. Oxford: Oxford University Press. Strawson, P.F. (1992). Analysis and Metaphysics. Oxford: Oxford University Press. Taylor, Shelley E. (1989). Positive Illusions. New York: Basic Books. Taylor, Shelley E. and Jonathan D. Brown (1994). “Positive illusions and well-​being revisited: Separating fact from fiction,” Psychological Bulletin 116: 21–​27. Vaihinger, Hans (1924). The Philosophy of “As If,” London: Kegan, Paul, Trench, Trubner & Co. Vargas, Manuel (2013). Building Better Beings. Oxford: Oxford University Press. Villhauer, Benjamin (2013). “Persons, punishment and free will skepticism,” Philosophical Studies 162: 143–​163. Wallace, R. Jay (1994). Responsibility and the Moral Sentiments. Cambridge, Ma: Harvard University Press. Waller, Bruce (2011). Against Moral Responsibility. Cambridge, MA: MIT Press. Waller, Bruce (2015). The Stubborn System of Moral Responsibility. Cambridge, MA: MIT Press. Waller, Bruce (2017). Syndicate; symposium on The Stubborn System of Moral Responsibility (online). Waller, Bruce (2018). The Injustice of Punishment. New York: Routledge. Wegner, Daniel M. (2002). The Illusion of Conscious Will. MIT Press. Williams, Bernard (1985). Ethics and the Limits of Philosophy. London: Fontana. Williams, Bernard (1995). “How free does the will need to be?,” in Making Sense of Humanity. Cambridge: Cambridge University Press. Wolf, Susan (1981). “The importance of free will,” Mind 90: 386–​405.

CHAPTER 11

F ree Will Skep t i c i sm a nd Crim inal J u st i c e The Public Health-​Quarantine Model Gregg D. Caruso

One of the most frequently voiced criticisms of free will skepticism is that it is unable to adequately deal with criminal behavior and that the responses it would permit as justified are insufficient for acceptable social policy. This concern is fueled by two factors. The first is that one of the most prominent justifications for punishing criminals, retributivism, is incompatible with free will skepticism. The second concern is that alternative justifications that are not ruled out by the skeptical view per se face significant independent moral objections (Pereboom 2014: 153). Despite these concerns, I maintain that free will skepticism leaves intact other ways to respond to criminal behavior—​in particular incapacitation, rehabilitation, and alteration of relevant social conditions—​ and that these methods are both morally justifiable and sufficient for good social policy. The position I defend is similar to Derk Pereboom’s (2001, 2013, 2014), taking as its starting point his quarantine analogy, but it sets out to develop the quarantine model within a broader justificatory framework drawn from public health ethics. The resulting model—​which I call the public health-​quarantine model (Caruso 2016, 2017a, 2021)—​ provides a framework for justifying quarantine and criminal sanctions that is more humane than retributivism and preferable to other non-​retributive alternatives. It also provides a broader approach to criminal behavior than Pereboom’s quarantine analogy does on its own since it prioritizes prevention and social justice.1 In section 1, I begin by (very) briefly summarizing my arguments against free will and basic desert moral responsibility. In section 2, I then introduce and defend my public health-​quarantine model, which is a non-​retributive alternative to criminal punishment that prioritizes prevention and social justice. In sections 3 and 4, I take up and respond

1 

This chapter includes some passages from Caruso (2016, 2017a).

Free Will Skepticism and Criminal Justice    223 to two general objections to the public health-​quarantine model. Since objections by Michael Corrado (2016), John Lemos (2016), Saul Smilanksy (2011, 2017), and Victor Tadros (2017) have been addressed in detail elsewhere (see Pereboom 2017a; Pereboom and Caruso 2018, 2021), I will here focus on objections that have not yet been addressed. In particular, I will respond to concerns about proportionality, human dignity, and victims’ rights. I will argue that each of these concerns can be met and that in the end the public health-​quarantine model offers a superior alternative to retributive punishment and other non-​retributive accounts.

1.  Free Will Skepticism To begin, it is important to first get clear on what type of moral responsibility is being doubted or denied by free will skeptics. Most skeptics maintain that our best philosophical and scientific theories about the world indicate that what we do and the way we are is ultimately the result of factors beyond our control, whether that be determinism, chance, or luck, and that because of this, agents are never morally responsible in the sense needed to justify certain kinds of desert-​based judgments, attitudes, or treatments—​such as resentment, indignation, moral anger, backward-​looking blame, and retributive punishment. This is not to say that there are no other conceptions of responsibility that can be reconciled with determinism, chance, or luck. Nor is it to deny that there may be good reasons to maintain certain systems of punishment and reward. Rather, it is to insist that to hold people truly deserving of blame and praise, punishment and reward, would be to hold them responsible for the results of the morally arbitrary or for what is ultimately beyond their control, which is fundamentally unfair and unjust. Derk Pereboom provides a very helpful definition of the kind of moral responsibility being doubted by skeptics, which he calls basic desert moral responsibility and defines as follows: For an agent to be morally responsible for an action in this sense is for it to be hers in such a way that she would deserve to be blamed if she understood that it was morally wrong, and she would deserve to be praised if she understood that it was morally exemplary. The desert at issue here is basic in the sense that the agent would deserve to be blamed or praised just because she has performed the action, given an understanding of its moral status, and not, for example, merely by virtue of consequentialist or contractualist considerations. (2014: 2)

Consistent with this definition, I have elsewhere argued that we should understand basic desert moral responsibility in terms of whether it would ever be appropriate for a hypothetical divine all-​knowing judge (who didn’t necessarily create the agents in question) to administer differing kinds of treatment (i.e., greater or lesser rewards or punishments) to human agents on the basis of actions that these agents performed

224   Gregg D. Caruso during their lifetime (see Caruso and Morris 2017). The purpose of invoking the notion of a divine judge in the afterlife is to instill the idea that any rewards or punishments issued after death will have no further utility—​be it positive or negative. Any differences in treatment to agents (however slight) would therefore seem warranted only from a basic desert sense, and not a consequentialist perspective. In the past, the standard argument for free will skepticism was based on the notion of determinism—​the thesis that every event or action, including human action, is the inevitable result of preceding events and actions and the laws of nature. Hard determinists argued that determinism is true and incompatible with free will and moral responsibility—​either because it precludes the ability to do otherwise (leeway incompatibilism) or because it is inconsistent with one’s being the “ultimate source” of action (source incompatibilism). Hard determinism had its classic statement in the time when Newtonian physics reigned, but it has very few defenders today. Most contemporary skeptics instead defend positions that are agnostic about determinism—​for example, Derk Pereboom (2001, 2014), Galen Strawson (1986, 1994), Saul Smilansky (2000), Neil Levy (2011), Richard Double (1991), Bruce Waller (2011, 2015), and myself (2012) (2021). Most maintain that while determinism is incompatible with free will and basic desert moral responsibility, so too is indeterminism, especially the variety posited by quantum mechanics. Others argue that regardless of the causal structure of the universe, we lack free will and moral responsibility because free will is incompatible with the pervasiveness of luck (Levy 2011). Others (still) argue that free will and ultimate moral responsibility are incoherent concepts, since to be free in the sense required for ultimate moral responsibility we would have to be causa sui (or “cause of oneself) and this is impossible (Strawson 1994). My own reasons for adopting free will skepticism amount to a rejection of both compatibilism and libertarianism. I maintain that the sort of free will required for basic desert moral responsibility is incompatible with causal determination by factors beyond the agent’s control and also with the kind of indeterminacy in action required by the most plausible versions of libertarianism. For this reason, I follow Pereboom in labeling my view hard incompatibilism (Pereboom 2001, 2014; Pereboom and Caruso 2018). Against the view that free will is compatible with the causal determination of our actions by natural factors beyond our control, I argue that there is no relevant difference between this prospect and our actions being causally determined by manipulators (see Pereboom 2001, 2014; Mele 2008; Todd 2011, 2013). I further argue that it is incompatible with an agent’s ability to do otherwise, a necessary condition for free will (Caruso 2012). Against event causal libertarianism, I advance the “luck” or “disappearing agent” objection, according to which agents are left unable to settle whether a decision/​action occurs and hence cannot have the control in action required for moral responsibility (see Caruso 2012, 2015, 2021; Pereboom 2001, 2014; 2017b; Waller 1990, 2011; Levy 2008, 2011; for non-​skeptics who advance similar objections see Mele 1999, 2017; Haji 2001). The same problem, I contend, arises for non-​causal libertarian accounts since these too fail to provide agents with the control in action needed for basic desert (see Pereboom 2014). While agent-​causal libertarianism could, in theory, supply this sort of control, I

Free Will Skepticism and Criminal Justice    225 argue that it cannot be reconciled with our best physical theories and faces additional problems accounting for mental causation (Caruso 2012, 2021). Since this exhausts the options for views on which we have the sort of free will at issue, I conclude that free will skepticism is the only remaining position. Since the arguments for hard incompatibilism have been spelled out and defended at great length elsewhere (see, e.g., Pereboom 2001, 2014; Caruso 2012, 2013, 2021; Levy 2011; Pereboom and Caruso 2018), and no solid refutation of them have yet been offered, I will not elaborate on them further here. Instead, my goal in this chapter is to explore the practical implications of free will skepticism. For many, it is not the philosophical arguments for free will skepticism that are the problem, it is the existential angst they create and the fear that relinquishing belief in free will and basic desert moral responsibility would undermine morality, negatively affect our interpersonal relationships, destroy meaning in life, and leave us unable to adequately deal with criminal behavior. Optimistic skeptics, however, respond by arguing that life without free will and basic desert moral responsibility would not be as destructive as these critics maintain, and, in fact, may be preferable in a number of important ways (see Pereboom 1995, 2001, 2014; Waller 2011, 2015; Caruso 2016, 2017a, 2017b, forthcoming; Pereboom and Caruso 2018). These optimistic skeptics argue that prospects of finding meaning in life or of sustaining good interpersonal relationships, for instance, would not be threatened. They further maintain that morality and moral judgments would remain intact. And although retributivism and severe punishment, such as the death penalty, would be ruled out, they argue that the imposition of sanctions could serve purposes other than the punishment of the guilty—​e.g., it can also be justified by its role in incapacitating, rehabilitating, and deterring offenders. In this chapter, I attempt to extend this general optimism about the practical implications of free will skepticism to the practices and policies surrounding criminal behavior.

2.  The Public Health-​Q uarantine Model It is important to begin by recognizing that retributive punishment is incompatible with free will skepticism because it maintains that punishment of a wrongdoer is justified for the reason that he deserves something bad to happen to him just because he has knowingly done wrong—​this could include pain, deprivation, or death. As Douglas Husak puts it, “Punishment is justified only when and to the extent it is deserved” (2000: 82). And Mitchell Berman writes, “A person who unjustifiably and inexcusably causes or risks harm to others or to significant social interests deserves to suffer for that choice, and he deserves to suffer in proportion to the extent to which his regard or concern for others falls short of what is properly demanded of him” (2008: 269). Furthermore, for the retributivist, it is the basic desert attached to the criminal’s immoral action alone that

226   Gregg D. Caruso provides the justification for punishment. The desert the retributivist invokes is basic in the sense that justifications for punishment that appeal to it are not reducible to consequentialist considerations nor to goods such as the safety of society or the moral improvement of the criminal. Free will skepticism undermines this justification for punishment because it does away with the idea of basic desert. If agents do not deserve blame just because they have knowingly done wrong, neither do they deserve punishment just because they have knowingly done wrong. The challenge facing free will skepticism, then, is to explain how we can adequately deal with criminal behavior without the justification provided by retributivism and basic desert moral responsibility. While some critics contend this cannot be done, free will skeptics point out that there are several alternative ways of justifying criminal punishment (and dealing with criminal behavior more generally) that do not appeal to the notion of basic desert and are thus not threatened by free will skepticism. These include moral education theories, deterrence theories, punishment justified by the right to harm in self-​defense, and incapacitation theories. While I have elsewhere argued that the first two approaches face independent moral objections—​ objections that, though perhaps not devastating, make them less desirable than their alternative (see Pereboom 2001, 2014; Caruso 2016, 2021; Pereboom and Caruso 2018)—​ I maintain that an incapacitation account built on the right to harm in self-​defense provides the best option for justifying a policy for treatment of criminals consistent with free will skepticism (Caruso 2016, 2017a 2021; Pereboom and Caruso 2018). The public health-​quarantine model is based on an analogy with quarantine and draws on a comparison between treatment of dangerous criminals and treatment of carriers of dangerous diseases. It takes as its starting point Derk Pereboom’s famous account (2001, 2013, 2014). In its simplest form, it can be stated as follows: (1) Free will skepticism maintains that criminals are not morally responsible for their actions in the basic desert sense; (2) plainly, many carriers of dangerous diseases are not responsible in this or in any other sense for having contracted these diseases; (3) yet, we generally agree that it is sometimes permissible to quarantine them, and the justification for doing so is the right to self-​protection and the prevention of harm to others; (4) for similar reasons, even if a dangerous criminal is not morally responsible for his crimes in the basic desert sense (perhaps because no one is ever in this way morally responsible) it could be as legitimate to preventatively detain him as to quarantine the non-​responsible carrier of a serious communicable disease (Pereboom 2014: 156). The first thing to note about the theory is that although one might justify quarantine (in the case of disease) and incapacitation (in the case of dangerous criminals) on purely utilitarian or consequentialist grounds, both Pereboom and I want to resist this strategy (see Pereboom and Caruso 2018). Instead, on our view incapacitation of the dangerous is justified on the ground of the right to harm in self-​defense and defense of others. That we have this right has broad appeal, much broader than utilitarianism or consequentialism has. In addition, this makes the view more resilient to a number of objections (see Pereboom 2017a; Pereboom and Caruso 2018).

Free Will Skepticism and Criminal Justice    227 Second, the quarantine model places several constraints on the treatment of criminals (see Pereboom 2001, 2014; Pereboom and Caruso 2018). First, as less dangerous diseases justify only preventative measures less restrictive than quarantine, so less dangerous criminal tendencies justify only more moderate restraints (Pereboom 2014: 156). In fact, for certain minor crimes perhaps only some degree of monitoring could be defended. Secondly, the incapacitation account that results from this analogy demands a degree of concern for the rehabilitation and well-​being of the criminal that would alter much of current practice. Just as fairness recommends that we seek to cure the diseased we quarantine, so fairness would counsel that we attempt to rehabilitate the criminals we detain (Pereboom 2014: 156). If a criminal cannot be rehabilitated, however, and our safety requires his indefinite confinement, this account provides no justification for making his life more miserable than would be required to guard against the danger he poses (Pereboom 2014: 156). Third, this account also provides a more resilient proposal for justifying criminal sanctions than other non-​retributive options. One advantage it has, say, over consequentialist deterrence theories is that it has more restrictions placed on it with regard to using people merely as a means. For instance, as it is illegitimate to treat carriers of a disease more harmfully than is necessary to neutralize the danger they pose, treating those with violent criminal tendencies more harshly than is required to protect society will be illegitimate as well. In fact, in all our writings on the subject, Pereboom and I have always maintained the principle of least infringement, which holds that the least restrictive measures should be taken to protect public health and safety (Caruso 2016, 2017a, 2021; Pereboom and Caruso 2018). This ensures that criminal sanctions will be proportionate to the danger posed by an individual, and any sanctions that exceed this upper bound will be unjustified. In addition to these restrictions on harsh and unnecessary treatment, the model also advocates for a broader approach to criminal behavior that moves beyond the narrow focus on sanctions. On the model I have developed, the quarantine analogy is placed within the broader justificatory framework of public health ethics (Caruso 2016, 2017a, 2021). Public health ethics not only justifies quarantining carriers of infectious diseases on the grounds that it is necessary to protect public health, it also requires that we take active steps to prevent such outbreaks from occurring in the first place. Quarantine is only needed when the public health system fails in its primary function. Since no system is perfect, quarantine will likely be needed for the foreseeable future, but it should not be the primary means of dealing with public health. The analogous claim holds for incapacitation. Taking a public health approach to criminal behavior would allow us to justify the incapacitation of dangerous criminals when needed, but it would also make prevention a primary function of the criminal justice system. So instead of myopically focusing on punishment, the public health-​quarantine model shifts the focus to identifying and addressing the systemic causes of crime, such as poverty, low social economic status, systematic disadvantage, mental illness, homelessness, educational inequity, abuse, and addiction (see Caruso 2017a).

228   Gregg D. Caruso In my recent Public Health and Safety: The Social Determinants of Health and Criminal Behavior (2017a), I argue that the social determinants of health (SDH) and the social determinants of criminal behavior (SDCB) are broadly similar, and that we should adopt a broad public health approach for identifying and taking action on these shared social determinants. I focus on how social inequities and systemic injustices affect health outcomes and criminal behavior, how poverty affects brain development, how offenders often have preexisting medical conditions (especially mental health issues), how homelessness and education affects health and safety outcomes, how environmental health is important to both public health and safety, how involvement in the criminal justice system itself can lead to or worsen health and cognitive problems, and how a public health approach can be successfully applied within the criminal justice system. I argue that, just as it is important to identify and take action on the SDH if we want to improve health outcomes, it is equally important to identify and address the SDCB. And I conclude by offering eight broad public policy proposals for implementing a public health approach aimed at addressing the SDH and SDCB (see Caruso 2017a for details). Furthermore, the public health framework I adopt sees social justice as a foundational cornerstone to public health and safety (Caruso 2016, 2017a, 2021). In public health ethics, a failure on the part of public health institutions to ensure the social conditions necessary to achieve a sufficient level of health is considered a grave injustice. An important task of public health ethics, then, is to identify which inequalities in health are the most egregious and thus which should be given the highest priority in public health policy and practice. The public health approach to criminal behavior likewise maintains that a core moral function of the criminal justice system is to identify and remedy social and economic inequalities responsible for crime. Just as public health is negatively affected by poverty, racism, and systematic inequality, so too is public safety. This broader approach to criminal justice therefore places issues of social justice at the forefront. It sees racism, sexism, poverty, and systemic disadvantage as serious threats to public safety and it prioritizes the reduction of such inequalities (see Caruso 2017a). While there are different ways of understanding social justice and different philosophical accounts of what a theory of justice aims to achieve, I favor a capability approach according to which the development of capabilities—​what each individual is able to do or be—​is essential to human well-​being (e.g., Sen 1985, 1999; Nussbaum 2011; Power and Faden 2006). For capability theorists, human well-​being is the proper end of a theory of justice. And on the particular capability approach I favor, social justice is grounded in six key features of human well-​being: health, reasoning, self-​determination, attachment, personal security, and respect (see Powers and Faden 2006; Caruso 2017a 2021).2 Following Powers and Faden (2006), I maintain that each of these six dimensions is an essential feature of well-​being such that “a life substantially lacking in any one is a life seriously deficient in what it is reasonable for anyone to want, whatever else they want”

2 

Note that this is a pared-​down list from the ones offered by Martha Nussbaum and other capability theorists (see Nussbaum 2011).

Free Will Skepticism and Criminal Justice    229 (Powers and Faden 2006: 8). The job of justice is therefore to achieve a sufficiency of these six essential dimensions of human well-​being, since each is a separate indicator of a decent life. The key idea of capability approaches is that social arrangements should aim to expand people’s capabilities—​their freedom to promote or achieve functionings that are important to them. Functionings are defined as the valuable activities and states that make up human well-​being, such as having a healthy body, being safe, or having a job. While they are related to goods and income, they are instead described in terms of what a person is able to do or be as a result. For example, when a person’s need for food (a commodity) is met, they enjoy the functioning of being well-​nourished. Examples of functionings include being mobile, being healthy, being adequately nourished, and being educated. The genuine opportunity to achieve a particular functioning is called a capability. Capabilities are “the alternative combination of functionings that are feasible for [a person] to achieve”—​they are “the substantive freedom” a person has “to lead the kind of life he or she has reason to value” (Sen 1999: 87). As Tabandeh, Gardoni, and Murphy describe: Genuine opportunities and actual achievements are influenced by what individuals have and what they can do with what they have. What they can do with what they have is a function of the structure of social, legal, economic, and political institutions and of the characteristics of the built-​environment (i.e., infrastructure). For example, consider the functioning of being mobile. The number of times an individual travels per week can be an indicator of mobility achievement. When explaining a given individual’s achievement or lack of achievement, a capability approach takes into consideration the conditions that must be in place for an individual to be mobile. For instance, the possession of certain resources, like a bike, may influence mobility. However, possessing a bike may not be sufficient to guarantee mobility. If the individual has physical disabilities, then the bike will be of no help to travel. Similarly, if there are no paved roads or if societal culture imposes a norm that women are not allowed to ride a bike, then it will become difficult or even impossible to travel by means of a bike. As this example makes clear, different factors will influence the number of times the individual travels. (Tabandeh, Gardoni, and Murphy 2017)

Thinking in terms of capabilities raises a wider range of issues than simply looking at the amount of resources or commodities people have, because people have different needs. In the example given above, just providing bicycles to people will not be enough to increase the functioning of being mobile if you are disabled or prohibited from riding because of sexist social norms. A capabilities approach to social justice therefore requires that we consider and address a larger set of social issues. Bringing everything together, my public health-​quarantine model characterizes the moral foundation of public health as social justice, not just the advancement of good health outcomes. That is, while promoting social goods (like health) is one area of concern, public health ethics as I conceive it is embedded within a broader commitment to secure a sufficient level of health and safety for all and to narrow unjust inequalities (see

230   Gregg D. Caruso Powers and Faden 2006). More specifically, I see the capability approach to social justice as the proper moral foundation of public health ethics. This means that the broader commitment of public health should be the achievement of those capabilities needed to secure a sufficient level of human well-​being—​including, but not limited to, health, reasoning, self-​determination, attachment, personal security, and respect. By placing social justice at the foundation of the public health approach, the realms of criminal justice and social justice are brought closer together. I see this as a virtue of the theory since it is hard to see how we can adequately deal with criminal justice without simultaneously addressing issues of social justice. Retributivists tend to disagree since they approach criminal justice as an issue of individual responsibility and desert, not as an issue of prevention and public safety. I believe it is a mistake to hold that the criteria of individual accountability can be settled apart from considerations of social justice and the social determinants of criminal behavior. Making social justice foundational, as my public health-​quarantine model does, places on us a collective responsibility—​ which is forward-​looking and perfectly consistent with free will skepticism—​to redress unjust inequalities and to advance collective aims and priorities such as public health and safety. The capability approach and the public health approach therefore fit nicely together. Both maintain that poor health and safety are often the byproducts of social inequities, and both attempt to identify and address these social inequities in order to achieve a sufficient level of health and safety. Summarizing the public health-​quarantine model, then, the core idea is that the right to harm in self-​defense and defense of others justifies incapacitating the criminally dangerous with the minimum harm required for adequate protection. The resulting account would not justify the sort of criminal punishment whose legitimacy is most dubious, such as death or confinement in the most common kinds of prisons in our society. The model also specifies attention to the well-​being of criminals, which would change much of current policy. Furthermore, the public health component of the theory prioritizes prevention and social justice and aims at identifying and taking action on the social determinants of health and criminal behavior. This combined approach to dealing with criminal behavior, I maintain, is sufficient for dealing with dangerous criminals, leads to a more humane and effective social policy, and is actually preferable to the harsh and often excessive forms of punishment that typically come with retributivism.

3.  Proportionality and Human Dignity One concern critics have with my approach to criminal behavior is that they fear it will not protect human dignity and respect for persons in the same way that retributivism does. Retributivists adopt something called the principle of proportionality. As Alec Walen describes, “Retributive justice holds that it would be bad to punish a wrongdoer more than she deserves, where what she deserves must be in some way proportional

Free Will Skepticism and Criminal Justice    231 to the gravity of her crime. Inflicting disproportionate punishment wrongs her just as, even if not quite as much as, punishing an innocent person wrongs her . . .” (Walen 2014). For retributivists, the principle of proportionality is needed to guarantee respect for persons since it treats them as autonomous, morally responsible agents and not just objects to be “fixed” or used as a means to an end. Hence, punishment administered because one is a morally responsible autonomous person who justly deserves punishment due to his or her own choices, preserves one’s status as a person and a member of the human community of responsible agents as long as it is not disproportionate (see Lewis 1971; Oldenquist 1988; and Morris 1968). Critics contend that without this principle in place, there will be no limit to the harshness of punishment meted out and no way to block treating individuals as a mere means to an end. Immanuel Kant, for example, famously argued that human beings possess a special dignity and worth which demands that they be treated as ends in themselves and never as mere means. According to Kant, imprisonment could only be justified on the grounds that the criminal conduct was a product of the free willed choices of the criminal making him/​her deserving of a punitive response. Kant, however, also believed that the death penalty was deserved, in fact obligatory, in cases of murder: But whoever has committed murder, must die. There is, in this case, no juridical substitute or surrogate that can be given or taken for the satisfaction of justice. There is no likeness or proportion between life, however painful, and death; and therefore there is no equality between the crime of murder and the retaliation of it but what is judicially accomplished by the execution of the criminal. His death, however, must be kept free from all maltreatment that would make the humanity suffering in his person loathsome or abominable. Even if a civil society resolved to dissolve itself with the consent of all its members—​as might be supposed in the case of a people inhabiting an island resolving to separate and scatter themselves throughout the whole world—​the last murderer lying in prison ought to be executed before the resolution was carried out. This ought to be done in order that every one may realize the desert of his deeds, and that blood-​guiltiness may not remain upon the people; for otherwise they might all be regarded as participators in the murder as a public violation of justice. (Kant 1790, Part II: 6)

While many retributivists disagree with Kant regarding the death penalty, they share his belief that punishment should not exceed what is deserved and that free will and basic desert moral responsibility are needed to maintain respect for persons. John Lemos, for example, has argued, “the human capacity for moral responsibility gives human beings a special dignity and worth that is fundamental to a proper system of morality grounded on the concept of respect for persons” (2013: 78), and theories of punishment that reject basic desert moral responsibility are incapable of protecting this special dignity and worth (see Lemos 2013, 2016). In response, I would argue three things: (1) It is unclear that the principle of proportionality in actual practice protects respect for persons any better than the alternatives; (2) what counts as proportional punishment is unclear and as a result several important

232   Gregg D. Caruso questions remain—​for example, how should we measure the gravity of a wrong, and how can punishment be “proportional” to it?; and (3) the public health-​quarantine model has a non-​desert-​based principle of proportionality of its own—​one which I maintain is capable of securing respect for persons and protecting innocent people from being used as a means to an end. Let me take each of these in turn. First, while concerns over proportionality are important ones, the worry that relinquishing the concept of just deserts will lead to harsh and inhumane treatment of persons is overblown. Before getting to the more philosophical responses to this objection, I would first like to examine the question empirically and ask whether belief in just deserts and retributive justice ensure punishment is proportional any better than the alternatives. Since the real-​life effects of free will skepticism is what is being questioned here, I think the empirical question is an important one. If the critics are wrong about the protective power of desert-​based moral responsibility and the constraints it places on proportional punishment, then this concern loses much of its force. Empirically speaking, then, does belief in just deserts and retributive justice ensure punishment is proportional? I contend that it does not. Of course, there are many reasonable retributivists who acknowledge that we imprison far too many people, in far too harsh conditions, but the problem is that retributivism remains committed to the core belief that criminals deserve to be punished and suffer for the harms they have caused. Recall Kant’s claim that we should execute the last prisoner on the island before we abandon it in order that everyone “realize the desert of his deeds.” This retributive impulse in actual practice—​despite theoretical appeals to proportionality by its proponents—​often leads to practices and policies that try to make life in prison as unpleasant as possible. Bruce Waller has done an excellent job examining this question empirically and he sets up the cultural expectations as follows: Belief in individual moral responsibility is deep and broad in both the United States and England; in fact, the belief seems to be more deeply entrenched in those cultures than anywhere else—​certainly deeper there than in Europe. That powerful belief in moral responsibility is not an isolated belief, existing independently of other cultural factors; rather, it is held in place—​and in turn, helps anchor—​a neo-​liberal cultural system of beliefs and values. At the opposite end of the scale are social democratic corporatist cultures like Sweden that have taken significant steps beyond the narrow focus on individual moral responsibility. With that picture in view, consider the basic protections which philosophers have claimed that the moral responsibility system afford: first, protection against extreme punitive measures; second, protection of the dignity and rights of those who are held morally responsible and subject to punishment; and third, a special protection of the innocent against unjust punishment. According to the claim that strong belief in individual moral responsibility protects against abuses, we would expect the United States and Great Britain (the neo-​liberal cultures with the strongest commitment to individual moral responsibility) to score best in providing such protections; and we would predict that Norway, Sweden, and Denmark (the social democratic corporatist cultures, with much more qualified

Free Will Skepticism and Criminal Justice    233 belief in individual moral responsibility) would be the worst abusers. (2014: 6; see also 2015)

What happens when we actually make the comparison, however, is that we find the exact opposite. That is, we find that the stronger the belief in basic desert moral responsibility (as in the United States) the harsher the punishment, the greater the skepticism of moral responsibility (as in Norway) the weaker the inclination toward punishment. A few cross-​cultural statistics should help make this point salient. In 2014, the Pew Research Center asked people whether they agreed or disagreed with the notion that personal success is determined by factors outside of oneself. While not exactly measuring belief in free will and moral responsibility, the survey was able to confirm that Americans are much more likely to see success or failure in personal terms. This is in line with the systems of thinking Waller describes and is unsurprising given the United States’ emphasis on rugged individualism and individual responsibility—​ which, of course, is closely aligned with attitudes about just deserts, praise and blame, punishment and reward. For example, 57% of Americans disagreed with the statement “Success in life is determined by forces outside our control,” which was the highest percentage among advanced countries. The United Kingdom was immediately behind the United States with 55% disagreeing. Unfortunately Scandinavian countries were not included in the survey but European nations Germany and Italy came in at 31% and 32% respectively. Now, retributivists would have us believe that given its strong commitment to individual moral responsibility, the United States can be expected to provide better protections against harsh and excessively punitive forms of punishment than countries with a weaker commitment to individual moral responsibility. The reality, however, is quite the opposite. Consider the problem of mass incarceration in the United States. While the United States makes up only 5% of the world’s population, it houses 25% of the world’s prisoners—​that’s one of the highest rates of incarceration known to mankind. Despite a steady decline in the crime rate over the past two decades, the Unites States imprisons more than 700 prisoners for every 100,000 of population, according to the International Centre for Prison Studies (ICPS).3 Compare that to the social democratic countries with a much weaker commitment to individual moral responsibility, such as Norway, Sweden, and Finland, where the imprisonment rate hovers around 70 per 100,000. As a proportion of the population, then, the United States has 10 times as many prisoners as these other countries. Furthermore, the United States not only imprisons at a much higher rate, it also imprisons in notoriously harsh conditions. American supermax prisons are often cruel places, using a number of harsh forms of punishment including extended solitary confinement. Prisoners are isolated in windowless, soundproof cubicles for 23 to 24 hours each day, sometimes for decades. Under

3 

International Centre for Prison Studies, “World Prison Brief,” accessed November 5, 2013, www.prisonstudies.org/​highest-​lowest.

234   Gregg D. Caruso such conditions, prisoners experience severe suffering, often resulting in serious psychological problems. Supreme Court Justice Anthony Kennedy, for instance, recently stated that, “solitary confinement literally drives men mad.”4 Looked at empirically, then, it’s nigh impossible to defend the claim that commitment to just deserts and retributivism ensures proportional and humane punishment. In fact, the opposite seems to be the case—​the problem of disproportionate punishment seems to grow more out of a desire for retribution and the belief that people justly deserve what they get than from free will skepticism. This claim is further supported by the fact that individual states within the United States with stronger belief in individual moral responsibility tend to have harsher forms of punishment (see Waller 2014, 2015). Given these cross-​cultural and interstate comparisons, I cannot help but conclude along with Waller that, “commitment to moral responsibility exacerbates rather than prevents excessively harsh punitive policies” (2014: 7). Recent work in experimental philosophy further revels that where belief in free will is strongest we tend to find increased punitiveness (see Shariff et al. 2014; Carey and Paulhus 2013). Perhaps the strongest evidence for this linking comes from a set of recent studies by Shariff et al. (2014). Shariff and his colleagues hypothesized that if free will beliefs support attributions of moral responsibility, then reducing these beliefs should make people less punitive in their attitudes about punishment. In a series of four studies, they tested this prediction. In Study 1 they found that people with weaker free will beliefs endorsed less retributive attitudes regarding punishment of criminals, yet their consequentialist attitudes were unaffected. In the study, 244 American participants completed the seven-​item Free Will subscale of the Free Will and Determinism Plus scale (FAD+) (Paulhus and Carey 2011), which measures belief in free will. In order to further measure attitudes toward retributivist and consequentialist motivations for punishment, Shariff and his colleagues had participants read descriptions of retributivism and consequentialism as motivations for punishment and then indicate on two separate Likert scales (1 = strongly disagree, 7 = strongly agree) how important retributivism and consequentialism should be in determining motivation for criminal punishments. As predicted, Shariff et al. found that stronger belief in free will predicted greater support for retributive punishment, but was not predictive of support for consequentialist punishment. The effects remained significant when statistically controlled for age, gender, education, religiosity, and economic and social political ideology. Study 1 therefore supports the hypothesis that free will beliefs positively predict retributive attitudes, yet it also suggests that “the motivation to punish in order to benefit society (consequentialist punishment) may remain intact, even while the need for blame and desire for retribution are forgone” (Shariff et al. 2014: 7). It is Study 2, however, that really highlights how stronger belief in free will and moral responsibility can lead to increased punitiveness. In the study, participants were 4  He made this statement before the House Appropriations Subcommittee on Financial Services and Federal Government, as reported on in the Huffington Post on 3/​24/​2015: http://​ www.huffingtonpost.com/​2015/​03/​24/​anthony-​kennedy-​solitary-​confinement_​n_​6934550.html

Free Will Skepticism and Criminal Justice    235 randomly assigned to one of two groups. In the anti-​free will condition, participants were given a passage from Francis Crick’s (1994) The Astonishing Hypothesis which rejected free will and advocated for a mechanistic view of human behavior. In the neutral condition, the passage was unrelated to free will. Next, participants read a fictional vignette involving an offender who beat a man to death. Acting as hypothetic jurors, participants recommended the length of the prison sentence (if any) that this offender should serve following a 2-​year, nearly 100%-​effective, rehabilitation treatment. As Shariff et al. describe: The notion that the offender had been rehabilitated was used in order to isolate participants’ desire for punishment as retribution. The passage further focused participants on retributive, rather than consequentialist, punishment by noting that the prosecution and defense had agreed that the rehabilitation would prevent recidivism and that any further detention after rehabilitation would offer no additional deterrence of other potential criminals. (Shariff et al. 2014: 4)

As predicted, participants who read the anti-​free will passage recommended significantly lighter prison sentences than participants who read the neutral passage. In particular, participants whose free will beliefs had been experimentally diminished recommended roughly half the length of imprisonment (~5 years) compared with participants who read the neutral passage (~10 years). This study helps further confirm that it is actually commitment to retributivism that increases punitiveness, contrary to what its proponent’s claim.5 Moving on to my second reply, the principle of proportionality does not provide us with any clear and unambiguous way of measuring the gravity of a wrong. Nor does it tell us how we should determine which punishment is “proportional” to the wrong done. There is no magic ledger we can look to that spells out the gravity of a wrong in one column and the punishment that is deserved in another. This is obvious from the fact that retributivists often disagree with one another about how to measure the gravity of a wrong—​consider, for instance, H.L.A. Hart’s question: “Is negligently causing the destruction of a city worse than the intentional wounding of a single policeman?” (1968: 162). And even when there is wide agreement on the gravity of a wrong, there is still often disagreement about what kind of punishment is deserved. For instance, all 5  Carey and Paulhus (2013) also found a relationship between beliefs about free will and punishment. In particular, they found that believing more strongly in free will was correlated with increased punitiveness—​i.e., free will believers were more likely to call for harsher criminal punishment in a number of hypothetical scenarios. In the third of their studies, for instance, Carey and Paulhus presented two scenarios portraying serious crimes (child molestation and the rape of an adult woman) and tested the degree to which subjects’ attitudes toward punishment of the criminals would be impacted by factors including the criminal having been abused as a child and assurance that a medical procedure would prevent the criminal from ever perpetrating similar crimes again. The fact that subjects who expressed the strongest belief in free will were essentially the only group of subjects whose attitudes toward punishment were not mitigated by environmental or consequentialist considerations led the researchers to conclude that “free will belief is related to retributivist punishment” (2013: 138).

236   Gregg D. Caruso retributivists can agree that murdering an innocent person is a grievous wrong, but they can, and often do, disagree on what count as “proportional” punishment. Kant proposes death. Others propose life in prison. Others still think life in prison is too harsh. How do we decide questions like these on the principle of proportionality? The problem of measuring gravity, for instance, is an important one for retributivists since what punishment is deserved is going to be determined by this. Yet the proportionality principle leaves unanswered several important questions. The first is “does it matter if harm is caused, or is the gravity of the wrong set fully by the wrong risked or intended” (Walen 2014). (For the position that harm does not matter, see Feinberg 1995; Alexander, Ferzan, and Morse 2009; for a criticism of that view, see Levy 2005; Walen 2010.) Second, what significance, if any, should be given to the difference between being punished for the first time, and having been punished before and then having committed the same or a similar wrong again (see Walen 2014)? Until retributivists can agree on how to resolve these problems it remains unclear how gravity should be measured, which needs to be settled if we are to know how to apply the proportionality principle in practice. Assuming for the moment, however, that a rank order of gravity is possible, there still remains the problem of determining what counts as proportional punishment. There are two basic senses of proportionality that can be found in the literature: cardinal and ordinal. Cardinal proportionality sets absolute measures for punishment that is proportional to a given crime; ordinal proportionality requires only that more serious crimes should be punished more severely (Walen 2014). There are, however, problems with both approaches. Cardinal proportionality, for instance, tends to lead to unacceptable extremes. For example: Lex Talionis . . . offers a theory of cardinal proportionality. In its traditional form—​an eye for an eye, a tooth for a tooth—​it seems implausible, both for being too lenient in some cases (take $10 from a thief who stole $10), and too extreme in others (repeatedly torture and rape someone who had committed many such acts himself). Kant proposed what might be thought a better version, saying that the thief should lose not just the value of what he stole, but instead all rights to property (1797: 142), and prohibiting those forms of “mistreatment that could make the humanity in the person suffering it into something abominable” (ibid.). Nonetheless, his measure for theft swings to the overly punitive side, leaving the convicted thief a dependent on the state, and thereby “reduced to the status of a slave for a certain time, or permanently if the state sees fit” (ibid.). Others have tried to rehabilitate lex talionis, arguing, for example, that it can be rendered plausible if interpreted to call for punishment that “possess[es] some or all of the characteristics that made the offense wrong” (Waldron 1992: 35). But however one spells out the wrong-​making characteristics, it seems likely that lex talionis will provide a measure either too vague to be of much help (see Shafer-​Landau 1996: 299–​302; 2000: 197–​198), or too specific to be plausible (at least in some cases). (Walen 2014)

Ordinal proportionality, on the other hand, faces a different problem:

Free Will Skepticism and Criminal Justice    237 If all that were required to do justice is to rank order wrongs by their gravity and then provide a mapping onto a range of punishments that likewise went from lighter to more serious—​respecting the norms of rank-​ordering and parity—​then neither the range of punishments from a fine of $1 up to a fine of $100, nor from 40 years to 60 years in prison, would provide disproportionate punishment, no matter what the crimes. This seems wrong. Murder should not be punished with a $100 fine, and littering should not be punished with 40 years in prison. Some vague degree of cardinality therefore seems to be called for, punishing grave wrongs with heavy penalties and minor wrongs with light penalties. (Walen 2014)

Such problems reveal that the principle of proportionality is too ambiguous to guarantee respect for persons since it is unable to draw a clear line in the sand between deserved punishments on the one hand and cruel and inhumane punishment on the other. As a result, cultural and societal pressure can easily affect how gravity and proportional punishment are measured, and this can easily lead (as highlighted above) to excessively punitive forms of punishment. Lastly, while rejecting the retributivist principle of proportionality, the public health-​ quarantine model has a proportionality principle of its own. It maintains that criminal sanctions should be proportionate to the danger posed by an individual, and any sanctions that exceed this upper bound will be unjustified. This is coupled with the principle of least infringement, which holds that the least restrictive measures should be taken to protect public health and safety. Together these two principles set strict limits on how individuals can and should be treated. Consider again the hypothetical scenario used in the Shariff et al. study. The fictional case involved an offender who beat a man to death but after serving two years in prison was nearly 100% effectively rehabilitated. The case further stipulated that “the prosecution and defense had agreed that the rehabilitation would prevent recidivism and that any further detention after rehabilitation would offer no additional deterrence of other potential criminals” (Shariff et al. 2014: 4). On my model, it would be unjust to continue to incapacitate this individual. Retributivists, on the other hand, will generally feel that this person deserves to be punished further since two years in prison is not proportional punishment—​although, as my comments on the proportionality principle above indicate, they will likely disagree on exactly what this additional punishment should amount to. Which of these views better respects human dignity? I have a hard time seeing how punishing someone who is no longer a threat to society, and in a way that exceeds effectiveness, respects human dignity. Instead, I maintain that the public health-​quarantine model actually respects human dignity more since it specifies that (a) individuals who are not a serious threat to society should not be incapacitated; (b) no one should be incapacitated longer than is absolutely necessary (where this is determined by the continued threat the individual poses to society); and (c) when it is necessary to incapacitate an individual, we must do so in a way that treats them humanely, with respect and dignity, and with rehabilitation as our goal. There is more to treating criminals with dignity than giving them their just deserts, even assuming the defensibility of retributivism. Human

238   Gregg D. Caruso dignity also demands that we not dehumanize, disenfranchise, and treat cruelly those we imprison. The public health-​quarantine model, I contend, does a better job at respecting human dignity since it prioritizes prevention, rehabilitation, and reintegration, and demands a level of concern for the well-​being of prisoners. Furthermore, the public health-​quarantine model can also respect human dignity by prohibiting the incapacitation of innocent people (see Pereboom and Caruso 2018). Neither Pereboom nor I set out our position in a strict consequentialist theoretical context. Rather, we justify incapacitation on the ground of the right to self-​defense and defense of others. That right does not extend to people who are non-​threats. It would therefore be wrong to incapacitate someone who is innocent, since they are not a serious threat to society. The aim of protection is justified by a right with clear bounds, and not by a consequentialist theory on which the bounds are unclear (see Pereboom and Caruso 2018). For this reason, the public health-​quarantine model provides a distinct advantage over consequentialist deterrence theories since it has more restrictions placed on it with regard to using people merely as a means. Concerns over the “use” objection, for example, count more heavily against punishment policy justified simply on consequentialist grounds than they do against incapacitation based on the quarantine analogy (see also Pereboom 2014, 2017a).

4.  Victims’ Rights A second objection is that victims of violent crime will never receive proper justice or satisfaction on my account since it rejects harsh punishment in favor of rehabilitating dangerous criminals and implementing the least restrictive forms of sanctions needed to secure public safety. Democratic Senator Dianne Feinstein and Republican Senator Jon Kyl have argued, for instance, that “for too long, our court system has tilted in favor of accused criminals and has proven appallingly indifferent to the suffering of crime victims.”6 I think this is a gross misrepresentation of the US criminal justice system over the last few decades—​evidenced by our current mass incarceration crisis, the heavy-​ handedness of mandatory minimums, the increased use of plea bargains, and the three-​ strikes-​you-​are-​out laws that have swept the nation—​but I mention it because it captures a common concern critics have with reformist proposals like my own. The concern is that such models put the rights of criminals above the concerns of victims, and worse still advocate for reforms that run contrary to the concerns of victims. While I take this objection seriously, I do not think the public health-​quarantine model is “indifferent to the suffering of crime victims.” Rather, I maintain that it better reflects the attitudes and preferences of most victims and does a better job preventing future victims. 6  As reported on in the Washington Post (2016): https://​www.washingtonpost.com/​news/​wonk/​wp/​ 2016/​08/​05/​even-​violent-​crime-​victims-​say-​our-​prisons-​are-​making-​crime-​worse

Free Will Skepticism and Criminal Justice    239 First, I contend that this objection is predicated on a mistaken assumption. The underlying assumption seems to be that most victims of violent crime want retribution above all else and that to deny them the satisfaction of seeing their perpetrators suffer is an injustice. Proponents of the death penalty and other forms of excessively punitive forms of punishment typically argue, for instance, that whatever deterrence factor such punishment may or may not have, such punishment provides justice for the victims and their families since it satisfies their desire for retribution and proportional punishment. Kant, for example, famously argued that if a people do not insist on the execution of murders, “blood guilt” would “cling” to them “as collaborators in this public violation of justice” (1797: 142). Setting aside the issue of what counts as proportional punishment raised above, it is an empirical question what victims actually want, what their preferences and attitudes are, and what kind of justice they would like to see from the criminal justice system. Fortunately, the Alliance for Safety and Justice has recently investigated exactly these questions. In its first-​of-​its-​kind national survey, they found that victims of violent crime say they want to see shorter prison sentences, less spending on prisons, and a greater focus on the rehabilitation of criminals (2016). The survey polled the attitudes and beliefs of more than 800 crime victims pooled from a nationally representative sample of over 3,000 respondents. According to the report: Perhaps to the surprise of some, victims overwhelmingly prefer criminal justice approaches that prioritize rehabilitation over punishment and strongly prefer investments in crime prevention and treatment to more spending on prisons and jails. These views are not always accurately reflected in the media or in state capitols and should be considered in policy debates. (2016: 4)

An examination of the data reveals that victims prefer an approach much closer to the public health-​quarantine model, with its focus on prevention, social justice, and rehabilitation, than retributivism. For instance, the survey found that: • By a 2 to 1 margin, victims prefer that the criminal justice system focus more on rehabilitating people who commit crimes than punishing them. • By a margin of 15 to 1, victims prefer increased investments in schools and education over more investments in prisons and jails. • By a margin of 10 to 1, victims prefer increased investments in job creation over more investments in prisons and jails. • By a margin of 7 to 1, victims prefer increased investments in mental health treatment over investments in prisons and jails. • By a margin of nearly 3 to 1, victims believe prison makes people more likely to commit crimes than to rehabilitate them. • By a margin of 7 to 1, victims prefer increased investments in crime prevention and programs for at-​risk youth over more investments in prisons and jails.

240   Gregg D. Caruso • 6 in 10 victims prefer shorter prison sentences and more spending on prevention and rehabilitation to prison sentences that keep people incarcerated for as long as possible. • By a margin of 4 to 1, victims prefer increased investments in drug treatment over more investment in prisons and jails. • By a margin of 2 to 1, victims prefer increased investments in community supervision, such as probation and parole, over more investments in prisons and jails. • 7 in 10 victims prefer that prosecutors focus on solving neighborhood problems and stopping repeat crime through rehabilitation, even if it means fewer convictions and prison sentences. • 6 in 10 victims prefer that prosecutors consider victims’ opinions on what would help them recover from the crime, even when victims do not want long prison sentences. The report also found that victims’ views remained consistent across demographics—​ that is, for each of the questions above, they found majority or plurality support across demographic groups, including age, gender, race, ethnicity, and political party affiliation. This skepticism of prisons is in line with most social science research, which has generally shown that mass incarceration causes more crime than it prevents, that institutionalizing young offenders makes them more likely to commit crime as adults, and that spending time in prison teaches people how to be better criminals (see, e.g., Weatherburn 2010). It would seem, then, that those tough-​on-​crime proponents who invoke the names of victims of violent crime and claim to speak for them, such as Feinstein and Kyl, often misrepresent their actual preferences, attitudes, and desires. To say that approaches like the public health-​quarantine model are “appallingly indifferent to the suffering of crime victims” is to discount what victims say they actually want. It also overlooks the fact that the best way to reduce crime, and the suffering caused by it, is to (a) prevent the crime from occurring in the first place by addressing the causal determinates of crime, and (b) to rehabilitate criminals so as to reduce the likelihood of recidivism. The public health-​quarantine model attempts to do both; retributivism by its very nature does neither. Since retributivism myopically focuses on justifying backward-​looking blame and punishment, it does not have the resources needed to address rehabilitation or preventative measures. I question, then, the claim that retributivism reflects a deeper concern for victims and their families. If one really cares about victims and their suffering, the best way to honor this concern is to reject retributivism and adopt a more holistic approach to criminal behavior that focuses on preventing crime, rehabilitating criminals, and reducing the number of people who become victims of violent crime. Second, even if victims of violent crime wanted to see criminals suffer and were on the whole indifferent to concerns about safety and rehabilitation—​contrary to what appears to be the case—​we should not conclude from this that we should inflict such harm and suffering, nor that denying victims the satisfaction of seeing their perpetrators suffer would be a violation of their rights. As Walen accurately points out, “the view that it

Free Will Skepticism and Criminal Justice    241 wrongs victims not to punish wrongdoers confuses vengeance, which is victim-​centered, with retributivism, which is agent-​centered: concerned with giving the wrongdoer the punishment he deserves” (2014). Paul Robinson (2008), for instance, has argued that retributivists must distinguish between vengeful and deontological conceptions of deserved punishment. The former urges punishing an offender in a way that mirrors the harm or suffering he/​she has caused: Because of this focus on the harm done, the vengeful conception of desert is commonly associated with the victim’s perspective. Retributive justice “consists in seeking equality between offender and victim by subjecting the offender to punishment and communicating to the victim a concern for his or her antecedent suffering” [(Fletcher 1999: 58).] . . . And the association with the victim’s suffering, in turn, associates vengeful desert with the feelings of revenge and hatred that we commonly see in victims. Thus, punishment under this conception of desert is sometimes seen as essentially an institutionalization of victim revenge; it is “injury inflicted on a wrongdoer that satisfies the retributive hatred felt by the wrongdoer’s victim and that is justified because of that satisfaction” [(Feinberg and Coleman 2000: 793)]. (Robinson 2008: 147–​148)

The problem, however, is that justifying punishment on the grounds of vengefulness or the satisfaction of retributive hatred fails to take into account the blameworthiness of the offender. The deontological conception of desert, alternatively, focuses at least not on the harm of the offense but on the blameworthiness of the offender, as drawn from the arguments and analysis of moral philosophy (Robinson 2008: 148). Thus, the criterion for assessing punishment is broader and richer than that for vengeful desert: Anything that affects an offender’s moral blameworthiness is taken into account in judging the punishment he deserves. The extent of the harm caused or the seriousness of the evil done will be part of that calculation but so too will be a wide variety of other factors, such as the offender’s culpable state of mind or lack thereof and the existing conditions at the time of offence, including those that might give rise to claims of justification, excuse, or mitigation. (Robinson 2008: 148)

To the extent, then, that retributivists want to appeal to moral blameworthiness rather than vengeful desires in justifying punishment, denying victims the vengeful satisfaction they seek would not be a violation of their rights. This brings me to my next reply. Punishment inflicts harm on individuals, and the justification for such harm must meet a high epistemic standard. If it is significantly probable that one’s justification for harming another is unsound, then, prima facie, that behavior is seriously wrong (Pereboom 2017a). But if free will skeptics are right, neither libertarians nor compatibilists satisfy this epistemic standard, and hence individuals do not justly deserve to be punished. And if individuals do not justly deserve to be punished, there is no violation of the rights of victims to deny them the revenge they seek. Even retributivists would acknowledge that the desire for revenge and retribution

242   Gregg D. Caruso has its limits. The principle of proportionality, despite its weaknesses, dictates that punishments that are disproportionate to the wrong done (whatever that ultimately amounts to) would be unjustified. Hence, if the victim of an armed robbery wanted to see their perpetrator executed, and this was deemed disproportionate punishment by the standards of retributivism, it would not be a violation of the victim’s rights on that theory to prohibit said execution. By extension, if free will skeptics are right, and retributive punishment itself is unjustified, then to deny victims their desire for revenge (conceived here in a purely backward-​looking, non-​consequentialist sense) would likewise not be a violation of their rights. For victims to have the right to see suffering and harm imposed on their perpetrators, it would have to be the case that such harm was justified. According to free will skeptics, however, neither victims of violent crime nor the state acting on their behalf are justified in causing more harm than is minimally required for adequate protection. Lastly, the public health-​quarantine model is able, I contend, to deal with the concerns of victims, acknowledge the wrongs done them, and help aid in recovery. As Pereboom and I have argued: Accepting free will skepticism requires rejecting our ordinary view of ourselves as blameworthy or praiseworthy in the basic desert sense. A critic might first object that if we gave up this belief, we could no longer count actions as morally bad or good. In response, even if we came to hold that a serial killer was not blameworthy due to a degenerative brain disease, we could still justifiably agree that his actions are morally bad. Still, secondly, the critic might ask, if determinism precluded basic desert blameworthiness, would it not also undercut judgments of moral obligation? If “ought” implies “can,” and if because determinism is true an agent could not have avoided acting badly, it would be false that she ought to have acted otherwise. Furthermore, if an action is wrong for an agent just in case she is morally obligated not to perform it, determinism would also undermine judgements of moral wrongness (Haji 1998). In response, we contend that even if the skeptic were to accept all of this (and she might resist at various points; cf. Pereboom 2014: ch.6; Waller 2011), axiological judgments of moral goodness and badness would not be affected (Haji 1998; Pereboom 2001). So, in general, free will skepticism can accommodate judgments of moral goodness and badness, which are arguably sufficient for moral practice. (Pereboom and Caruso 2018: 200)

There is nothing preventing free will skeptics, then, from acknowledging the moral wrongness of criminal acts. There is also nothing preventing them from acknowledging the harm done to victims by these morally bad acts. Given that free will skeptics can retain axiological judgments of moral goodness and badness, the public health-​quarantine model can recommend that one way to help aid victims in recovery is to have the wrong done to them acknowledged and a commitment made to rehabilitate the offender and protect others from similar crimes. This brings me to my final point. On the forward-​looking account of moral responsibility developed by Pereboom (2013, 2014), non-​desert-​based blame and the

Free Will Skepticism and Criminal Justice    243 acknowledgment of wrong can be used for the purposes of protection, moral formation, and reconciliation. This forward-​looking approach to moral responsibility can be used to aid victims in their recovery and perhaps even achieve some form of reconciliation. Restorative justice models, for example, have been employed around the country over the last few decades with great success (see, e.g., Camp et al. 2013; Walgrave 2002). Restorative justice is an approach that emphasizes repairing the harm caused by criminal behavior by bringing together members of the community, victims, and offenders. As the Centre for Justice and Reconciliation describe it: Restorative justice views crime as more than breaking the law—​it also causes harm to people, relationships, and the community. So a just response must address those harms as well as the wrongdoing. If the parties are willing, the best way to do this is to help them meet to discuss those harms and how to bring about resolution. Other approaches are available if they are unable or unwilling to meet. Sometimes those meetings lead to transformational changes in their lives.7

The restorative approach maintains that the best way to repair the harms caused by criminal behavior is to bring together all stakeholders for the purpose of making amends and reintegration. It focuses on repairing the harm caused by crime and reducing future harm through crime prevention. Now it is true that most restorative justice methods require offenders to take responsibility for their actions and for the harm they have caused, but such responsibility need not be conceived in terms of basic desert. Most current restorative justice models probably do assume backward-​looking blame and basic desert moral responsibility (e.g., Sommers 2016), but these are not essential components of the restorative approach. The same ends, I contend, can be achieved on a model that does not appeal to basic desert moral responsibility. A conversational model of forward-​looking moral responsibility like that proposed by Pereboom (2013, 2014) could, for example, serve as a basis for an exchange between victim and offender in a way that does not invoke backward-​looking blame or basic desert. Such an exchange could aid both in the rehabilitation of offenders and in the recovery of victims. To use slightly different lingo, we can say that a restorative justice model consistent with free will skepticism could appeal to answerability and attributability conceptions of moral responsibility rather than accountability.

5.  Conclusion In this chapter I have argued that the public health-​quarantine model not only justifies the incapacitation of dangerous criminals, it also demands that we identify and 7 

http://​restorativejustice.org/​restorative-​justice/​about-​restorative-​justice/​tutorial-​intro-​to-​ restorative-​justice/​lesson-​1-​what-​is-​restorative-​justice/​

244   Gregg D. Caruso take action on the social determinants of criminal behavior—​such as poverty; socioeconomic status; abuse and violence in the home; housing; mental health; access to healthcare, education, environmental health, and nutrition. I argued that the right to self-​protection and prevention of harm to others justifies incapacitating the criminally dangerous with the minimum harm required for adequate protection. But I also argued that a more comprehensive approach to criminal justice, one which views public safety as akin to public health, demands that we reject retributivism and purely punitive approaches to criminal justice and shift the focus to prevention, rehabilitation, and reintegration. The public health-​quarantine model prioritizes prevention and social justice and maintains that the goal of the criminal justice system should be the rehabilitation and reintegration of offenders back into society. Along the way, I also responded to several concerns about proportionality, human dignity, victims’ rights, and the incapacitation of innocent people. I argued that each of these concerns can be met and that the public health-​quarantine model is consistent with free will skepticism, morally justifiable, and sufficient for good social policy. If what I’ve argued is correct, the public health-​quarantine model offers free will skeptics a way forward and a suitable conception of justice without retribution.

References Alexander, L., K. Ferzan, and S. Morse. 2009. Crime and Culpability: A Theory of Criminal Law. New York: Cambridge University Press. Alliance for Safety and Justice. 2016. Crime survivors: The first-​ever national survey of victims’ views on safety and justice. Report available: https://​www.allianceforsafetyandjustice.org/​ wp-​content/​uploads/​documents/​Crime%20Survivors%20Speak%20Report.pdf Berman, M. N. 2008. Punishment and justification. Ethics 18: 258–​290. Camp, V. et al. 2013. Victim satisfaction with restorative justice: More than simply procedural justice. International Review of Victimology 19(2): 117–​143. Carey, Jasmine M., and Delroy L. Paulhus. 2013. Worldview implication of believing in free will and/​or determinism: Politics, morality, and punitiveness. Journal of Personality 81(2): 130–​141. Caruso, G. D. 2012. Free Will and Consciousness: A Determinist Account of the Illusion of Free Will. Lanham, MD: Lexington Books. Caruso, G. D. (Ed.). 2013. Exploring the Illusion of Free Will and Moral Responsibility. Lanham, MD: Lexington Books. Caruso, G. D. 2015. Kane is not able: A reply to Vicens’ “Self-​forming actions and conflicts of intention.” Southwest Philosophy Review 31(2): 21–​26. Caruso, G. D. 2016. Free will skepticism and criminal behavior: A public health-​quarantine model. Southwest Philosophy Review 32(1): 25–​48. Caruso, G. D. 2017a. Public Health and Safety: The Social Determinants of Health and Criminal Behavior. UK: ResearchersLinks Books. Caruso, G. D. 2017b. Free will skepticism and the question of creativity: Creativity, desert, and self-​creation. Ergo 3(23): 23–​39. Caruso, G. D. 2019. Free will skepticism and its implications: An argument for optimism. In Elizabeth Shaw (Ed.), Free Will Skepticism in Law and Society. New York: 43–​72.

Free Will Skepticism and Criminal Justice    245 Caruso, G. D., and S. Morris. 2017. Compatibilism and retributivist desert moral responsibility: On what is of central philosophical and practical importance. Erkenntnis 82(4):837–​855. DOI 10.1007/​s10670-​016-​9846-​2. Caruso, G.D. 2021. Rejecting Retributivism: Free Will, Punishment, and Criminal Justice. New York: Cambridge University Press. Corrado, M. L. 2016. Two models of criminal justice. Available at SSRN: http://​ssrn.com/​ abstract=2757078 Crick, Francis (1994). The Astonishing Hypothesis: The Scientific Search for the Soul. Scribners. Double, R. 1991. The Non-​reality of Free Will. Oxford: Oxford University Press. Feinberg, J. 1995. Equal punishment for failed attempts: Some bad but instructive arguments against it. Arizona Law Review 37: 117–​133. Haji, I. 1998. Moral Accountability. New York: Oxford University Press. Haji, I. 2001. Control conundrums: Modest libertarianism, responsibility, and explanation. Pacific Philosophical Quarterly 82: 178–​200. Hart, H. L. A., 1968. Punishment and Responsibility. Oxford: Oxford University Press. Husak, D. 2000. Holistic retributivism. California Law Review 88: 991–​1000. Kant, Immanuel. 1790. The Metaphysics of Morals. Translated by W. Hastie. Edinburgh: Clark (1887). Kant, Immanuel. 1797. The Metaphysics of Morals. Translated by M. Gregor. New York: Cambridge University Press (1991). Lemos, John. 2013. Freedom, Responsibility, and Determinism: A Philosophical Dialogue. Indianapolis, IN: Hackett. Lemos, J. 2016. Moral concerns about responsibility denial and the quarantine of violent criminals. Law and Philosophy 35(5): 461–​483. Levy, K. 2005. The solution to the problem of outcome luck: Why harm is just as punishable as the wrongful action that causes it. Law and Philosophy 24: 263–​303. Levy, N. 2008. Bad luck once again. Philosophy and Phenomenological Research 77(3): 749–​754. Levy, N. 2011. Hard Luck: How Luck Undermines Free Will and Moral Responsibility. New York: Oxford University Press. Lewis, C. S. 1971. The humanitarian theory of punishment. In God in the Dock: Essays in Theology and Ethics, pp. 287–​294. Cambridge: William B. Eerdmans. Mele, A. 1999. Kane, luck, and the significance of free will. Philosophical Explorations 2: 96–​104. Mele, A. 2008. Manipulation, compatibilism, and moral responsibility. The Journal of Ethics 12: 263–​286. Mele, A. 2017. Aspects of Agency. New York: Oxford University Press. Morris, H. 1968. Persons and punishment. Monist 52: 475–​501. Nussbaum, M. 2011. Creating Capabilities: The Human Development Approach. Cambridge, MA: Harvard University Press. Oldenquist, A. 1988. An explanation of retribution. Journal of Philosophy 85: 464–​478. Paulhus, D. L., & Carey, J. M. (2011). The FAD-​Plus: measuring lay beliefs regarding free will and related constructs. Journal of personality assessment, 93(1), 96–​104. Pereboom, D. 1995. Determinism al dente. Noûs 29: 21–​45. Pereboom, D. 2001. Living Without Free Will. New York: Cambridge University Press. Pereboom, D. 2013. Free will skepticism and criminal punishment. In Thomas Nadelhoffer (ed.), The Future of Punishment, pp. 49–​78. New York: Oxford University Press. Pereboom, D. 2014. Free Will, Agency, and Meaning in Life. Oxford: Oxford University Press.

246   Gregg D. Caruso Pereboom. D. 2017a. A defense of free will skepticism: Replies to commentaries by Victor Tadros, Saul Smilansky, Michael McKenna, and Alfred R. Mele on Free will, agency, and meaning in life. Criminal Law and Philosophy 11(3): 617–​636. Pereboom, D. 2017b. Responsibility, agency, and the disappearing agent objection. In Jean-​ Baptiste Guillon (ed.), Le Libre-​Arbitre: Approaches Contemporaries, Paris: College de France. Pereboom, D., and G. D. Caruso. 2018. Hard-​incompatibilist existentialism: Neuroscience, punishment, and meaning in life. In G. D. Caruso and O. Flanagan (eds.), Neuroexistentialism: Meaning, Morals, and Purpose in the Age of Neuroscience, pp. 193–​222. New York: Oxford University Press. Powers, M., and R. Faden. 2006. Social Justice: The Moral Foundations of Public Health and Health Policy. New York: Oxford University Press. Robinson, P. H. 2008. Competing conceptions of modern desert: Vengeful, deontological, and empirical. Cambridge Law Journal 67(1): 145–​175. Sen, A. 1985. Commodities and Capabilities. Amsterdam: North-​Holland. Sen, A. 1999. Development as Freedom. New York: Oxford University Press. Shafer-​Landau, R. 1996. The failure of retributivism. Philosophical Studies 82: 289–​316. Shariff, A. F., J. D. Greene, J. C. Karremans, J. Luguri, C. J. Clark, J. W. Schooler, R. F. Baumesiter, and K. D. Vohs. 2014. Free will and punishment: A mechanistic view of human nature reduces retribution. Psychological Science, published online June 10: 1–​8. Smilansky, S. 2000. Free Will and Illusion. New York: Oxford University Press. Smilansky, S. 2011. Hard determinism and punishment: A practical reductio. Law and Philosophy 39: 353–​367. Smilansky, S. 2017. Pereboom on punishment: Funishment, innocence, motivation, and other difficulties. Criminal Law and Philosophy 11(3): 591–​603. Sommers, Tamler. (2016). The three r’s: Retribution, revenge, and reparation. Philosophia 44: 327–​342. Strawson, G. 1986. Freedom and Belief. Oxford: Oxford University Press. Strawson, G. 1994. The impossibility of moral responsibility. Philosophical Studies 75(1): 5–​24. Tabandeh, A., P. Gardoni, and C. Murphy. 2018. A reliability-​based capability approach. Risk Analysis 38(2): 410–​424. Tadros, V. 2017. Doing without desert. Criminal Law and Philosophy 11(3): 605–​616. Todd, P. 2011. A new approach to manipulation arguments. Philosophical Studies 152(1): 127–​133. Todd, P. 2013. Defending (a modified version of) the Zygote Argument. Philosophical Studies 164(1): 189–​203. Waldron, J. 1992. Lex Talionis. Arizona Law Review 34: 25–​51. Walen, A. 2010. Crime, culpability and moral luck: Comment on Alexander, Ferzan and Morse. Law and Philosophy 29: 373–​384, Walen, A. 2014. Retributive justice. Stanford Encyclopedia of Philosophy (online). http://​ plato.stanford.edu/​entries/​justice-​retributive/​#Pro Walgrave, Lode. (Ed.) 2002. Restorative Justice and the Law. Devon, UK: Willan Publishing. Waller, B. 1990. Freedom Without Responsibility. New York: Temple University Press. Waller, B. 2011. Against Moral responsibility. Cambridge, MA: MIT Press. Waller, B. 2014. The culture of moral responsibility. Southwest Philosophical View 30(1): 3–​17. Waller, B. 2015. The Stubborn System of Moral Responsibility. Cambridge, MA: MIT Press. Weatherburn, Don. 2010. The effects of prison on adult re-​offending. Crime and Justice Bulletin NSW Bureau of Crime Statistics and Research: Contemporary Issues in Crime and Justice 143: 1–​12.

CHAPTER 12

M etaskep t i c i sm Tamler Sommers*

The statement, therefore, that men can be held responsible solely for ­individual conduct freely willed is certainly wrong; it mistakes a principal  characteristic of individualistic ages for an eternal law of human nature. —​Heinrich Gomperz1

Philosophers have approached the problem of moral responsibility systematically, developing theories with conditions that must be met for agents to deserve blame and praise, and perhaps punishment and reward, for their actions. They interpret the meaning of “deserve” in the non-​consequentialist sense that is common in the literature: An agent deserves blame [praise, punishment, reward] if and only if it is fair to blame A [praised, punished, rewarded] independent of any practical or consequentialist benefits that may arise from doing so. Although the term “moral responsibility” is used in many other ways, it is this desert-​entailing sense that is at the center of the contemporary philosophical debate.2 Theories of responsibility have two features in common. First, they posit conditions or criteria for moral responsibility that are meant to apply universally, for all agents, for all societies. Second, they appeal to intuitions to justify these conditions. Consequently, the theories rely at least implicitly on empirical assumptions concerning the universality of certain attitudes and intuitions. In Sommers (2012) I develop an empirical and philosophical challenge to these assumptions. First, I offer evidence of radical differences in attitudes and intuitions about moral responsibility across cultures and throughout history. And I show that it is implausible to think that the variation in intuitions can be explained away as the product of irrationality or disagreement over non-​moral facts. I conclude that we have no principled way of establishing *  Thanks to Dana Nelkin for valuable critical feedback on this chapter. 1 

Gomperz, 1939, p. 342. Recent work by Pereboom and others have further narrowed down the notion of “basic desert” to exclude contractualist as well as consequentialist considerations. See e.g. Pereboom (2014). 2 

248   tamler sommers universally true conditions of moral responsibility, and that the compatibility question (an obsession of the contemporary philosophical literature) has no objectively correct answer. I called the position “metaskepticism” about moral responsibility, because if the conclusion is true it would undermine not only positive theories of responsibility, but also hard incompatibilist or skeptical theories which argue that human beings everywhere cannot deserve praise or blame for their actions.3

1.  The Appeal to Intuition When I embarked on this project, I would often hear objections along the lines of “who cares about people’s intuitions about free will and moral responsibility? I’m interested in the truth about free will and moral responsibility. Your account doesn’t tell me anything about that.” This objection, familiar to many experimental or empirically inclined philosophers, might appear to have a sensible ring to it. In debates over, say, group selection in evolutionary theory, we don’t examine folk intuitions or historical attitudes. We just develop theories and see how they hold up. Why shouldn’t philosophers regard free will and moral responsibility the same way? The answer is simple: Unlike evolutionary biologists, philosophers have thus far investigated the nature of their topic through an appeal to the intuitions of their audience. Arguments for incompatibilism typically employ at least one of two key principles to show that the conditions for moral responsibility cannot be satisfied if determinism is true: (1) the Principle of Alternate Possibilities (PAP) which states that we cannot be morally responsible for an act if we lack the ability to do otherwise; and (2) the “Transfer of Non-​Responsibility” (TNR) principle, which states (roughly) that we cannot be morally responsible for an act if we are not morally responsible for any of the determining factors of that act. (The non-​responsibility for the determining factors “transfers” to the act itself.)4 In support of these principles, incompatibilists often appeal directly to their own intuitions and the intuitions of the reader. Van Inwagen (1983), for example, offers the following defense of the TNR principle (or what he calls “Rule B”): I must confess that my belief in the validity of [The TNR principle] has only two sources, one incommunicable and the other inconclusive. The former source is what philosophers are pleased to call “intuition” . . . latter source is the fact that I can think of no instances of [the TNR Principle] that have, or could possibly have, true premises and a false conclusion. (Van Inwagen 1983, pp. 97–​99) 3  With benefit of hindsight, I wish I had chosen a different label—​relativism or even better pluralism about moral responsibility. 4  Fischer and Ravizza (1998) provide the following more careful formulation of this principle: “(1) p obtains and no one is even partly morally responsible for p; and (2) if p obtains, then q obtains, and no one is even partly morally responsible for the fact that if p obtains, then q obtains; then (3) q obtains, and no one is even partly morally responsible for q.” (p. 152).

Metaskepticism   249 Van Inwagen’s two sources are in fact nearly identical, since someone with different intuitions from van Inwagen’s could come up with counterexamples rather easily. And indeed, the evidence suggests that van Inwagen’s intuitions are more the exception than the rule, confined mostly to WEIRD (Western Educated Industrialized Rich Democratic) societies.5 Other incompatibilists (e.g., Pereboom 1995, 2001) describe specific cases in which an agent is intuitively not morally responsible and then argue that there are no relevant differences between those cases and all instances of fully determined human behavior. But these “generalization strategies” (as R.J. Wallace (1994) has called them) cannot get off the ground unless the reader shares the starting intuitions that the agents in the original cases are not morally responsible for their behavior. Compatibilists are no less reliant on appeals to intuitions, both to develop counterexamples to incompatibilist principles6 and to develop sufficient conditions for morally responsible behavior. Consider, for example, Susan Wolf ’s argument for the “Sane Deep Self View.” Wolf relates the case of JoJo, the son of a brutal dictator, who has been trained from early childhood to value arbitrary expressions of cruelty, such as executing and torturing people for no reason at all. Consequently, JoJo acquires a character that both values cruel behavior, and endorses it upon reflection, thus satisfying the conditions for blameworthiness in the compatibilist accounts of Frankfurt and Watson.7 According to Wolf, however, “in light of JoJo’s heritage and upbringing, it is dubious at best that he should be regarded as responsible for what he does” (1987, pp. 379–​380). This leads Wolf to add another condition to the Deep Self View: the “condition of sanity,” the capacity to understand the difference between morally right and wrong actions. Wolf ’s argument has two premises that require intuitive agreement. First, readers must share with Wolf the (quite controversial) intuition that JoJo is not morally responsible for his cruel behavior because of his upbringing. Second, we must agree that the addition of the “sanity requirement” provides the Deep Self View with sufficient conditions for moral responsibility. And of course, to argue against Wolf ’s “sane deep self ” theory, someone must come up with an intuitively plausible counterexample. The cycle continues. 5  This term comes from Henrich et al. (2010). See Sommers (2012) for a survey of research suggesting that the emphasis on control and sourcehood as a condition for moral responsibility is more prominent in individualist societies in the modern West. I discuss some of this research below. Of course, there is variation within individualist WEIRD societies too. Sommers (2018), for example, examines many examples of honor cultures in the American South, as well as the military, team sports, and urban gang culture. 6  Frankfurt’s (1969) famous case of the assassin, for instance, can only be a successful counterexample to the principle of alternate possibilities if we find the assassin intuitively blameworthy when there is no interference. Fischer and Ravizza employ the same strategy in their arguments against the “transfer of non-​responsibility” principle (see Ravizza (1994), Fischer and Ravizza (1998); and the “erosion” and erosion* cases). 7  Though the details of Frankfurt and Watson’s accounts differ, they share the basic view that we can be morally responsible for actions that are reflectively endorsed by a deeper self than a mere first-​order desire.

250   tamler sommers

1.1  Must Theories of Responsibility Appeal to Intuition? One might concede that most theories in the contemporary debate appeal to intuition and reply “so much the worse for those theories—​moral responsibility “done right” will not involve such an appeal.” But we would then need a sense of what this intuition-​free approach to responsibility would look like—​and this is a difficult and perhaps impossible task. A priori approaches that attempt to conceptually link moral responsibility to necessary and sufficient conditions seem doomed to failure. Human beings competently employ the concept of moral responsibility while at the same time disagreeing about conditions of its application. It’s not clear how one might go about defending something like the sanity condition without appealing to intuition somewhere along the line. Even the most “self-​evident” condition such as a basic control condition is difficult to support on a priori grounds. Throughout history human beings have employed, and continue to employ, the concept of moral responsibility (in the desert-​entailing sense) in cases where control is absent. A common interpretation of the doctrine of original sin, for example, is that human beings are guilty, and deserve punishment, for the actions of Adam and Eve. Martin Luther explicitly defends the view that human beings can deserve damnation even though they have no control over their sinful nature. You may find this wildly counterintuitive, but there’s nothing contradictory about this description of moral responsibility and desert. The same is certainly true for the more hotly contested conditions (such as sanity, indeterministic deliberation, self-​determination) that characterize the contemporary debate. One account that arguably does not rely on intuitions or folk beliefs is P. F. Strawson’s in “Freedom and Resentment.” Strawson grounds our attributions of responsibility in the human attitudes and practices that we express in our relationships. For Strawson, blame and praise judgments are expressions of attitudes like resentment, gratitude, guilt, and pride. Many contemporary Strawsonians reject the extreme naturalism of his account, claiming that it blurs the distinction between being responsible and holding responsible.8 They agree that responsibility has important connections with the reactive attitudes, but argue that responsibility judgments should be seen as judgments about whether the expressions are “fair,” “appropriate,” or “fitting.”9 This move takes them back to intuition-​grounded accounts, intuitions about when an attitude like resentment is fair or fitting. But if we accept Strawson’s approach in all of its naturalistic glory, then the metaskeptical position becomes easier to defend. Attitudes and practices about blame vary tremendously across cultures and throughout history, as do the kinds of situations that trigger resentment. The nature of these differences would make universal judgments about the conditions for moral responsibility impossible. Strawson might

8 

9 

See McKenna and Russell (2008). See, for example, Wallace (1994) and Shoemaker (2015).

Metaskepticism   251 welcome this relativistic implication of his theory, in which case my thesis would be consistent with his position and perhaps a development of it.10

2.  The Empirical Challenge to Universalist Theories The empirical challenge has two parts. The first provides evidence of radical differences in attitudes and intuitions about the conditions for moral responsibility. The second argues that the variation cannot be explained away as the product of irrationality or ignorance of the relevant empirical facts. The challenge is in one sense a version of the argument from disagreement familiar in metaethical debates (e.g. Mackie 1983; Loeb 1998; and Doris and Plakias 2007).11 But there are some important differences as well. Philosophers working on free will and moral responsibility proceed almost exclusively by appealing to intuitions about principles and cases, and then developing theories to accommodate the intuitions. The metaethical literature features a far more diverse set of methodologies for defending moral realism, including those that do not rely on appeals to moral intuitions. Since theories of moral responsibility are intuition driven, the existence of deep variation in intuitions raises significant challenges. Furthermore, the metaethical literature features an enormous number of articles and books that focus on the issue of disagreement. Indeed, some of the new methodologies in metaethics have been developed in part to deal with the challenge. By contrast, as far as I know, there is not a single article or book in the last 50 years that addresses the challenge of disagreement in the free will and moral responsibility debate. (Bernard Williams’s excellent book Shame and Necessity may be an exception, but it is not specifically directed at the free will problem.) One of my goals in developing a detailed version of such a challenge is to encourage philosophers working on responsibility to grapple with the problem of disagreement—​as moral realists of all stripes (intuitionists, rationalists, and naturalists) have done in the metaethical literature. (See, e.g., Audi 2005, 2008; Brink 1984; Boyd 1997; Smith 1994). Finally, the success of the metaskeptical challenge regarding moral responsibility is not dependent on the success of broader metaethical challenges to moral realism for the simple reason that the metaskeptical challenge relies on a smaller set of empirical claims. Arguments from disagreement in metaethics attempt to establish a more sweeping and ambitious conclusion than the one described here. Certain areas of morality may admit of a realist treatment while others may not. Perhaps there would be agreement under ideal conditions of rationality about certain moral questions (the

10  Revisionism, as defended by Vargas (2013), is likely also compatible with the broader metaskeptical thesis. 11  See Mackie (1977). And see Loeb (1998), Doris and Stich (2005), and Doris and Plakias (2007) for excellent discussions of the role of empirical research in these arguments.

252   tamler sommers wrongness of causing unnecessary suffering), but not others (whether it is permissible to sacrifice one life to save five). In these cases, the proper metaethical conclusion may then be what Doris and Plakias (2007) term a “patchy moral realism.” The metaskeptical challenge, however, is only affected if the patches of moral truth cover questions about the conditions of moral responsibility.12

2.1  Responsibility and Control Most Western theories of responsibility—​compatibilist and incompatibilist—​place a robust “control condition” on responsibility, meaning that agents must have a strong degree of control over their behavior to be justly blamed or praised for it. Theories that lack a control condition entirely are referred to as “strict liability” theories according to which individuals may be blamed or punished whether or not there was intent or negligence on the individual’s part. Most Westerners regard strict liability approaches as fundamentally unfair, even if they may be appropriate under certain circumstances for pragmatic reasons. As Thomas Nagel puts it: “strict liability may have its legal purposes, but seems irrational as a moral position” (Nagel 1979, p. 31). The association of responsibility with control seems so intuitive as to seem self-​ evident. But in fact, it is a culturally local and historically recent development, confined as I mentioned to the WEIRD people of the world. In many honor cultures, for example, the condition is considerably de-​emphasized or in certain cases absent. Consider the common practice of collective punishment—​retaliation against people who played no part in committing the offense that is to be avenged. The anthropologist Ram Nath Sharma writes that in most tribal societies, punishment given for murder is death, but this punishment may not be given to him who has murdered. In his place, some other member of his family, group, or clan may be killed since the group is collectively responsible for the criminal acts of each of its members.” (Sharma 1997, p. 377)

Christopher Boehm (1984) describes the retaliation killings of non-​ offenders in Montenegro as not only morally permissible but morally necessary. He emphasizes that it’s appropriate to kill any member of the offender’s group, no matter their connection to the offense or the offender. Regarding the revenge norms of an Inuit tribe, Balicki (1970) writes that

12  For the same reason, the reader should not assume that the metaskeptic is committed to a broader form of moral relativism. It is entirely possible that metaskepticism about moral responsibility is true but skepticism or relativism about moral facts in general is false. Arguments from disagreement do not stand or fall together. It all depends on the values in question and the nature of the disagreement.

Metaskepticism   253 the objective of the revenge party was not just to kill the original murderer but members of his kindred as well. In a sense the members of the kindred shared responsibility of the murder. (p.184)

And, Gellner (1981) writes of the Bedouin societies that individuals in a tribe are responsible . . . for the conduct of any one of their members sharing in the risk of becoming object of retaliation, or of contributing to blood-​money as an alternative to feud in anyone in the group commits an act of aggression against a member of another group, and similarly being morally bound aggression against any fellow member. (pp. 37–​38)

Individuals in non-​honor cultures have a hard time understanding collective responsibility, let alone justifying it. Many will attempt to connect the judgments to something else besides responsibility. Maybe the retaliators are trying to punish and hurt the real offender by harming his friends and family—​and thus do not truly deem their targets responsible or blameworthy. But this wouldn’t explain why such attacks occur even when the offender is dead and so cannot himself be affected by the death of a family member. So maybe the retaliators are trying to send a message to deter future attacks. But this explanation is implausible as well. Retaliation in honor cultures is never described as “collateral damage” or “regrettable but necessary.” Members of honor cultures don’t turn to consequentialist justifications, and they don’t express the sentiment or belief that the non-​offenders are being unjustly punished. Furthermore, individuals in honor cultures experience all the attitudes associated with blame and responsibility—​gratitude, resentment, moral outrage—​toward their targets.13 Indeed, many of these norms are recognizable in non-​honor societies in certain contexts. Consider beanball norms in baseball for example. If one pitcher hits your superstar, the other pitcher (even in the national league, where the pitcher bats) does not go after the pitcher but rather the superstar of the other team. Rarely do you find the sentiment that it’s unfair to punish an opposing player for the actions of his pitcher. The National Hockey League has a complex set of collective responsibility and collective punishment norms as well. Honor subcultures are present in larger WEIRD societies, and individuals seem capable of adjusting their attitudes according to their context. (See Sommers (2018) for more detailed accounts of these norms and groups). A related principle in Western theories of responsibility is that agents must intend to perform an action or be culpably negligent in some way to be genuinely blameworthy for it. They must have mens rea, a guilty mind, as the principle is described in criminal

13 

Actually, to be clear, members of honor culture may appeal to their non-involvement as a way of avoiding retaliation but these appeals are typically regarded as poor excuses.

254   tamler sommers law.14 This condition is also considerably de-​emphasized in honor cultures. Sharma (1997) writes that “intent is not very important in consideration of punishment in tribal society” (p. 376). He tells the story of a Hupa woman who made a fire to heat water and wash clothes. A child happened to get entangled in the fire and died as a result. Though the death was accidental, the woman was obligated to sacrifice her own son as punishment. We also find cultural differences about the relative importance of intention for responsibility in collectivist cultures as well. In a series of wide-​ranging studies, Hamilton and Sanders (1981, 1992) probed explicitly for responsibility judgments in the United States and Japan. In the studies, subjects from Detroit and two Japanese cities—​Yokahama and Kanazawa—​were given vignettes featuring wrongdoing and asked to assess responsibility and assign punishment. The authors outline three factors that they believed would influence assignments of responsibility in both Japan and America: (1) The deed. How severe was the harm? What was the agent’s mental state, to what degree was the act intentional? (2) The context. Was there was past pattern of wrongful behavior? To what extent was the act was influenced by another person? (3) The role of relationships. What is the relationship of the offender to the victim? The authors hypothesized that in both cultures, offenders would be held more responsible for acts that showed a high degree of intention on the part of the agent. But they also predicted systematic variation in the weight assigned to intention when determining overall responsibility. The results support this prediction. Americans placed far more importance on issues relating to the agent’s mental state or intentions than the Japanese. In fact, the actor’s state of mind when wrongdoing or harm doing occurred was approximately twice as important a determinant of responsibility for Detroiters as it was for Japanese respondents (Hamilton and Sanders 1992, p. 122). Thus, while subjects in both countries take intention and motive into account, the Americans placed far greater weight on this factor.15

14 

I’m assuming here that the legal requirements reflect the moral attitudes about the conditions for culpability. See the discussion in Sommers (2018) concerning the Supreme Court and Victim Impact Statements in which this connection between moral attitudes and legal requirements are made explicit. It’s worth noting too that strict liability laws are almost entirely confined to civil cases, misdemeanors, and corporate crimes. There are very few strict liability laws for criminal felonies. 15  One might think that Japanese are simply less discerning about mental state considerations, about whether or not an act truly reflects the intentions of the offender. But the data do not support this explanation. The one type of case in which Japanese did place great weight on mental state factors is when the offender was in a position of authority over the victim. In those cases, the Japanese showed that they understood the intentions of the offender. Thus, it would be surprising if they failed to do so in other cases. According to the authors, Japanese make less use of information about the actor’s mental state in determining responsibility, but they do not ignore mental state information.

Metaskepticism   255

2.2  Manipulation as Excuse? Maybe Not Most theories of moral responsibility hold that if agents are manipulated to behave wrongly, then they are not blameworthy for the behavior.16 Once again, this condition seems to be de-​emphasized in honor cultures. Perhaps the best illustration is found in the culture of Ancient Greece and the story of Agamemnon. Agamemnon is compelled to murder his daughter in order to set sail for Troy—​which is required by the will of Zeus. For good measure, Zeus sends Ate (or passion) to Agamemnon, removing his wits before he makes the decision. Lloyd-​Jones (1962) writes: Zeus is indeed determined that the fleet must sail; Agamemnon has indeed no choice. But how has Zeus chosen to enforce his will? Not by charging Calchas or some other accredited mouthpiece to inform the king of his decision; but by sending Ate to take away his judgment so that he cannot do otherwise. Does it follow that Agamemnon is not held responsible for his action? Certainly not. (1962, p. 192)

In spite of the constraint and the manipulation, Clytemnestra and the Chorus (in Aeschylus’s version) holds Agamemnon morally responsible for the act. Clytemnestra murders Agamemnon in retaliation for the act. Contemporary scholars have trouble making sense of this seemingly irrational attitude about moral responsibility. The classicist Albin Lesky, for example, writes: If one makes a clear logical distinction, of course, one will say: “A man who acts under necessity is not acting voluntarily.” But to insist upon logical consistency would mean that we should have to reject considerable parts of Aeschylus’ tragedies . . . Is not the campaign against Troy a just punishment inflicted on behalf of the highest god, Zeus, who protects the rights of hospitality? Thus, Agamemnon acts on behalf of the god who wills this punishment. And yet the price for this punishment is a terrible guilt, for which the king has to atone with his death. Here there is no logical consistency. (Lesky 1966; from Nussbaum 1985, p. 235. My italics)

Lesky’s remarks are revealing of Western mindset because there is in fact no logical inconsistency. There would only be logical inconsistency if something like the PAP principle were built into the concept of guilt—​and clearly (for the Greeks) it is not. The judgment of Agamemnon is not illogical; rather, it is counterintuitive from a contemporary Western perspective. Nussbaum offers an interesting way of explaining away the variation. She claims that Agamemnon is being judged not for the crime itself, but for his emotional response to these crimes. He performs them too clinically, without sufficient reluctance.

16  Pereboom’s (2001) first case in his four-​case argument, for example, features overt manipulation. Pereboom is counting on his readers to find it intuitively obvious that the agent is not morally responsible for his immoral behavior.

256   tamler sommers Consequently, the Chorus and Clytemnestra view Agamemnon as lacking virtue and blame him accordingly.17 This is an attractive reconciliation between Western attitudes about manipulation and the Greek view of Agamemnon’s guilt. But this analysis fails to capture the heart of the honor-​based attitudes about responsibility. Recall that Zeus has not only presented Agamemnon with a Sophie’s choice-​style dilemma, he has also manipulated his feelings and attitudes. He sends Ate to Agamemnon, taking away his wits. As Lloyd Jones points out, this occurs in the Iliad as well when Zeus wants Agamemnon to start a quarrel with Achilles. If Agamemnon’s emotions are being manipulated to perform the action, and he regrets the act soon after, it isn’t plausible to say he identifies with them in the sense that Nussbaum means. It seems, therefore, that guilt and responsibility are assigned even when identification is absent. In the Homeric Greek honor culture, one could be responsible and justly punished for acts that were fated, constrained, or brought about by manipulation. Being free of manipulation is not a necessary condition for moral responsibility.

3.  Can the Variation Be Explained Away? Critics of metaskepticism may object that the variation to be described in this section is the result of conceptual confusion or conflation.18 Perhaps when honor cultures engage in collective punishment, they are working with a notion of responsibility that does not involve desert or “fault.” And perhaps East Asians employ a notion of responsibility that is not essentially tied to blameworthiness, focusing instead on restoring relations within the group. In other contexts, honor cultures and collectivist societies either apply a robust control condition for responsibility just as we do in individualist non-​honor societies. Perhaps other cultures are working with a different concept, such as legal responsibility, which would explain the variation in attitudes or intuitions about the conditions for applying the concept. All of this is certainly possible. But it is just as possible that philosophers raised on particular norms and values in a certain place and time have mistaken their culturally and historically local perspectives on moral responsibility for self-evident truths. It would not be the first time that human beings have made a mistake of this kind. Consequently, it is not enough for the critic to point out the mere possibility of a favorable interpretation of the data. This would make the universalist thesis unfalsifiable, for no interpretation can ever be ruled out conclusively. The universalist must show why it’s plausible that the variation is a product of conceptual confusions. In each of the examples here 17  According to Lacey (2001), this view corresponds with the attitudes about moral responsibility in Victorian England. 18  Thanks to Dana Nelkin for offering a clear articulation of this worry.

Metaskepticism   257 (and there is more detail in Sommers 2012), I have tried to provide examples in which the judgments are tied to the same practices that we associate with deserved blame and punishment. Even if the metaskeptic begins with the burden of proof, there must be some way to eventually discharge it. Thus the metaskeptic offers positive evidence for variation in intuitions about something that looks very much like moral responsibility and desert—​fair assignments of blame, praise, punishment, or reward. Universalists who wish to deny this must provide support for their view, support that goes beyond merely pointing out that it is possible that the other cultures are working with a different concept. The critic might concede (at least for the sake of argument) that the variation is over roughly the same concept, but claim that the attitudes and judgments of other cultures are irrational. After all, it’s rare to find universal agreement about any issues at all. People still disagree about our place in the solar system, yet no one argues for metaskepticism about astronomy. The majority of Americans do not believe that we have evolved from common ancestors of other species, so should we be metaskeptics about evolution? Of course not. In our search for the truth about moral responsibility, we don’t have to concern ourselves with primitive conceptions any more than the modern astronomer should worry about Ptolemaic views of the solar system. The consensus we require is the consensus of rational people who are well-​informed about the relevant empirical facts, and who have given careful thought to competing views and arguments. This approach will lead us to the truth about moral responsibility, just like it has in other areas of inquiry such as logic, mathematics, and science.19 I believe that a response along these lines is the most viable strategy for defending universalist theories of moral responsibility. Sturgeon (2006) has written that a large part of moral disagreement would disappear if everyone adopted the position of philosophical naturalism, something it is reasonable to expect all competent moral judges to do. Smith (1994) agrees, arguing that a good chunk of entrenched moral disagreement derives from some parties forming their moral beliefs “in response to the directives of a religious authority rather than as a result of the exercise of their own free thought in concert with their fellows” (pp. 188–​189). And Boyd (1988) writes that “careful examination will reveal . . . that agreement on non-​moral issues would eliminate almost all disagreement about the sorts of moral issues which arise in ordinary moral practice” (p. 123). As Loeb (1998) notes, however, “conjecture about what careful examination will reveal is no substitute for careful examination” (p. 290). It is not enough to simply assume that disagreement is a result of superstition, errors in reasoning, or ignorance about non-​moral facts. One must provide evidence for this claim. And with disagreement about moral responsibility, most of the evidence points the other way. The norms of collective responsibility in the honor cultures of Saga Iceland, for example, are elaborate

19 

This objection is adapted from Sturgeon (2009), who is replying to arguments from disagreement in metaethics more broadly.

258   tamler sommers and complex. They’re not grounded in exotic metaphysical beliefs or implausible empirical assumptions. They are the subject of endless discussion, analysis, and revision from within.20 In one story, the hero Hrafnkels takes revenge on the brother of his torturer, a man who had nothing to do with the original offense. The brother complains to the patrons, who serve as quasi-​judges; the patrons supported Hrafnkels’s decision and gave reasons for their endorsement. The result is a considered judgment that does not seem to arise from superstition or incomplete knowledge about the relevant non-​moral facts. We have no basis for thinking that more rational reflection would have caused the Icelanders to revise their intuitions or judgments about the conditions for responsibility. If there are such grounds, the universalist should point them out and not merely assume they exist. Another compelling reason to doubt that differences can be explained away emerges when we examine the origin of the differences and the psychological mechanisms that underlie them. The legal theorist Nicola Lacey argues that the Western commitment to individual (rather than communal) responsibility is at bottom a response to coordination problems faced by systems of criminal law. Our attitudes can be expected to change according to the environment in which the system operates (Lacey 2001). Research in cultural psychology and evolutionary biology support this view. Recent theories of human cooperation suggest that our moral intuitions and beliefs arise from norms that emerge as responses to features of the social and physical environment. Different norms are better suited to different environments.21 Finally, consider that our responsibility-​related attitudes are embedded within a complex web of social norms and institutions that involve our conceptions of identity and the self. In the Japanese/​American studies already mentioned, Hamilton and Sanders made their predictions in light of the different cultural models of the self. They expected that in a society like the United States, where actors are viewed as isolated individuals, information about individual attributes and deeds would be relatively more important in assessing responsibility than in societies where actors are viewed as contextual. Because collective influence is perceived as both more real and more legitimate in a contextual culture, the Japanese should be more sensitive than Americans to the impact of influence from others on responsibility (Hamilton and Sanders, 1992, p. 87). These theoretical models are in line with recent research in cultural psychology on the different conceptions of the self in individualist (e.g., United States, Western European) and collectivist (e.g., East Asian) societies. The individualist self is regarded as discrete, autonomous, and rational (Triandis 1989; Markus and Kitayama 1991). There is “a sense of self with a sharp boundary that stops at one’s skin and clearly demarcates self from non-​self ” (Spence, 1985, p. 1288, cited in Kim et al., 1994, p. 28). The collective self is conceived of as contextual, embedded, and situated within certain roles and statuses that make up the group. Clear lines separate groups, but there is a “porous boundary between the

20  21 

See Miller (1990, 1993) for descriptions of these norms. See Richerson and Boyd (2005) and Henrich (2007), for example.

Metaskepticism   259 individual self and the group to which it belongs. The self is defined in terms of its in-​ group relationships (Triandis 1995; see also Miller 1984; and Shweder et al. 1985 for excellent analyses). Universalists may still insist that holding individuals responsible only for what is in their control is self-​evidently morally superior to an alternative conception. They may concede that under certain conditions, when there is scarcity of resources, for example, and no dependable form of law enforcement, it may be necessary to de-​emphasize the control condition and punish people for unintended actions, or engage in collective punishment. But these practices are still fundamentally unfair, a necessary evil or at most “second best” morally speaking, and can only be justified in such unfavorable conditions.22 Here again, however, universalists make an unwarranted generalization from their own intuitions. Take the case of Oedipus Rex, for example. Oedipus takes every step he could to avoid fulfilling the prophecy that he would kill his father and marry his mother, but through a series of accidents ends up doing both. Yet once he discovers what happened, he doesn’t make excuses—​he doesn’t let himself off the hook. He blinds himself and wanders away from his kingdom in shame. Oedipus is not unaware that his actions were unintentional, the result of cruel fate. He just doesn’t believe these facts absolve him of responsibility. From a modern perspective, Oedipus’s attitude may appear irrational or even pathological. But we should not overlook the moral benefits of this way of thinking. Indeed, Oedipus’s refusal to absolve himself may seem positively admirable when compared to the modern temptation to grasp at every possible excuse for our conduct. In Milan Kundera’s (1999) great novel The Unbearable Lightness of Being, the protagonist Tomas condemns this self-​exonerating mindset. Tomas is appalled that his Czech countrymen who had collaborated with the Soviets feel no guilt, no sense of responsibility—​that they let themselves off the hook simply because they lacked knowledge of the atrocities the Soviets would commit. Tomas compares them unfavorably to Oedipus. Kundera writes: When Tomas heard Communists shouting in defense of their inner purity, he said to himself, “as a result of your ‘not knowing,’ this country has lost its freedom, lost it for centuries, perhaps, and you shout that you feel no guilt? How can you stand the sight of what you’ve done? How is it that you aren’t horrified? Have you no eyes to see? If you had eyes, you would have to put them out and wander away from Thebes.

Tomas (and Kundera too, almost certainly) believe it’s immoral for them to excuse themselves simply because they were ignorant of what the Soviets would do. Yes, they were unaware of the extent of the atrocities, but according to Tomas, this is no excuse. They should feel deep shame for what they did to their own country. Tomas doesn’t see Oedipus’s perspective as irrational or pathological. He sees it as morally superior to that of his countrymen. 22 

Thanks to Saul Smilansky for articulating this objection.

260   tamler sommers One might claim, as Nussbaum (1993) does, that Tomas’s and Oedipus’s way of regarding responsibility neglects crucial facts about an agent’s particularity, his good character, and fine intentions—​and that rational responsibility judgments must take agent particulars into account. But there is an important difference between neglecting particulars, and being aware of the particulars but deeming them irrelevant or less relevant to the agent’s responsibility. Recall in the discussion of differences between Japanese and American perspectives that the Japanese were equally aware of the empirical facts relevant to the agent’s state of mind, they just did not find the agent’s state of mind to be as important as Americans for determining the offender’s responsibility. The same seems to be true about the Greeks and most honor cultures. They do not ignore questions about motive and intention. They simply disagree with Nussbaum over how critical they are in just sentencing. To assume that an epistemic or control condition is essential for issuing just assignments of responsibility simply begs the question. Or consider practices related to collective responsibility and punishment. As Feinberg notes in his classic paper on the subject, collective responsibility not only expresses the solidarity of the group but strengthens it as well (Feinberg 1968, p. 677). Collective responsibility therefore promotes a moral good to the extent that group solidarity is a moral good. Furthermore, feelings of collective responsibility are bound up with feelings of sympathetic identification. Consider the feelings of responsibility we feel for the acts of our children. We take pride in their accomplishments and feel shame and responsibility for any wrongdoing on their part. A cynic might claim that we are only feeling pride or shame for our own role in raising them, but as most parents recognize, this is simply not the case. As Feinberg notes, the responsibility we feel for our children is genuinely vicarious—​it is an expression and a reinforcement of love and solidarity. In my view, this is a real moral good, not one that is morally “second best” but necessary for practical reasons. Feinberg goes on to discuss the “frankpledge” system of collective responsibility that existed in England between the 11th and 15th centuries. According to Anglo-​Saxon law, groups of approximately 10 people, connected by kin or friendship, were jointly responsible for the criminal acts of each of its members. Feinberg rejects the view that such a system is the “barbarous expedient of a people who had no conception of individualistic justice.” He writes: The frankpledge system was a genuine system of criminal law. There was nothing arbitrary, post-​hoc, or ex post facto about it. It was also a system of compulsory group self-​policing in an age when professional policing was impossible. Moreover, it reinforced a pre-​existing group solidarity the like of which cannot occur in an age of rapid movement like our own. Most importantly, the system worked; it prevented violence and became generally accepted as part of the expected natural order of things. (Feinberg 1968, p. 680. My italics.)

To Feinberg, it is not at all clear that collectivist approaches to responsibility are obviously false and morally inferior. Yet, Feinberg is skeptical of Gomperz’s claim that a

Metaskepticism   261 wholesale rejection of collective responsibility “mistakes a principal characteristic of individualistic ages for an eternal law of human nature” (see the opening quote for this chapter). Gomperz’s claim, Feinberg writes, “misleadingly suggests a kind of historical relativism according to which individualistic and collectivist ages alternate like fashions in ladies skirts” (1968, p. 680). According to Feinberg, something like the frankpledge system could never function in our modern age of anonymous interactions and professional policing. The changes in modern times, Feinberg claims, have come quite inevitably and are here to stay “unless massive destruction forces the human race to start all over again” (1968, p. 680). There are two things worth noting here. First, Feinberg makes an empirical claim here rather than a moral one. According to Feinberg, the social and political structure of our current environment is one for which individualistic perspectives are suitable—​ and nothing in the remotely foreseeable future will change this. He is not claiming that this is morally the best state of affairs; indeed, he goes on to note some moral and practical costs of the individualist perspective. Like Gomperz, he is skeptical of regarding the principle of individual responsibility as some sort of eternal law or truth. Second, Feinberg’s empirical claim is false, if he means it to apply to the entire world rather than just to the United States. There are many cultures populating the world at present with environments that more closely resemble the age of 12th-​century England than 21st-​ century urban America. So it may be that neither set of norms and attitudes is morally “second-​best.” They each have moral advantages and disadvantages, and each is only appropriate in certain environments. When functioning properly, the competing conceptions of moral responsibility may be “tied for first,” morally speaking. Finally, it’s helpful to remember that differences in collectivist and individualistic perspectives are differences of degree rather than kind. There are Americans on the far end of the individualist side, but there is further for Americans to go. We might, for example, become people who no longer take pride in the achievements of our children except to the extent that we played a conscious role in bringing about their achievements ourselves. We might no longer feel responsible and ashamed for the wrongdoing of our children and siblings unless we consciously contributed in some way to its occurrence. Perhaps this move toward an even greater extreme will occur. But would this be something to celebrate? Would this count as further moral progress? Is this view of responsibility more rational? In my view, no. We should be wary of too confidently asserting that the views closer to the collectivist side of the spectrum are less rational and morally inferior to our own.23 None of this is to deny that sorting out errors of reasoning, and learning the relevant non-​moral facts, can have important and salutary effects. Philosophical analysis, moral reflection, and dialogue often leads us to revise considered judgments about responsibility, and there may be substantial changes within a particular culture or group.

23 

Of course, suitably adjusted, these remarks would apply equally well to the overconfident universalist on the collectivist side of the spectrum as well.

262   tamler sommers But given that much of the disagreement in intuitions and attitudes is not, at bottom, a product of errors of reasoning or ignorance of non-​moral facts, it’s implausible to suppose that different cultures would arrive at the same considered judgments about the conditions of moral responsibility. Nor is there any reason to think that they should. If a culture’s norms of responsibility de-​emphasize a control condition, and these norms are well adapted to a particular environment, then it’s unclear how we can accuse individuals who observe these norms of being irrational in those environments. It may be that the conditions for true blameworthiness that arise from intuitions formed in Western environments simply have no place in other kinds of environments.

4.  Implications of Metaskepticism Universalist theories of moral responsibility cannot be justified without appealing to intuitions about crucial principles and cases involving issues relating to moral desert. These theories are therefore grounded in empirical assumptions about the universality or convergence of considered intuitions about moral responsibility. The evidence suggests that no such convergence over considered judgments would occur, even under conditions of ideal rationality. And if that is the case, there does not seem to be a principled way of establishing conditions of moral responsibility that would apply across cultures. What are the social and political implications of this position? When I first embarked on this project, I had grandiose notions about how to apply metaskeptical principles to problems in foreign policy and criminal justice. I soon realized that these issues were manifold and complex and involved entirely different literatures than the ones I had been examining. So my conclusions are far more tentative now, raising questions and hopefully serving as a point of departure. Rhetoric that accompanies military intervention in the affairs of other countries often involves appeals to principles of autonomy, responsibility, and justice. Some politicians and foreign policy experts justified the invasion of Iraq, for example, as fulfilling a moral imperative to spread democratic values and ideals of individual responsibility. Colonialist Western governments have offered similar justifications for hundreds of years. The truth of metaskepticism would cast doubt or at least complicate these kinds of justifications. On a more practical level, the variation in perspectives on responsibility and justice has important implications for the ways in which governments engage in their campaigns. Some members of the US military are beginning to recognize this (as they did during war with Japan, when they commissioned Ruth Benedict to conduct a study of Japanese values). Former Army Major William McCallister, a consultant for the Marines, has argued that military policies in Iraq did not reflect an adequate understanding of cultural norms. According to McCallister, the concepts of shame and honor are as important to the Iraqis as land or water, as they form the moral currency of the country (McCallister 2005). He attributes our unpopularity with Iraqi tribes in large

Metaskepticism   263 part to our failure to understand this feature of their society. Similar observations about the role of honor have been made about all the countries in that region. Thanks in large part to McCallister’s efforts, the US Army in 2006 released a 125-​page handbook entitled Through the Lens of Cultural Awareness: A Primer for US Armed Forces Deploying to Arab and Middle Eastern Countries. In order to shape policies properly, we need a comprehensive analysis of the role that honor and collective responsibility plays in these countries and the role that individual responsibility may not play. Once alerted to differences in ethical perspectives, we may no longer expect people in honor cultures to immediately embrace American notions of justice. The metaskeptical thesis also raises deep philosophical questions about the justification of punishment. Criminal justice in America is in part grounded in abstract retributive principles that require impartial justice based on individual culpability. The US system severs the connection between the victim and offender when doling out punishment. Members of honor cultures find the commitment to total impartiality and “blind justice” to be extremely unsatisfying (Daly and Wilson 1988; Miller 2006). Greater sensitivity to variation may dictate more of a balanced view about the relationship between justice and the victim. Indeed, one of the most promising movements in the philosophy of punishment is restorative justice, which incorporates victims into the sentencing process.24 Let me conclude by saying that metaskepticism is not a pessimistic position. Recognizing the diversity of legitimate responsibility practices makes us more open to learning about the attitudes of other communities and to adjusting our own views in light of what we learn. After all, even for individuals, the balance of intuitions about moral responsibility is a delicate one. Changes in environment or life circumstances (having children, getting married, being the victim of a crime, losing someone you love, or simply getting older) can alter the balance in either direction. It might be that responsibility-​related attitudes continually evolve and shift, depending on the context of our relationships and circumstances. Maybe it is philosophers, with long-​standing commitments to particular theories, who are the exceptions.

References Audi, Robert (2005). Practical Reasoning and Ethical Decision. Routledge. Audi, Robert (2008). Intuition, Inference, and Rational Disagreement in Ethics. Ethical Theory and Moral Practice 11 (5):475–​492. Balikci, A. 1970. The Netsilik Eskimo. Boston: The Natural History Press. Boehm, C. 1984. Blood Revenge. Philadelphia: University of Pennsylvania Press. Boyd, R. 1997. “How to Be a Moral Realist.” Moral Discourse and Practice: Some Philosophical Approaches: 105.

24 

For research on the moral benefits of restorative justice, see, e.g., Christie (1977), Braithwaite (1989), Koss (2013), Sommers (2016), and Sommers (2018).

264   tamler sommers Braithwaite, J. 1989. Crime, Shame and Reintegration. Cambridge: Cambridge University Press. Brink, David O. (1984). Moral realism and the sceptical arguments from disagreement and queerness. Australasian Journal of Philosophy 62 (2):111–​125. Christie, Nils. 1977. “Conflicts as Property*.” The British Journal of Criminology 17(1): 1–​15. Daly, M., and Wilson, M. 1988. Homicide. Aldine de Gruyter. Doris, J.M., and Plakias, A. 2007. “How to Argue About Disagreement: Evaluative Diversity and Moral Realism.” In W. Sinnot-​Armstrong (ed.), The Biology and Psychology of Morality. New York: Oxford University Press. Doris, J.M., and Stich, S.P. 2005. “As a Matter of Fact: Empirical Perspectives on Ethics.” In F. Jackson and M. Smith (eds.), The Oxford Handbook of Contemporary Philosophy. Oxford: Oxford University Press. Feinberg, Joel (1968). Collective responsibility. Journal of Philosophy 65 (21):674–​688. Fischer, J., and Ravizza, M. 1998. Responsibility and Control. Cambridge: Cambridge University Press. Frankfurt, H. 1969. “Alternate Possibilities and Moral Responsibility.” Journal of Philosophy 66: 829–​839. Gellner, Ernest (1981). Nationalism. Theory and Society 10 (6):753–​776. Gomperz, H. 1939. “Individual, Collective, and Social Responsibility.” Ethics 49(3): 329–​342. Hamilton, V.L., and Sanders, J. 1992. Everyday Justice. New Haven, CT: Yale University Press. Henrich, Joseph et al. 2010. “The Weirdest People in the World?” Behavioral and Brain Sciences 33(2–​3): 61–​83, doi:10.1017/​s0140525x0999152x. Henrich, N., and Henrich, J.P. Why Humans Cooperate: A Cultural and Evolutionary Explanation. Oxford: Oxford University Press, 2007. Kim, U.E. et al., 1994. Individualism and Collectivism: Theory, Method, and Applications. Thousand Oaks, CA: SAGE Publications. Koss, Mary P. 2013. “The RESTORE Program of Restorative Justice for Sex Crimes.” Journal of Interpersonal Violence 29(9): 1623–​1660, doi:10.1177/​0886260513511537. Kundera, M, 1999. The Unbearable Lightness of Being. New York: Harper Collins. Lacey, N. 2001. “In Search of the Responsible Subject: History, Philosophy and Social Sciences in Criminal Law Theory.” The Modern Law Review 64(3): 350–​371. Lesky, A. 1966. Decision and Responsibility in the Tragedy of Aeschylus. Journal of Hellenic Studies 86: 78–​85. Lloyd-​Jones, H. 1962. “The Guilt of Agamemnon.” The Classical Quarterly 12(2): 187–​199. Loeb, D. 1998. Moral Realism and the Argument from Disagreement. Philosophical Studies 90 (3):281–​303. Luther, M. 2007. The Bondage of the Will. Ambassador-​Emerald International; First edition Mackie, J.L. 1977. Ethics: Inventing Right and Wrong. London: Penguin. Markus, H., and Kitayama, S. 1991. “Culture and the Self: Implications for Cognition, Emotion, and Motivation.” Psychology Review 98: 224–​253 McCallister, W. 2005. “The Iraq Insurgency: Anatomy of a Tribal Rebellion,” First Monday, 10:3. McKenna, M., and Russell, P. 2008. Free Will and Reactive Attitudes. Burlington, VT: Ashgate. 1–18 Miller, J.G. 1984. Culture and the Development of Everyday Social Explanation. Journal of Personality and Social Psychology 46(5):961–​978. Miller, W.I. 1993. Humiliation: And Other Essays on Honor, Social Discomfort, and Violence. Ithaca, NY: Cornell University Press.

Metaskepticism   265 Nagel, T. 1993. “Moral Luck.” In Daniel Statman (ed.), Moral Luck. State University of New York Press. pp. 141–​166. Nussbaum, M. 1985. “Aeschylus and Practical Conflict.” Ethics 95(2): 233–​267. Nussbaum, M. 1993. “Equity and Mercy.” Philosophy & Public Affairs 22(2): 83–​125. Pereboom, D. 2001. Living Without Free Will. Cambridge: Cambridge University Press. Pereboom, Derk (2014). Free Will, Agency, and Meaning in Life. Oxford University Press. Ravizza, Mark (1994). Semi-​compatibilism and the transfer of non-​responsibility. Philosophical Studies 75 (1–​2):61–​93. Richerson, P., and Boyd, R. 2005. Not by Genes Alone. Chicago: University of Chicago Press. Sharma, R.K., and Sharma, R.M. 1997. Anthropology. New Delhi: Atlantic Books. Shoemaker, D. 2015 Responsibility from the Margins. Oxford: Oxford University Press. Shweder, R.A., and Miller, J.G. 1985. The Social Construction of the Person: How Is It Possible. The Social Construction of the Person: Springer: New York. 41–​69. Smith, M. 1994. The Moral Problem. Blackwell. Sommers, T. 2012. Relative Justice. Princeton, NJ: Princeton. Sommers, T. 2016. “The Three Rs: Retribution, Revenge, Reparation.” Philosophia. 44(2): 327–​342. Sommers, T. 2018. Why Honor Matters. New York: Basic Books. Sturgeon, N. 2006. “Ethical Naturalism.” In The Oxford Handbook of Ethical Theory. Oxford: Oxford University Press. Sturgeon, Nicholas L. 2009. Doubts about the Supervenience of the Evaluative. In Russ Shafer-​ Landau (ed.), Oxford Studies in Metaethics. pp. 53–​92. Triandis, H. 1989. “Self and Social Behavior in Differing Cultural Contexts.” Psychology Review 96: 269–​289. Van Inwagen, P. 1983. An Essay on Free Will. Oxford: Clarendon. Vargas, M. 2013. Building Better Beings. Oxford: Oxford University Press. Wallace, R. Jay (1994). Responsibility and the Moral Sentiments. Harvard University Press. Williams, B. 1993. Shame and Necessity. Berkeley: University of California Press. Wolf, S. 1987. “Sanity and the Metaphysics of Responsibility.” In F. Schoeman (ed.), Responsibility, Character and the Emotions: New Essays in Moral Psychology. Cambridge: Cambridge University Press

PA RT V I

BLAME

CHAPTER 13

Bl am e and H ol di ng Resp onsi bl e Angela M. Smith

1.  Introduction The last two decades have witnessed a remarkable blossoming of philosophical interest in the concept of blame, and in the associated idea of holding persons responsible. That interest can be traced to a number of important developments in contemporary theorizing about the nature of moral responsibility. The first development, which has been well-​documented, is the increasing influence (some today might say hegemony) of P. F. Strawson’s seminal 1962 essay “Freedom and Resentment.” Though this article did not receive a great deal of attention at the time it was written, work in the late eighties and early nineties by Gary Watson (1987), T. M. Scanlon (1988), Susan Wolf (1990), and R. Jay Wallace (1994) brought Strawson’s distinctive approach to theorizing about moral responsibility to the forefront of philosophical reflection on this topic. That distinctive approach emphasizes, among other things, that our theorizing about moral responsibility should begin with a clear understanding of the types of moral reaction that are involved when we hold one another responsible for our attitudes or conduct. We must understand the nature of these moral reactions, according to this approach, in order to determine what conditions an agent must meet in order to be an eligible target of such reactions. In that sense, our understanding of what is required in order to be responsible depends crucially on our understanding of what is involved in holding one another responsible. This first development led philosophers interested in questions of moral responsibility to focus more attention on the nature of our moral responsibility practices, and in particular on the nature of moral blame. While many have followed Strawson in identifying blame with certain negative reactive attitudes such as resentment, indignation, and guilt, others have interpreted blame as a distinctive form of moral appraisal, as a stance involving a combination of cognitive and motivational elements, or as a set

270   Angela M. Smith of reactions that play a particular functional role in moral life. This increased focus on moral psychological questions about how best to interpret the nature of moral blame constitutes the second important development in contemporary theorizing about moral responsibility. A third important development has been an increasing interest in what has come to be known as “the ethics of blame.” Strawson’s emphasis on the interpersonal nature of our moral responsibility practices highlights the fact that “holding responsible” is a three-​ term relationship, in which one party holds another party responsible by responding to her with certain distinctive moral responses (Watson 1996, p. 235). In addition to asking questions about an agent’s basic eligibility for these distinctive moral responses, therefore, we can also ask questions about the conditions under which a would-​be blamer would be justified in responding to others in these distinctive ways. Certain facts about the would-​be blamer and/​or about her relation to the blameworthy agent may be relevant to determining both whether and, if so, when and how she is justified in blaming others. My aim in what follows is to give a brief overview of some of the recent literature on blame and holding responsible, highlighting some of the most important ongoing debates relating to this topic.1

2.  What Is Blame? As already mentioned, much of the recent literature on questions of moral responsibility has focused on trying to get clear about the nature of the moral responses embodied in our actual moral responsibility practices. This focus reflects an assumption that is shared by many who write on this topic, which we might call assumption MR: MR: S is morally responsible for X if and only if S is an eligible candidate for certain distinctive moral responses on the basis of X.2 This assumption is meant to capture the widely held view that agents who are morally responsible for their activities are potential targets of an especially deep or

1  For three excellent general overviews of this literature, see Coates and Tognazzini (2012, 2013), and Tognazzini and Coates (2014). 2  Versions of some such biconditional can be found in the work of Brink and Nelkin (2013, p. 287), Fischer and Ravizza (1993, pp. 18–​19), Macnamara (2015, p. 213), McKenna (2012, pp. 34–​55), Nelkin (2016, p. 183), Rosen (2015, p. 67), Scanlon (1998, p. 48), Shoemaker (2015, p. 17), and Wallace (1994, p. 91), among many others. There is debate, however, over whether it makes sense to say that agents can be morally responsible for things that are themselves morally neutral or indifferent (and that therefore do not call for any particular moral response in the context). Vargas (2013, p. 213 and pp. 307–​309) and Rosen (2015, p. 66) restrict moral responsibility to things that are either praiseworthy or blameworthy. McKenna (2012, pp. 16–​17) restricts moral responsibility to things that are “morally significant” (even if not praiseworthy or blameworthy). Smith (2015, pp. 106–​107) argues that one can be morally responsible for things that are morally neutral or insignificant.

Blame and Holding Responsible    271 meaningful kind of moral response. They are not just causally responsible for good or bad happenings, as a Category 5 hurricane might be “responsible” or “to blame” for the massive flooding. Rather, the actions of morally responsible agents are thought to reflect specially upon them, and to open them up to responses with a distinctive kind of “force” or “depth” (Wolf 1990; Wallace 1994; Hieronymi 2004). The question is, what are these moral responses, and what gives them this special force or depth?3 One natural answer to these questions is that blame must be some sort of punitive activity we direct toward those whom we regard as morally blameworthy. If we take our cue from the criminal legal system, a judgment of moral blameworthiness might be seen as akin to a finding of criminal guilt, and moral blame might be seen as akin to criminal punishment. To blame someone, on this view, is simply to engage in negative overt behavior (such as scolding, rebuking, telling off, or ostracizing) in response to someone’s objectionable conduct. According to many Utilitarian philosophers writing in the first half of the twentieth century, moral blame should be understood as a form of negative treatment whose purpose is to bring about socially beneficial outcomes; and blame is justified, on this view, so long as it is possible to influence a person’s future conduct or character through such sanctioning activities (Schlick 1939; Smart 1961). While this overt sanction account of blame was once widely accepted in the philosophical literature, I think it is fair to say that it has now fallen decisively out of favor. Even philosophers who see close connections between blame and certain forms of punishment or unpleasant treatment generally insist that blame itself is distinct from any of these forms of punitive activity (Wallace 1994, pp. 55–​56; Rosen 2015, p. 83). One obvious problem with this account is that it does not appear to allow for the phenomenon of private or unexpressed blame. But it seems clear that we can blame people privately, without ever expressing that blame through overt sanctioning activities. Nor does it appear to allow for blame of people who are outside the reach of our sanctioning activities (either because they are now dead or distant from us in time or space). But it also seems quite possible for us to blame people, such as brutal dictators in distant countries, whom we are not in a position to personally sanction in any way. Both of these objections point toward a more general objection, which is that even when blame is overt, its force seems to reside not in the outward conduct itself, but rather in the negative attitude that is expressed by this outward conduct. Indeed, we generally do not think that the kind of overt sanctioning activities directed at small children or pets count as instances of “moral blame” precisely because these activities do not express the relevant sort of moral attitudes or judgments. But if that is correct, then it seems that the sanctioning activities many people identify with blame are really just vehicles for the expression of moral blame, and we still need an account of what blame, itself, is. More recent accounts of the nature of moral blame tend toward four general types of approach: Appraisal Views, Reactive Attitude Views, Hybrid Views, and Functional

3 

The next several paragraphs draw heavily from my discussion in Smith (2013, pp. 29–​32).

272   Angela M. Smith Views.4 According to the Appraisal View, to blame a person for something is essentially to take that thing to reveal something negative about that person’s character. Defenders of the appraisal view often appeal to the idea of a moral “ledger” or “balance sheet,” and suggest that when we blame a person we are, in effect, making subtractions to our assessment of their overall “moral record” (Feinberg 1970, pp. 125–​127; Glover 1970, p. 64; Haji 1998, p. 8; Zimmermann 1988, p. 38). Wrongdoers are judged to have a black mark, or a demerit, in their moral ledger, which in turn lowers our overall assessment of their moral worth. Gary Watson, likewise, has suggested that one significant type of moral blame, which he calls “aretaic blame,” involves an appraisal of an agent’s “faults—​. . . or vices—​as manifested in thought and action” (Watson 1996, p. 231). To evaluate an agent as cruel, insensitive, or boorish, Watson suggests, is to judge that she has failed to live up to “certain standards of human excellence in human affairs” (p. 230). Such aretaic appraisals, he argues, should be distinguished from responses of “accountability blame,” which involve negative responses that go beyond such “fault-​finding appraisal.” One advantage of the Appraisal View is that it can explain how blame goes beyond a mere judgment of blameworthiness, but in a way that, unlike the sanction view, allows for the phenomenon of private or unexpressed blame. In order to blame a person, I must not only judge that she is blameworthy for something; I must also judge that she is diminished or disfigured or tarnished in some way in virtue of her misconduct. But I can make such an appraisal without ever expressing that judgment to the person in any way, through word or action. Yet critics of the Appraisal View have argued that this simply does not capture what most of us have in mind when we blame someone for her misconduct. In particular, it has been argued that this view cannot capture the distinctive force of blame, or explain why it is anything more than “a pointless assignment of moral grades” (Scanlon 2008, p. 127). The notion that in blaming we may be dispassionately evaluating a person’s moral record simply does not do justice to the emotional significance of blame and to the important role this attitude plays in structuring our moral relations with one another. Reflection on this significance and this role brings us to the second, and probably most influential, contemporary account of the nature of blame, the Reactive Attitude View. In his landmark essay, “Freedom and Resentment,” P. F. Strawson drew attention to a set of attitudes that he argued is intimately bound up with our practices of holding one another responsible (Strawson 1962/​1993). These “reactive attitudes,” as he called them, are essentially emotional reactions to the good-​or ill-​will that people manifest toward us (or others) in their behavior. Strawson, and Strawson-​inspired theorists such as R. Jay Wallace, put particular emphasis on the negative attitudes of resentment, indignation, and guilt as the characteristic emotional responses to perceived manifestations of ill-​ will (Wallace 1994, pp. 29–​30). According to these philosophers, these reactive attitudes 4  Tognazzini and Coates (2014) propose a similar taxonomy using the labels “cognitive theories, emotional theories, conative theories, and functional theories.” Given that most of the theories in question acknowledge that blame typically involves a combination of cognitive, emotional, and conative elements, however, I find the labels proposed above somewhat less misleading.

Blame and Holding Responsible    273 are the key to understanding the nature and significance of blame (and self-​blame). Blame, on this view, is not simply a negative assessment of someone’s character, nor is it an explicit activity we engage in in order to sanction someone for bad behavior. Blame is a way of responding emotionally to the perceived disregard or disrespect manifested in someone’s behavior toward oneself or others. These reactions, according to Strawson, “rest on, and reflect, an expectation of, and demand for, the manifestation of a certain degree of goodwill or regard on the part of other human beings toward ourselves; or at least on the expectation of, and demand for, an absence of the manifestation of active ill-​will or indifferent disregard” (1962/​1993, p. 57). And these reactive attitudes, in turn, “tend to inhibit or at least limit our goodwill toward the object of these attitudes, tend to promote an at least partial and temporary withdrawal of goodwill” (p. 63). When we blame someone, then, we are emotionally exercised by what they have done, and this emotional disturbance carries with it a certain amount of hostility toward the offender. As Strawson puts it, the reactive attitudes entail “the modification . . . of the general demand that another should, if possible, be spared suffering” (p. 63). The Reactive Attitude View has seemed to many philosophers to capture quite nicely the distinctive force of blame, and to explain how blame differs from a simple judgment of blameworthiness. The person who says, “I know X is blameworthy, but I just can’t bring myself to blame him” is indicating that he does not feel the usual attitudes of resentment, indignation, or hostility toward a person who has manifested ill-​will. And the person who forgives can be understood as disavowing the sort of “partial withdrawal of goodwill” that would normally be justified in response to a blameworthy agent. This view clearly allows for the possibility of private or unexpressed blame, but it also gives a plausible account of what the reactive attitudes express when they are communicated to others: They express a kind of hostility toward the agent for violating the “basic moral demand” for goodwill or reasonable regard in her interactions with others. This, in turn, explains why blame plays such an important role in our interpersonal relationships, for blame and the other reactive attitudes embody at a deep level the basic expectations we hold one another to as members of a shared moral community. Finally, it has been argued that this account of blame can make sense of the widely held view that robust moral responsibility requires that an agent have a certain kind of control over that for which she is held responsible. Because it is unpleasant to be the target of the reactive attitudes, and because these attitudes involve dispositions to adverse treatment, it is only fair to direct such attitudes toward agents who have had a “fair opportunity to avoid” being subject to these negative responses (Watson 1996, p. 239; cf. Nelkin 2015).5

5 

Recent defenders of Reactive Attitude Accounts of blame include Cogley (2013); Graham (2014); Menges (2017); Nelkin (2016); Pickard (2013); Rosen (2015); Tognazzini (2013); Wallace (2011); and Wolf (2011). Watson (1996) identifies the negative reactive attitudes with “accountability blame,” which he distinguishes from the kind of negative appraisal involved in “aretaic blame.” These distinct forms of moral blame, he argues, presuppose different agency conditions, and thus point toward different conceptions of moral responsibility (which have come to be known as “responsibility as attributability” and “responsibility as accountability”). For skepticism about the attributability/​accountability

274   Angela M. Smith While many have endorsed this basic Strawsonian account of the nature of moral blame, others have found it less than fully satisfactory. Both George Sher (2006) and T. M. Scanlon (2008) have objected that blame does not necessarily involve reactive emotions such as resentment and indignation. We can genuinely blame a loved one, for example, or a distant wrongdoer, without having any feelings of resentment, indignation, or anger. They have also both questioned the assumption that the reactive attitudes necessarily embody a kind of hostility, a withdrawal of goodwill, or a suspension of our general desire that offenders be spared suffering.6 Both have argued that, rather than identifying blame with a narrow set of emotional reactions, we would do better to offer a more flexible account that can explain why blame can take so many different forms—​in some cases hot and hostile, in other cases cool and distancing—​in different contexts. This brings us to what I have called the Hybrid View of moral blame. Sher’s suggestion is that we should understand blame as a set of dispositions to have certain attitudinal and behavioral reactions, and these dispositions should be understood as traceable to a single desire-​belief pair that includes (1) a belief that the person in question has acted badly or has a bad character; and (2) a corresponding desire that the person not have acted badly or not have a bad character (p. 112). This results in what Sher calls a “two-​tiered account of blame” (p. 138), with the core desire-​belief pair forming the first tier and some collection of blame-​related behavioral and attitudinal dispositions forming the second tier. This has the advantage of explaining why anger, hostile behavior, reproach, disappointment, and withdrawal all count as blame-​related reactions, while denying that any of these activities or attitudes is essential to blame. To blame someone, then, is to have this desire-​belief pair and to be disposed to react in these distinctive ways in appropriate circumstances. Similarly, Scanlon (2008) offers an account of blame that identifies it with a core belief plus an associated motivational element that can take a variety of different forms depending on the particular context. As he puts it, To claim that a person is blameworthy for an action is to claim that the action shows something about the agent’s attitudes toward others that impairs the relations that others can have with him or her. To blame a person is to judge him or her to be blameworthy and to take your relationship with him or her to be modified in a way that this judgment of impaired relations holds to be appropriate. (pp. 128–​129)

For example, if a friend abandons me in a time of need, I may judge that she has attitudes (of disloyalty, lack of care, etc.) that impair my relationship with her. If I make no other modifications to my expectations, intentions, and attitudes toward my friend, then it

distinction, see Smith (2015). For a defense of this distinction, and an argument that there is a third type of moral blame associated with a third type of responsibility, see Shoemaker (2015). 6  Pereboom (2014, ch. 6 and ch. 8) also rejects the retributive elements of the negative reactive attitudes, but argues that blame may involve demands for justification and nonreactive emotions such as disappointment, sorrow, and regret.

Blame and Holding Responsible    275 may be said that I judge her blameworthy, but I do not blame her. Alternatively, I may modify my expectations, intentions, and attitudes toward her in any number of ways. I may no longer trust her, seek her company, or take pleasure in her successes; I may feel resentful or angry or disappointed; I may complain to her or seek a justification or explanation. In Scanlon’s view, all of these possible modifications to my own attitudes, intentions, and expectations count as ways of “blaming” another, and what unites them is that they are all appropriate responses to a judgment of impaired relations. Like Sher, then, Scanlon thinks it is a mistake to identify blame with any single attitude or type of behavior; rather, blame can take any number of different forms depending upon the nature of the relationship that is impaired, the nature of the impairment itself, and the specific relation between the blamer and the one blamed. These Hybrid Views thus seek to accommodate the variety of responses we are prepared to call responses of moral blame while still providing a unifying rationale that can explain what all such responses have in common. These two hybrid accounts of the nature of blame have generated a great deal of discussion, elaboration, and criticism over the last decade.7 The most common criticism that has been raised to both accounts is that they are, as Victoria McGeer (2013) colorfully puts it, far too “sanitized” to capture the blaming reactions embodied in our actual moral practices.8 While she applauds the effort to provide an account of blame that can be seen as a normatively acceptable (and perhaps even valuable) response to moral wrongdoing, she argues that an adequate account of blame also needs to be psychologically realistic. That is, it must “direct us toward a psychologically plausible phenomenon, however unsavory, that answers to our ordinary conception of blame: it satisfies the ordinary connotations of the term, including connotations of anger, resentment, and desire for payback” (p. 163). Others have objected to these accounts on the grounds that they do not adequately capture the essentially relational and communicative nature of blame, in particular the fact that blame seems both to presuppose and to issue a kind of moral demand to its target. Blame, according to these critics, presupposes that another has violated a moral demand, and it also (at least implicitly or incipiently) issues a demand for an appropriate moral response to that violation (Darwall 2006; McKenna 2012; Nelkin 2016; Walker 2006; Watson 1987, 1996). The notion that the core of blame is a “desire that the agent not have acted badly or not have a bad character,” as Sher claims, or that it consists in a set of adjustments to one’s attitudes, intentions, and expectations toward another in light of her relationship-​impairing attitudes, as Scanlon claims, has seemed to many insufficiently confrontational to capture the basic phenomenology of blame as it is ordinarily experienced. These critiques, in turn, have led a number of philosophers to propose that we should abandon the search for a set of necessary and sufficient conditions for moral blame, and 7  Several of the contributions in the seminal Coates and Tognazzini (2013) volume on blame raise difficulties for one or both of these accounts, including those by Smith, McGeer, Franklin, Pereboom, and Bennett. 8  Similar criticisms are raised by Wallace (2011) and Wolf (2011).

276   Angela M. Smith to look instead at the basic “point” of blame—​that is to say, the functional role it plays in our moral lives (Fricker 2016; McGeer 2013; McKenna 2012; Vargas 2013). According to Functional Views, the search for a strict analysis of blame is bound to be frustrated, because our practices of blame are too thoroughly disunified; yet we can understand these diverse practices as explanatorily related to one another in virtue of their sharing a distinctive underlying aim, which different defenders of Functional Views understand in different ways. McGeer (2013), for example, seeks to provide a broadly naturalistic, evolutionary, and functionalist account of blame that interprets it as having both a backward-​ looking element of appraisal as well as a forward-​looking, action-​priming element. She hypothesizes that a species of anger is characteristic of blame responses in part because such emotions are effective norm-​enforcement devices. Since large-​scale human cooperation could emerge only with the development of social norms, it is plausible to suppose that there would need to be certain cognitive/​motivational mechanisms in place to police such norms. Thus we can understand blame as a state that is “apt for being caused by perceived wrongdoing and apt for producing certain behavioral effects” (p. 169). These behavioral effects include aggressive (punitive) responses aimed at changing or inhibiting behavior that is seen as transgressing personally or socially valued norms (p. 172). The ultimate function of blame, she goes on to suggest, is to spark a dynamic trajectory in which, ideally, the wrongdoer comes to feel guilt and remorse, which in turn leads to the blamer’s mollification, understanding, and perhaps even forgiveness. In that way, this process has the potential to enhance the moral understanding of both the wrongdoer and the blamer, as well as others in the moral community. In a somewhat similar vein, Manuel Vargas (2013) argues that our practices of “holding responsible” in general, and blaming in particular, serve the function of fostering a valuable form of agency. On what he calls the “Agency Cultivation Model,” when we hold one another responsible, “we participate in a system of practices, attitudes, and judgments that support a special kind of self-​governance, one whereby we recognize and suitably respond to moral considerations” (2013, p. 2). The variety of “blaming reactions” (p. 119) we engage in, therefore, are unified by the fact that they tend to foster and sustain this “valuable form of agency,” a form of agency that “we ordinarily have reason to care about” (p. 2). Michael McKenna (2012) and Miranda Fricker (2016), like McGeer, stress the distinctively communicative nature of blame and its functional role in our moral lives. McKenna defends what he calls a “Conversational Theory of Moral Responsibility” according to which blame can be understood as a stage in a broader “conversational exchange” that often takes place within our moral responsibility practices. Blame directed at an agent invites a response, or a “moral account,” which in turn often gives rise to further moral responses (such as exoneration, forgiveness, or punishment). Miranda Fricker (2016), likewise, offers what she calls a “paradigm based explanation” of blame, which involves identifying an explanatorily basic form of blame that can be seen as serving a particular point and purpose in our moral lives. She calls this paradigm form of blame “Communicative Blame,” and argues that the basic function of such blame is

Blame and Holding Responsible    277 to bring about an increased alignment of moral understandings between the blamer and the wrongdoer. She then argues that other forms of blame (e.g., self-​blame, private blame, third-​party blame, etc.) can all be understood as derivative forms of blame that still make sense in light of blame’s fundamental communicative purpose. Finally, a number of philosophers have suggested that the underlying function of blame might be to register protest, and that this is what fundamentally distinguishes “non-​blaming” from “blaming” responses to moral offense or wrongdoing (Hieronymi 2001; Talbert 2012; Smith 2013). Since it is possible to register such protest through a wide variety of responses—​from a private and unemotional decision to no longer trust another, to a very public and emotional expression of rebuke—​such views allow for a wider range of responses to count as instances of moral blame than some other functional accounts. What all such accounts appear to have in common, however, is the idea that blame has an implicitly “communicative” function in moral life: Though any particular instance of blame need not be expressed, blame is the sort of state that is fitted to communicate the blamer’s fundamental expectations about how she (and others) are to be treated. While functional accounts seem to be gaining some traction in the field, they too have been subject to a variety of questions and objections. One question concerns how we should understand the relation between the (often) naturalistic, evolutionary story about the emergence of our blaming attitudes, on the one hand, and certain normative concerns one might have about some lingering punitive or retributive elements of these attitudes, on the other. If our blaming attitudes are, at base, simply evolved mechanisms of social control, are these attitudes we have reason to continue to endorse? Others have worried that these views, insofar as they stress the positive effects of our blaming practices, are susceptible to many of the same objections that were raised to earlier Utilitarian social influence accounts.9 One might also question whether it is possible to make sense of the notion of “protest” without making a prior appeal to the notion of moral blame (in which case, the protest account might look fundamentally circular). Finally, some have argued that it is a mistake to claim that attitudes of blame are, even incipiently, communicative, since this might seem to imply that it is inappropriate to direct such attitudes to agents who are, for whatever reason, unable to give appropriate “uptake” to such attitudes through responses of guilt or remorse (Scanlon 2008, fn. 53, pp. 233–​34). The question of whether and in what way blame is communicative, and how the answer to that question might be relevant to our understanding of the basic conditions of moral responsibility, is currently one of the most hotly debated questions in the field today (Darwall 2006; Macnamara 2015; Nelkin 2015; Smith 2019; Talbert 2012; Watson 2011).

9  Vargas (2013, pp. 167–​198) gives careful attention to these objections, and explains why he does not think they apply to his own “Agency Cultivation” account. McGeer (2015) suggests that Vargas’s view cannot entirely escape these worries.

278   Angela M. Smith

3.  The Ethics of Blame In parallel with the recent explosion of literature on the nature of blame, philosophers have also begun to address a variety of questions about whether, when, and in what way it is appropriate for particular individuals to blame particular others for particular transgressions. Of course, one obvious condition on the appropriateness of blaming another has to do with facts about the person being blamed: Is she, in fact, blameworthy? That is, does she meet those conditions, whatever they are, that render her morally responsible for her conduct, and has her conduct, in fact, violated, without sufficient excuse or justification, certain moral norms or expectations that legitimately apply to her? There is extensive debate in the literature over what the basic conditions of moral responsibility are, over how to understand the basic moral norms and expectations that apply to us, and over what counts as a sufficient excuse or justification for violating such norms. These important debates are beyond the scope of this discussion. But once one grants that an agent is clearly blameworthy for something she has done (however that is defined), we can still ask questions about whether we are all equally entitled to blame her for it, or to blame her in the same way. After all, different people stand in different relations to the blameworthy agent, to her victim, and/​or to the transgression in question, and it seems that these facts may be relevant to determining how it is appropriate for particular individuals to respond to the blameworthy party. Consider, for example, the following simple example. Abby, Bruce, and Cathy are friends who have committed to running together every Saturday morning at 8:00 a.m. Abby always arrives at the track on time, properly stretched and ready to go. Bruce is usually, but not always, on time, but Cathy regularly arrives 10–​15 minutes late. Neither Bruce nor Cathy ever apologizes to Abby on the occasions in which they do arrive late. One Saturday, however, Abby and Cathy arrive on time, and Bruce is not there. When he does arrive, five minutes late, Cathy heatedly rebukes him and accuses him of not respecting her and Abby’s time. In this situation, many would argue that Cathy lacks standing to blame Bruce for his tardiness.10 The fact that she herself has frequently engaged in similar wrongdoing, without acknowledgment, excuse, or apology, seems to undermine her authority to “hold” Bruce responsible for his blameworthy conduct. Abby, however, seems to have standing to blame both Bruce and Cathy for their tardiness. Finally, Bruce might think it appropriate to blame himself for inconveniencing Abby, but not for inconveniencing Cathy, given how often she has inconvenienced him in similar ways in the past. (For good measure, we might also note that Douglas, a complete stranger who just happens to be running around the track that day, seems to lack standing to blame either Bruce or Cathy in this context.) 10 

Bell (2013) is one of the few philosophers who has argued explicitly against a “standing” condition on appropriate blame.

Blame and Holding Responsible    279 This simple example is meant to show that the question of how it is appropriate for an individual to respond to herself or others on the basis of her/​their blameworthy conduct is a complex ethical question that must take into consideration factors that go beyond whatever factors are relevant to determining an agent’s blameworthy status. It is unclear, however, just what those relevant factors are, and how exactly they are supposed to function in rendering certain instances of blame appropriate or inappropriate. These questions are further complicated by the fact that our answers may depend upon what blame is, which we have already seen is a topic of intense debate. If blame is simply a form of negative moral appraisal, for example, perhaps it is less subject to contextual norms of appropriateness. If, alternatively, blame consists in hostile reactive attitudes and/​or their public expression, then it is more plausible to claim that they are subject to further norms of appropriateness. In what follows, I will assume that what is at issue are forms of blame that go beyond mere negative appraisal, but will otherwise try to remain neutral about the precise content of blame. Philosophers who have written on this topic have highlighted two general types of consideration that are relevant to assessing the appropriateness of particular blaming responses on the part of particular individuals in particular contexts. The first set of considerations concern issues of standing: Does X have the moral standing to blame Y for Z? The second set of considerations concern what Tognazzini and Coates (2014) have dubbed issues of procedure: Once it is determined that X has standing to blame Y for Z, we can still ask a further set of questions about whether she ought to do so, and if so in what way. In some cases, we may judge that it would be cruel, gratuitous, or excessive to blame another (or to do so in a particular way), while in other cases we may judge that not to blame would itself be morally objectionable. Let me say something briefly about each of these types of consideration.

3.1  Issues of Standing As already mentioned, many philosophers have argued that certain facts about a would-​ be blamer can undermine her standing to blame an otherwise blameworthy agent. The most commonly discussed standing conditions are the Non-​Hypocrisy Condition, the Non-​Complicity Condition, and the Personal Business Condition.11 According to the Non-​Hypocrisy Condition, one’s standing to blame another may be undermined if one is guilty of the very same transgression for which one would be blaming the other. Phrases like, “Pot calling the kettle black!,” “Oh, that’s rich coming from you!” and “Look who’s talking!” attest to the fact that we think certain personal failings on the part of a would-​be blamer can undermine her authority to call others to account. It is

11 

These labels come from Bell (2013).

280   Angela M. Smith important to note that such responses do not, in fact, attempt to rebut the claim that the responder is worthy of blame or criticism. Rather, they are meant to challenge the authority of a particular blamer to adopt a stance of criticism toward her in the circumstances. There is a lively debate in the literature over why hypocrisy should undermine one’s standing to blame. According to Scanlon (2008), the fundamental problem with hypocritical blame is that it rests on a false presumption. On his view, to blame another is to adjust one’s attitudes, intentions, and expectations in response to the judgment that another has attitudes that impair her relations with others. A hypocritical blamer, then, is one who claims that the other has attitudes that have impaired their relationship. And yet, “[t]‌here is something false” in the hypocritical blamer’s suggestion that the other is responsible for the impaired relationship in which they find themselves. After all, if I have engaged in similar wrongdoing in the past, “I cannot claim that the attitudes revealed in your [behavior] constitute an impairment in our relations, because the mutual intentions and expectations that constitute those relations were already impaired by my own similar attitudes, revealed repeatedly in my past conduct” (p. 176). The central problem with hypocritical blame, then, is that it presupposes a false picture of the relationship that obtains between the blamer and the one blamed. Wallace (2010), by contrast, has argued that in order to understand the fundamental problem with hypocritical blame, we must recognize that “blame carries with it a kind of practical commitment to self-​scrutiny” (p. 326). When we blame others, we are also making a commitment to scrutinize our own attitudes and behavior to ensure that they are consistent with the standards to which we are holding others. This commitment, in turn, rests on a presumption of equal moral standing: that we equally waive our right to protection from moral opprobrium if we violate the moral norms for which we blame others. The problem with hypocritical blamers, then, is that they “have failed to live up to the commitment that they have undertaken through the attitudes that constitute their blame” (pp. 326–​327). In so doing, they violate a fundamental principle of equal moral standing that lies at the very heart of morality. Marilyn Friedman (2013), picking up on Wallace’s suggestion, proposes a somewhat simpler diagnosis of what is fundamentally wrong with hypocritical blaming. She claims that “the key problem consists of applying, in an arbitrarily discriminatory manner, moral norms that should be applied in the same way for a wider population. It is unjustified evaluative discrimination that constitutes the key problem in hypocritical moral evaluation and that undermines moral standing to blame for that particular sort of wrongdoing” (p. 280). In this way, Friedman draws out similarities between hypocritical blaming and other forms of “inconsistent” blaming—​such as blaming members of a racial minority for certain sorts of actions while refraining from blaming members of the racial majority for the very same sorts of actions. The problem is not so much that the hypocritical blamer arbitrarily favors herself, then, but simply that she applies moral norms in an arbitrarily discriminatory manner. (An interesting question we might ask about this diagnosis is whether we would find it similarly objectionable for an agent to blame herself for things that she does not blame any others.)

Blame and Holding Responsible    281 However we account for the inaptness of hypocritical blame, there are a number of additional questions that can be raised about the nature and scope of the Non-​ Hypocrisy Condition. If a would-​be blamer has sincerely apologized for her own transgressions in the past, does she regain her standing to blame others who engage in similar transgressions? Does a would-​be blamer’s transgressions in one domain (say, regarding questions of trustworthiness) undermine her standing to blame others for transgressions in another domain (say, regarding truthfulness)? And does the severity of the transgression matter to determining whether one’s standing is undermined (e.g., if a would-​be blamer has engaged in a mild flirtation with a work colleague in the past, does she lose her standing to blame her spouse for engaging in a sexual affair with one of his colleagues?). Given the pervasiveness of charges of “hypocrisy” in social and political life today, this is likely to remain an active area of scholarship for philosophers interested in questions about “the ethics of blame.”12 Turning, then, to the Non-​Complicity Condition, a number of philosophers have argued that one does not have standing to blame others for actions that one has in some way been complicit in encouraging, ordering, or forcing the blameworthy agent to perform. G. A. Cohen discusses the way in which this criticism might be raised by Palestinian terrorists to Israeli officials who have deprived them of any other reasonable means for having their legitimate grievances addressed (Cohen 2006). Tommy Shelby, similarly, has argued that certain kinds of criticisms cannot be made by relatively affluent citizens of a society about the behavior of the ghetto poor, given that the circumstances in which they have been asked to make their life choices are often the result of serious forms of injustice (Shelby 2007). And Antony Duff has proposed that the state’s right to put a criminal defendant on trial can be undermined in some circumstances by the prior misconduct of the state or its officials, if that misconduct left the defendant little choice but to engage in the wrongdoing for which the state is calling upon him to answer (Duff 2010). As with the Non-​Hypocrisy Condition, there is still much debate over how, exactly, to interpret the nature and scope of the Non-​Complicity Condition. For example, it is unclear just how complicit one must be in facilitating an agent’s wrongdoing in order to undermine one’s own standing to blame that agent. Does it matter whether one is aware of one’s complicity, and does it matter whether one actually endorses the wrongdoing in which one is complicit? (Todd 2012, pp. 10–​11). As a general matter, a full defense of the Non-​Complicity Condition will require philosophers to address a number of difficult questions arising in social and political philosophy about the nature and legitimizing force of consent and about the conditions under which social and political institutions may count as unjustly coercive. The third commonly referenced standing condition is the Personal Business Condition. This condition was illustrated above with the example of Douglas: Many of

12 

Johann Frick (2016) has recently raised a closely related, but distinct, question about the possible “speaker-​relativity of justification” and its implications for contractualist moral theory.

282   Angela M. Smith us would find it outrageous if complete strangers felt themselves entitled to blame us or our intimates for moral failures that have absolutely nothing to do with them. This is particularly clear when it comes to expressions of blame. If Douglas were to rebuke Bruce for being late for his running date, it would be natural for Bruce to tell him to “mind his own business.” It is a puzzle, however, why it should be the case that one’s standing to express blame should (sometimes) depend upon one’s relationship to the wrongdoer, to the victim, or to both. After all, if Bruce had slapped Abby rather than simply arriving late to the track, we would be less inclined to think that Douglas should refrain from expressing blame on the grounds that it is “none of his business.” Just what types of moral transgression it is or is not “one’s business” to blame or rebuke is itself an interesting first-​order normative question, and one about which there may well be significant intra-​cultural and cross-​cultural disagreement. Linda Radzik (2014) notes that it is not unusual for strangers on the street in Germany to deliver lengthy rebukes to those who violate minor social rules, while that would be viewed as inappropriate in most parts of the United States (p. 653). It is not clear whether any general account of the “Personal Business Condition” could be given to explain why we may lack standing in certain situations to express blame. But that there is some such standing condition, and that it is distinct from both the Non-​Hypocrisy Condition and the Non-​Complicity Condition, seems evident when we think about what our moral lives would be like if every moral agent took every moral transgression to be “their business” to call out. To sum up, there is widespread agreement in the literature that certain individuals lack “standing” to engage in certain kinds of blame in certain kinds of context. Much more work still needs to be done, however, in order to cash out what the notion of “standing” itself comes to, and to determine whether that notion functions in the same way in all of the types of cases discussed earlier in the chapter. Is having standing a matter of having certain rights, certain privileges, certain powers, or a certain kind of authority? Or is it better to think of the notion of standing as just a catch-​all for referring to a range of first-​order (moral) norms about when it is and is not appropriate to engage in certain kinds of blaming activities?13 As we will see in a moment, there do seem to be additional first-​order (moral) norms governing other aspects of our blaming practices. Whether our ideas about standing can be accounted for in a similar way is still up for debate.

3.2  Issues of Procedure Even if we think that an agent has standing to blame a blameworthy agent, we might still think that there are further norms governing whether, when, and how she should

13 

For such a “deflationary” view of standing, see Matthew King, “Skepticism About the Standing to Blame,” David Shoemaker (ed.), in Oxford Studies in Agency and Responsibility.

Blame and Holding Responsible    283 do so. Blame of a genuinely blameworthy agent may be grossly disproportional to the seriousness of the wrong, may be delivered at an inappropriate time or place, or may be insufficiently sensitive to the agent’s own self-​critical responses to her moral failure. For example, even if Cathy had standing to hold Bruce responsible for his tardiness at the track, we might still think that attitudes and expressions of righteous indignation would be overblown in this context (particularly if we stipulate that this was a one-​time failure on Bruce’s part). We also seem to recognize certain “time, place, and manner” restrictions when it comes to the appropriate expression of moral blame. While intimates surely have standing to hold one another responsible for their moral failures, doing so in very public settings often seems objectionable. Loudly berating one’s spouse for her terrible investment decisions or her slovenly habits at the company Christmas party seems inappropriate, even if the criticism is warranted and one has standing to deliver it. It would be similarly objectionable to deliver a searing moral rebuke to one’s best friend just as she is about to undergo life-​saving surgery. Finally, in some cases a blameworthy agent’s own sincere self-​reproach seems to render further blame on the part of others inappropriate or excessive. If Bruce sincerely apologized for his (one-​time) tardiness and vowed not to let it happen again, it would seem exceedingly churlish for Cathy to continue to harbor and/​or to express resentment to him for his minor transgression. All of these considerations suggest that the blaming attitudes we take and express toward others are themselves subject to various norms of appropriateness, and we can be subject to (moral) criticism for blaming people in ways or at times that violate these norms (Smith 2007, 482–​483). By the same token, failures to blame (or to do so in certain ways) may also leave one open to (moral) criticism. If I do not blame those who engage in serious wrongdoing against others, this seems to imply that I condone those activities, which in turn may show a lack of respect for the moral standing of those who have been wronged (Scanlon 2008, 169–​170). Just what these further norms of appropriateness may be is again an interesting and underdeveloped question in first-​order normative theory. There is further work to be done in determining (1) how to cash out the relevant notion of “appropriateness” (e.g., is it a matter of certain responses being “fitting” or “rational” or “deserved” or “morally obligatory”?), (2) how to go about justifying the existence and the content of such norms (e.g., on prudential, deontological, utilitarian, contractualist, or virtue ethical grounds), and (3) whether the norms of appropriateness embodied in our current moral practices are themselves morally justifiable. On this last issue, an interesting literature is beginning to emerge that highlights the ways in which our practices of “holding responsible” are influenced by social and political factors such as oppression, structural injustice, and hierarchies of power structured around markers of race, gender, and class (Hutchison, Mackenzie, and Oshana 2018; Fricker 2016, 180–​ 182; Mason 2019). It may be that the current “norms of appropriateness” implicit in our blaming practices are themselves ethically problematic and in need of revision. This is yet another reason why the recent blossoming of philosophical interest in the concepts of blame and holding responsible should be regarded as a welcome development in recent theorizing about moral responsibility.

284   Angela M. Smith

References Bell, Macalester. 2013. “The Standing to Blame: A Challenge,” in D. Justin Coates and Neal Tognazzini (eds.), Blame: Its Nature and Norms. New York: Oxford University Press, 263–​281. Brink, David O. and Dana Nelkin. 2013. “Fairness and the Architecture of Responsibility,” in D. Shoemaker (ed.), Oxford Studies in Agency and Responsibility, vol. 1. New York: Oxford University Press, 284–​213. Coates, D. Justin and Neal Tognazzini. 2012. “The Nature and Ethics of Blame,” Philosophy Compass 7(3): 197–​207. Coates, D. Justin and Neal Tognazzini. 2013. Blame: Its Nature and Norms. New York: Oxford University Press. Cohen, Gerald. 2006. “Casting the First Stone: Who Can, and Who Can’t, Condemn the Terrorists?,” Royal Institute of Philosophy Supplement 58: 113–​136. Cogley, Zac. 2013. “The Three-​Fold Significance of Blaming Emotions,” in David Shoemaker (ed.), Oxford Studies in Agency and Responsibility. New York: Oxford University Press, 205–​224. Darwall, Stephen. 2006. The Second-​Person Standpoint: Morality, Respect, and Accountability. Cambridge, MA: Harvard University Press. Duff, R. A. 2010. “Blame, Moral Standing, and the Legitimacy of the Criminal Trial,” Ratio n.s. 23: 123–​140. Feinberg, Joel 1970. Doing and Deserving. Princeton, NJ: Princeton University Press. Fischer, John Martin and Mark Ravizza, eds. 1993. Perspectives on Moral Responsibility. Ithaca, NY: Cornell University Press. Frick, Johann. 2016. “What We Owe to Hypocrites: Contractualism and the Speaker-​Relativity of Justification,” Philosophy and Public Affairs 44(4): 223–​265. Fricker, Miranda. 2016. “What’s the Point of Blame? A Paradigm Based Explanation,” Noûs 50(1): 165–​183. Friedman, Marilyn. 2013. “How to Blame People Responsibly,” The Journal of Value Inquiry 47: 271–​284. Glover, Jonathan 1970. Responsibility. London: Routledge and Kegan Paul. Graham, Peter A. 2012. “A Sketch of a Theory of Moral Blameworthiness,” Philosophy and Phenomenological Research 88(2): 388–​409. Haji, Ishtiyaque. 1998. Moral Appraisability: Puzzles, Proposals, and Perplexity. New York: Oxford University Press. Hieronymi, Pamela. 2001. “Articulating an Uncompromising Forgiveness,” Philosophy and Phenomenological Research 62(3): 529–​555. Hieronymi, Pamela 2004. “The Force and Fairness of Blame,” Philosophical Perspectives 18: 115–​148. Hutchison, Katrina, Catriona Mackenzie, and Marina Oshana, eds. 2018. Social Dimensions of Moral Responsibility. New York: Oxford University Press. King, Matthew. 2019. “Skepticism About the Standing to Blame,” in David Shoemaker (ed.), Oxford Studies in Agency and Responsibility. New York: Oxford University Press. Macnamara, Coleen. 2015. “Blame, Communication, and Morally Responsible Agency,” in R. Clarke, M. McKenna, and A. M. Smith (eds.), The Nature of Moral Responsibility: New Essays. New York: Oxford University Press, 211–​235.

Blame and Holding Responsible    285 Mason, Elinor. 2019. “Between Strict Liability and Blameworthy Quality of Will: Taking Responsibility,” in David Shoemaker (ed.), Oxford Studies in Agency and Responsibility. New York: Oxford University Press. McGeer, Victoria. 2013. “Civilizing Blame,” in D. Justin Coates and Neal Tognazzini (eds.), Blame: Its Nature and Norms. New York: Oxford University Press, 162–​188. McGeer, Victoria. 2015. “Building a Better Theory of Responsibility,” Philosophical Studies 172: 2635–​2649. McKenna, Michael. 2012. Conversation and Responsibility. New York: Oxford University Press. Menges, Leonhard. 2017. “The Emotion Account of Blame,” Philosophical Studies 174: 257–​273. Nelkin, Dana. 2015. “Psychopaths, Incorrigible Racists, and the Faces of Responsibility,” Ethics 125(2): 357–​390. Nelkin, Dana. 2016. “Accountability and Desert,” Journal of Ethics 20: 173–​189. Pereboom, Derk. 2014. Free Will, Agency, and Meaning in Life. New York: Oxford University Press. Pickard, Hanna. 2013. “Irrational Blame,” Analysis 73(4): 613–​626. Radzik, Linda. 2014. “Moral Rebukes and Social Avoidance,” Journal of Value Inquiry 48: 643–​661. Rosen, Gideon. 2015. “The Alethic Conception of Moral Responsibility,” in R. Clarke, M. McKenna, and A. M. Smith (eds.), The Nature of Moral Responsibility: New Essays. New York: Oxford University Press, 65–​87. Scanlon, T. M. 1988. “The Significance of Choice.” Tanner Lectures on Human Values 8: 149–​216. Scanlon, T. M.. 1998. What We Owe to Each Other. Cambridge, MA: Harvard University Press. Scanlon, T. M.. 2008. Moral Dimensions: Permissibility, Meaning, Blame. Cambridge, MA: Harvard University Press. Schlick, Moritz 1939. “When Is a Man Responsible?,” in D. Rynin (trans.), Problems of Ethics. New York: Prentice Hall, Inc. Shelby, Tommie. 2007. “Justice, Deviance, and the Dark Ghetto,” Philosophy and Public Affairs 35(2): 126–​160. Sher, George 2006. In Praise of Blame. Oxford: Oxford University Press. Shoemaker, David. 2015. Responsibility from the Margins. New York: Oxford University Press. Smart, J. J. C. 1961. “Free Will, Praise, and Blame,” Mind 70: 291–​306. Smith, Angela M. 2007. “On Being Responsible and Holding Responsible,” Journal of Ethics 11: 465–​484. Smith, Angela M. 2013. “Moral Blame and Moral Protest,” in D. Justin Coates and Neal Tognazzini (eds.), Blame: Its Nature and Norms. New York: Oxford University Press, 27–​48. Smith, Angela M. 2015. “Responsibility as Answerability,” Inquiry 58(2): 99–​126. Smith, Angela M. 2019. “Who’s Afraid of a Little Resentment,” in David Shoemaker (ed.), Oxford Studies in Agency and Responsibility. New York: Oxford University Press. Todd, Patrick. 2012. “Manipulation and Moral Standing: An Argument for Incompatibilism,” Philosopher’s Imprint 12. Strawson, P. F. 1962. “Freedom and Resentment,” Proceedings of the British Academy 48: 1–​25; Reprinted in John Martin Fischer and Mark Ravizza (eds.), 1993. Perspectives on Moral Responsibility. Ithaca, NY: Cornell University Press, 45–​66. Talbert, Matthew. 2012. “Moral Competence, Moral Blame, and Protest,” Journal of Ethics 16: 89–​109. Tognazzini, Neal. 2013. “Blameworthiness and the Affective Account of Blame,” Philosophia 41: 1299–​1312.

286   Angela M. Smith Tognazzini, Neal and D. Justin Coates. 2014. “Blame,” The Stanford Encyclopedia of Philosophy (Spring 2016 ed.), Edward N. Zalta (ed.), URL=https://​plato.stanford.edu/​archives/​spr2016/​ entries/​blame/​. Vargas, Manuel. 2013. Building Better Beings: A Theory of Moral Responsibility. New York: Oxford University Press. Walker, Margaret Urban. 2006. Moral Repair: Reconstructing Moral Relations After Wrongdoing. Cambridge: Cambridge University Press. Wallace, R. Jay. 1994. Responsibility and the Moral Sentiments. Cambridge, MA: Harvard University Press. Wallace, R. Jay. 2010. “Hypocrisy, Moral Address, and the Equal Standing of Persons,” Philosophy and Public Affairs 38: 307–​341. Wallace, R. J. 2011. “Dispassionate Opprobrium: On Blame and the Reactive Sentiments,” in Samuel Freeman, Rahul Kumar, and R. Jay Wallace (eds.), Reasons and Recognition: Essays on the Philosophy of T. M. Scanlon. New York: Oxford University Press, 348–​372. Watson, Gary. 1987. “Responsibility and the Limits of Evil: Variations on a Strawsonian Theme,” in Ferdinand Schoeman (ed.), Responsibility, Character, and the Emotions: Essays in Moral Psychology. Cambridge: Cambridge University Press, 256–​286. Watson, Gary. 1996. “Two Faces of Responsibility,” Philosophical Topics 24(2): 227–​248. Watson, Gary 2011. “The Trouble with Psychopaths,” in Samuel Freeman, Rahul Kumar, and R. Jay Wallace (eds.), Reasons and Recognition: Essays on the Philosophy of T. M. Scanlon. Oxford: Oxford University Press, 307–​331. Wolf, Susan. 1990. Freedom Within Reason. New York: Oxford University Press. Wolf, Susan (2011). Meaning in Life and Why It Matters (Markus Rüther). Philosophischer Literaturanzeiger 64 (3):308. Wolf, Susan (2011). “Blame, Italian Style,” in Samuel Freeman, Rahul Kumar, and R. Jay Wallace (eds.), Reasons and Recognition: Essays on the Philosophy of T. M. Scanlon. New York: Oxford University Press, 332–​347. Zimmerman, Michael 1988. An Essay on Moral Responsibility. Totowa, NJ: Rowman and Littlefield.

chapter 14

Resp onsibil i t y a nd Reactive At t i t u de s R. Jay Wallace

One of the most influential contributions to modern discussions about freedom and responsibility is P. F. Strawson’s essay “Freedom and Resentment.”1 First published in 1963, this paper sketched a new way of thinking about the issues in this area, one that emphasized the significance to our interpersonal practices of the reactive attitudes (as Strawson called them). Strawson’s central idea is that emotions such as resentment and indignation are constitutive of treating others as morally accountable individuals, and that a susceptibility to such attitudes is given to us with our participation in interpersonal relationships. Many contemporary philosophers consider themselves Strawsonians, though they differ amongst themselves as to what exactly that means. One commitment that Strawsonians indubitably share, however, is to the significance of the reactive attitudes to moral responsibility. My objective in the present contribution is to explore this interesting and important idea. What exactly are the reactive attitudes? What role do these attitudes play in our practice of holding people accountable for what they do? What implications might the reactive attitudes have for our understanding of the conditions of responsibility (including the question of whether individuals can be responsible for their actions in a deterministic world)? My main suggestion will be that there is an interpretation of the reactive attitudes on which they are plausibly understood to be central to our practice of holding one another morally to account. To be a responsible agent, on this way of understanding things, is to be a fit target of the reactive attitudes, so construed. I shall argue, further, that this conclusion helps us to understand better what debates about the conditions of responsibility are really about, even if it does not resolve those debates by itself. 1  P. F. Strawson, “Freedom and Resentment,” as reprinted in Gary Watson, ed., Free Will, Second Edition (Oxford: Oxford University Press, 2003), pp. 72–​93. Parenthetical references in the text of the chapter will be to pages in this edition of Strawson’s article.

288   R. Jay Wallace

1.  Reactive Attitudes and Human Relationships There are different ways we might understand the reactive attitudes as a class. Strawson himself thinks of them as emotions we are subject to in virtue of our involvement in human relationships, and this determines a relatively capacious understanding of their extension. Thus what he calls “personal reactive attitudes” include “such things as gratitude, resentment, forgiveness, love, and hurt feelings” (75), as well as anger and “personal antagonism” (79–​80). Emotions of these kinds are attitudes we take toward other individuals to whom we stand in significant interpersonal relationships. But Strawson thinks there are other reactive attitudes in addition to these, including “vicarious” and “self-​reactive” emotions; the former category encompasses indignation and moral disapprobation, and the latter extends to guilt, remorse, and the sense of obligation (83–​85). Strawson makes several further claims about the reactive attitudes, of which the following are perhaps most significant. First, these emotions are “reactive” insofar as they are reactions to the attitudes of other people toward us. They respond to the quality of will that is displayed toward us by the actions or omissions of other people, and this is in some sense what they are fundamentally about. Second, he often associates reactive attitudes with demands that we make of each other to a certain level of consideration or regard. About the personal reactive attitudes, for instance, he writes that they “rest on, and reflect, an expectation of, and demand for, the manifestation of a certain degree of goodwill or regard on the part of other human beings toward ourselves” (84). So our susceptibility to reactive attitudes is connected to our adopting toward each other a stance of demanding that we adopt attitudes of goodwill or consideration. Third, when these demands are flouted, the result amounts to a kind of injury or offense by one party against another. What is meant, I think, is not that the injured party is necessarily harmed or caused to suffer in some independent, physical, or emotional sense, but that the failure to meet the demand itself constitutes a kind of moral injury or offense. One injures another party, in the relevant sense, just in case one fails to display toward them that level of consideration or regard that we expect or demand of each other in our ordinary interactions. These three features of the reactive attitudes—​we might call them quality of will, demand, and injury—​are most clearly met with in a subset of the cases that Strawson himself mentions. This subset includes the attitude that features in the title of Strawson’s article, namely resentment, as well as indignation, guilt, and remorse. These are the emotions that seem most plausibly to be about the kind of moral injury that involves the flouting of the demands we hold each other to for a certain level of consideration or regard. We feel resentment, paradigmatically, when we have been wronged by another party, where to be wronged in the relevant sense is to be treated without that level of consideration or goodwill that we legitimately expect people to show us. A similar structure is implicit in the vicarious and self-​reactive attitudes, it seems. With indignation,

Responsibility and Reactive Attitudes    289 for instance, we react vicariously on behalf of a third party who has been wronged by someone else. The wrong involves the flouting of demands that the party might make of others on their own behalf, and in responding vicariously we show that we expect people to treat each other in accordance with the same standards we expect them to uphold in their treatment of us. The sense of guilt or remorse likewise shows that we hold ourselves to interpersonal expectations, and respond with distress when we realize that we have not lived up to those expectations in our treatment of others. If this way of understanding the reactive attitudes is to be plausible, however, we will need to understand demands or expectations in the right way. There is an epistemic sense of expectation, for instance, on which to expect something to happen is to believe with a sufficiently high probability that it will come to pass. Expectations of this epistemic kind render us vulnerable to characteristic emotions when they are frustrated, but the emotions are typically in the key of surprise or disappointment. The reactive attitudes, by contrast, have an angry character that reflects a different way of expecting or demanding something from a person. It is the kind of expecting or demanding that is at issue when we address demands to someone interpersonally (by saying, e.g., “Hey, you can’t do that!”). This normative way of making a demand or holding someone to an expectation, I believe, ultimately cannot be made sense of apart from the reactive attitudes.2 That is, to make a demand of this kind just is, inter alia, to be disposed to the reactive emotions in case the demand should be flouted, or to believe those emotions would be fitting or appropriate under those circumstances. Normative expectations are thus not available to be analyzed independently of the reactive emotions. Rather, the two need to be understood together. As Strawson puts the point: “attitudes of disapprobation and indignation are precisely the correlates of the moral demand in the case where the demand is felt to be disregarded. The making of the demand is the proneness to such attitudes” (90). On this plausible approach, normative expectations function to define the level of consideration or regard below which reactive emotions are apt to be triggered (or viewed, at any rate, as fitting or appropriate). But they are not in place, interpersonally, independently of the reactive attitudes whose conditions they specify.3 The interpretation I have been developing applies nicely to the cases of resentment, indignation, and guilt. But there is a question about whether it can be extended straightforwardly to all of the items in the broader class of attitudes that Strawson characterizes as reactive. Love and gratitude, for instance, are not direct responses to anything that might be characterized as a moral injury; nor do they obviously involve the kind of normative demands that we have seen to connect to reactions of angry disapprobation. So

2 

For further development of this suggestion, see my Responsibility and the Moral Sentiments (Cambridge, MA: Harvard University Press, 1994), chap. 2. 3  I have recently argued, more specifically, that the moral demands to which the narrowly reactive attitudes are connected should be understood to have a relational content, defining directed requirements that are owed to other individuals. See my “Emotions and Relationships: On a Theme from Strawson,” in David Shoemaker and Neal A. Tognazzini, eds., Oxford Studies in Agency and Responsibility 2 (2014), pp. 119–​142; also my The Moral Nexus (Princeton, NJ: Princeton University Press, 2019), chap. 3.

290   R. Jay Wallace what might be meant in characterizing them as reactive attitudes? A promising answer to this question might appeal to the idea that, though love and gratitude are not themselves immediate responses to the flouting of interpersonal demands, they are emotions that presuppose that the person subject to them addresses such demands to the individual at whom the emotions are targeted. On this approach, to love someone in the relevant way is to be involved with them in a relationship that is structured around normative expectations, and hence to be vulnerable to more narrowly reactive attitudes when the beloved treats one with a lack of consideration and regard. And similarly for gratitude and (perhaps) personal antagonism. This may be what Strawson had in mind in adopting the capacious conception of reactive attitudes implicit in the examples that are cited in “Freedom and Resentment.” But it is not entirely clear whether the commitment that underlies this approach is really plausible. The question is whether one can stand in an ordinary loving relationship with someone, or feel grateful to them, without at the same time holding them to normative demands or expectations, in the way that renders one vulnerable to attitudes of angry disapprobation when the expectations are breached. Reason to question whether this is the case is provided by familiar Christian characterizations of the believer’s attitudes toward God, which include love and gratitude, but which do not obviously implicate the demanding stance that renders the believer susceptible to resentment or indignation toward God. This is admittedly a special case, but it suggests the coherence of love and gratitude in the absence of any susceptibility to the more narrowly reactive attitudes. In the case of relationships between people—​to extend the skeptical line of thought—​some have at least aspired to a kind of love that takes its objects as they are, and does not direct toward them demands to be or to act otherwise. If someone could manage to pull this off, they might also feel gratitude toward the person whom they love, without being vulnerable on that account to the more narrowly reactive emotions already discussed. A different response to this challenge would be to relax still further the criteria for what counts as a reactive attitude. Instead of connecting them definitionally to the normative demands and expectations that Strawson himself tends to emphasize, we might interpret them simply as attitudinal responses to the attitudes of other people toward us. This relaxed interpretation would accommodate most cases of love and gratitude and personal antagonism. But the cost is a less structured understanding of reactivity, which renders it difficult to make sense of some of the other important claims that have been made about the significance of the reactive attitudes; this is a point to which I shall return in the next section. In the meantime, I would like to address another issue that has figured prominently in discussions of the reactive attitudes. In his original paper on the topic, Strawson drew a contrast between two different attitudes that could be taken toward human beings. There is, first, the “participant” attitude of involvement in ordinary interpersonal relationships, which is the default stance we adopt in our ordinary interactions with other people. And there is, second, the “objective” attitude that we take when we view people as objects to be studied or controlled or cured or trained. The objective stance is associated with certain familiar professional roles, including those of researchers or therapists or managers; but

Responsibility and Reactive Attitudes    291 it is also occasionally available to ordinary individuals outside of these contexts and roles, for instance as a refuge from the “strains of involvement” (79–​80). Strawson thinks that objectivity of this kind is not sustainable indefinitely. For better or worse, we are involved in ordinary interpersonal relationships with people, and we cannot persistently avoid the attitude of participation that goes along with such relationships. Furthermore, though objectivity is opposed to the reactive attitudes, those attitudes are endemic to the participant stance. As Strawson writes, “being involved in inter-​personal relationships as we normally understand them precisely is being exposed to the range of reactive attitudes and feelings” (81). The upshot is that the reactive attitudes are virtually inevitable for us, given to us with our necessary participation in interpersonal relationships with other humans. But how plausible is the connection Strawson postulates between the participant stance and the reactive attitudes? The answer arguably depends on how we understand the criteria for reactivity. As noted earlier, if we emphasize normative expectations or demands as essential to the reactive attitudes as a class, it becomes correspondingly questionable whether exposure to them really is part and parcel of involvement in interpersonal relationships. Those who are able to accept their loved ones as they are, without demanding or expecting that they should be otherwise, would be participants in human relationships, without being disposed to react to the individuals to whom they are related with emotions such as resentment or indignation. The more relaxed interpretation of the reactive attitudes sketched earlier, by contrast, yields a more convincing account of the connection Strawson postulates between reactivity and the participant stance. If reactive attitudes are simply emotional responses to the attitudes that other people take toward us, then it is indeed hard to see how one could stand in an interpersonal relationship without being susceptible to at least some states of this general kind.

2.  Reactive Attitudes and Responsibility Strawson himself took the points just discussed, about the connections between participation in human relationships and the reactive attitudes, to be significant for debates about the conditions of responsibility. Before we can assess this suggestion, we need to get clear about the role of the reactive attitudes in our responsibility practices. What does moral responsibility have to do with the reactive emotions? A natural answer to this question starts with the phenomenon of moral blame. The most basic sign that you hold someone accountable for their actions is the tendency to blame them when they act impermissibly. But blame is a protean phenomenon, which is difficult to pin down.4 On the one hand, to blame someone in the relevant sense is not 4  A helpful collection of recent contributions on moral blame, which illustrates the range of approaches taken in contemporary discussions, is D. Justin Coates and Neal A. Tognazzini, eds., Blame: Its Nature and Norms (New York: Oxford University Press, 2013).

292   R. Jay Wallace merely to judge that they have done something that is morally wrong. You can deem someone to have acted impermissibly without blaming them for their infraction. You might for instance think that they are not fully accountable for their actions; or you might simply not be in the mood to judge them for what they have done. So blame seems to go beyond believing another to have acted wrongly (even if that is a necessary element in blame). On the other hand, blame does not go beyond a judgment of wrongdoing by involving punishment or the imposition of sanctions on the wrongdoer. This is in fact something that adults rarely do in their interactions with each other, and yet blame seems pretty commonplace in the context of such interactions. A central challenge for an account of blame is to explain how it goes beyond judgments of wrongdoing without assimilating it to punishment or sanctioning behavior. The reactive attitudes seem potentially significant in this connection. Take the central cases of resentment, indignation, and guilt. These intuitively seem to be attitudes through which we blame people for their moral infractions. As we saw in the preceding section, they are naturally understood to be responses to the flouting of normative expectations or demands, so they presuppose judgments of wrongdoing on the part of the agent who is subject to them. But they also clearly go beyond such judgments. They are reactive emotions, ways of being exercised about what the wrongdoer has done, which are broadly in the key of angry disapprobation. One can clearly be emotionally exercised in this way without acting to punish the targets of one’s attitude or to impose sanctions on them. At the same time, the imposition of sanctions and punishments, when we do go in for it, is plausibly understood to express attitudes of angry disapproval toward the individuals who are sanctioned or punished. This helps us to understand how such actions might connect to the emotional attitudes that are in the first instance constitutive of moral blame, as they also clearly seem to do. Blame thus begins to become tractable, philosophically, if we understand it in terms of Strawson’s core reactive emotions.5 But if blame can be understood in these reactive terms, it is not clear that the same goes for moral responsibility more broadly. For we can regard people as morally accountable for their actions without blaming them in this reactive sense. For one thing, it might be that they have not (yet) done anything that is

5  For criticisms of the suggestion that anger and related attitudes are central to blame, see, e.g., George Sher, In Praise of Blame (Oxford: Oxford University Press, 2006), chap. 5; others who de-​emphasize reactive emotion in their accounts of blame include T. M. Scanlon in Moral Dimensions: Permissibility, Meaning, Blame (Cambridge, MA: Harvard University Press, 2008), chap. 4; Derk Pereboom in Free Will, Agency, and Meaning in Life (Oxford: Oxford University Press, 2014), chap. 6; and Angela Smith in “Moral Blame and Moral Protest,” in Coates and Tognazzini, eds, Blame, pp. 27–​48. A sticking point for critics of reactive blame is the fact that we blame people for infractions that occur in the context of loving and warm relationships between family members and intimates, where overt hostility and ill will seem out of place. A vigorous case for the compatibility of angry blame with relationships of this kind is made by Susan Wolf in “Blame, Italian Style,” in R. Jay Wallace, Samuel Freeman, and Rahul Kumar, eds., Reasons and Recognition: Essays on the Philosophy of T. M. Scanlon ((New York: Oxford University Press, 2011), pp. 332–​347.

Responsibility and Reactive Attitudes    293 morally objectionable. For another, even when they have acted wrongly, we might not be able to work up any opprobrium about what they have done; perhaps we feel that the infraction does not sufficiently implicate us, or perhaps there are other things that make claims on our attention and emotional energy. What are we doing when we treat people as morally accountable agents, given that doing so does not necessarily involve responding to their actions with the reactive attitudes that are arguably constitutive of blame? To answer this question, we might return to the idea of a normative expectation or demand that seems implicit in the attitudes of resentment, indignation, and guilt. These emotions, I have suggested, seem to be occasioned by violations of the demands we hold people to for consideration or regard. Following Strawson, I suggested further that holding someone to a demand of the relevant kind is constitutively connected to the narrowly reactive emotions; “the making of the demand is the proneness to such attitudes” (90). Perhaps, then, we regard people as morally accountable agents when we address normative demands to them in this distinctive way. This is clearly a stance that we can take toward someone regardless of whether they have (yet) acted in violation of the demands that are addressed to them. And it is also compatible with our not experiencing any occurrent reactive emotion in response to infractions that the responsible agent has in fact committed: we can be prone to such attitudes, and deem them appropriate responses to the agent’s flouting of the relevant demands, even on occasions when the attitudes in question are not actually experienced by us. Under these conditions, we might be said to regard someone as blameworthy despite the fact that we do not ourselves blame them, precisely insofar as it would be fitting to respond to what they have done with the attitudes constitutive of blame. An account of moral accountability and blameworthiness can thus be constructed out of the reactive interpretation of attitudinal blame. The reactive approach to responsibility strikes me as highly plausible, and I have defended it in other work.6 The approach starts with the idea that blame paradigmatically involves a kind of angry disapprobation, an idea that makes good sense of the way in which blame goes beyond judgments of wrongdoing. To regard someone as responsible, then, is to hold them to moral expectations or demands, in the way that renders one susceptible to the reactive attitudes when those demands are violated. Note, however, that the approach trades on the assumption that the reactive attitudes are structured around normative expectations and demands, an assumption that applies most clearly to the core examples of resentment, indignation, and guilt. A different way of thinking about responsibility might be built on the more relaxed interpretation of the reactive attitudes that was sketched earlier. On the approach I have in mind, to be a responsible agent is to be someone whose actions have a certain kind of normative significance for other parties. Specifically, it is to be someone who

6 See Responsibility and the Moral Sentiments, chap. 3 and my “Dispassionate Opprobrium,” in

Wallace et. al., eds., Reason and Recognition, pp. 348–​72; see also Wolf, “Blame, Italian Style.”

294   R. Jay Wallace is capable of attitudes in the broad spectrum ranging from good will to disregard, including such familiar attitudes as consideration, solicitude, indifference, hostility, and contempt. An individual who displays attitudes toward you on the negative end of this spectrum thereby alters their relationship to you, giving you reason to adjust your behavior in turn. If you are treated with hostility or contempt, for instance, then it seems you have good reason to withdraw from trust-​based relationships with the person who has treated you in this way, and perhaps to avoid the person altogether if it is open to you to do so. To modify your comportment in these ways in response to attitudes of disregard or ill will, it has been suggested, is the essence of moral blame.7 Understood along these lines, blame is primarily a matter of adjusting your behavior in regard to someone, in response to attitudes on their part that impair their relationship with you. Call this the relationship-​impairment conception of blame. This approach places behavioral responses at the center of moral blame, including such responses as withdrawal, avoidance, and disengagement. But behavioral responses of these kinds have an attitudinal dimension, insofar as they involve intentions to conduct yourself in certain ways. These intentions, furthermore, are formed in reaction to the attitudes of disregard or ill will that others adopt toward you. On the relaxed interpretation introduced in section 1, they might therefore be considered reactive attitudes; those who are subject to them will naturally be prone to experience other emotions as well, such as sadness or distress, which reflect the fact that they care about the relationships that the wrongful parties have impaired. Blame goes beyond judgments of wrongdoing, on the relationship-​impairment conception, because it essentially involves reactive attitudes of this extended kind. There is a real question, however, about how plausible this way of developing a Strawsonian approach would be. As I just mentioned, the relaxed interpretation of reactive attitudes with which it operates includes states such as beliefs and intentions to act, which do not necessarily have any emotional coloration at all. Thus it is possible to believe that someone has shown contempt or disregard for your interests, and deliberately adjust your behavior toward them in response, in a completely dispassionate spirit, without being emotionally exercised (as I put it above) by what has happened to you. Or if one is not completely dispassionate about it, the emotions one experiences might be feelings of sadness or sorrow or melancholy.8 According to the relationship-​impairment approach, these reactions to wrongdoing would count as episodes of moral blame, exhibiting the characteristic modifications in behavior in response to the impairment of your relationship with someone through their attitudes toward you. The most familiar episodes of moral blame, however, are hotter than this in their emotional temperature; they are ways of being exercised by the ill will of the wrongdoer, which reflect the fact that one’s normative expectations have been flouted. This 7 

This positive account of blame is developed by Scanlon in Moral Dimensions, chap. 4. A point Scanlon emphasizes, in Moral Dimensions, p. 136. Pereboom also emphasizes that his forward-​looking, non-​angry kind of blame can be emotionally infused in these ways, in Free Will, Agency, and Meaning in Life, chap. 6. 8 

Responsibility and Reactive Attitudes    295 important dimension of blame seems to be captured better by what I earlier called the reactive account, which interprets blame in terms of the core reactive attitudes of resentment, indignation, and guilt. These emotions are forms of angry disapprobation, which reflect the interpersonal demands we make of each other for a basic level of consideration and regard; these are features that correspond well to central elements in the most paradigmatic cases of blame. For this reason, many philosophers believe the more narrowly reactive emotions to be important to a broadly Strawsonian approach to responsibility, and would take the most iconic instances of moral blame to involve the angrier attitudes that are structured through interpersonal expectations. This would be in the spirit of the reactive account of responsibility sketched earlier.

3.  Optimism About Responsibility Strawson characterized himself as a kind of “optimist” about moral responsibility (93). He thought our practices of holding one another responsible could be seen to be in basically good order, once we understand those practices along the lines of the reactive account. Many contemporary Strawsonians follow him on this point, though there are very different ways of developing the implications of the reactive approach for traditional debates about freedom and responsibility. One family of positions starts from the idea that the reactive attitudes as a class are essential to the participant stance of involvement in interpersonal relationships. If we take this idea as given, it can easily seem as if there is no genuine question of whether we might overcome or cast aside our practices of holding people responsible, and hence no need to grapple seriously with critical challenges to those practices. Humans are inherently social creatures, and there is no realistic prospect of living a human life that does not involve the participant stance of involvement in interpersonal relationships. If reactive attitudes are given to us with the participant stance, it appears to follow that we are committed, in virtue of our inherently social nature, to treating those we interact with as morally responsible agents. This would be one way of protecting responsibility from worries of the kind that are prominent of the pessimistic tradition of incompatibilism about freedom and the conditions of responsibility. We might think of this as a naturalistic variant of Strawsonian optimism, since it emphasizes our nature as inherently social beings. A different position that starts from a similar premise emphasizes the advantages to human life of the kind of interpersonal relationships that go together with the participant stance. Thus, even if it were in our nature to abandon the participant stance and adopt a thoroughgoing objectivity of attitude, this is not something that any sane human being would want to do. It would mean the sacrifice of the kinds of relationships that most of us live for, and that are essential to anything recognizable as a flourishing human life. If things like friendship and attachment are in this way crucial to human well-​being, and if having such relationships with people constitutively involves treating

296   R. Jay Wallace them as morally accountable for what they do, then it would seem superlatively important that we hang onto our responsibility practices in something like their current form. This could be understood as a pragmatic justification of our responsibility practices, which appeals to considerations about the prospective “gains and losses to human life” (83) that would follow if we endeavored to retain those practices or to overcome them comprehensively. Neither of these ways of developing the optimist position seems entirely compelling, however.9 For one thing, they both rest on the idea that the participant stance of involvement in human relationships necessarily renders us susceptible to reactive attitudes in our interactions with each other. As we saw in section 1, this idea seems plausible only if the reactive attitudes are interpreted in the most capacious way possible, simply as attitudes that we have about the attitudes that other people harbor toward us. But this relaxed interpretation of them detaches from the reactive attitudes the element of normative expectation or demand discussed in the earlier sections of this chapter. This is the element that distinguishes the forms of angry disapprobation that figure in many of the most prototypical episodes of blame, such as resentment, indignation, and guilt. It may well be that participation in human relationships goes together with a susceptibility to reactive attitudes in the most relaxed sense, such as sorrow about the fact that someone has treated one with contempt or hostility, and an intention to withdraw from trust-​ based interactions with that person in the future. As noted in the preceding section, however, it isn’t clear that a susceptibility to attitudes of this kind suffices for holding the target of those attitudes accountable. To the extent this is the case, it may well be doubted whether the participant stance constitutively involves a disposition to hold people morally accountable for what they do. There is a long philosophical tradition of skepticism about anger and related emotional reactions to wrongdoing. A large and diverse group of thinkers, including Stoics, Buddhists, Confucians, utilitarians, and others, have argued that anger, resentment, and related emotional formations do not make evaluative sense, and that they are implicated in forms of interpersonal relation that are problematic both for the bearers of the emotions and for those at whom they are targeted.10 Without endeavoring to assess this important and multifaceted tradition of skeptical reflection, two observations seem to me in order about it. First, insofar as they urge us to overcome angry forms of disapprobation, of the kind associated with what I have called normative expectations and demands, these thinkers are naturally understood to be expressing skepticism about our practices of blaming people for actions that disregard important moral requirements. A social world without resentment and indignation is one in which people do not

9  A more sympathetic discussion of what I called the naturalistic argument is Gary Watson, “Peter Strawson on Responsibility and Sociality,” in Shoemaker and Tognazzini, eds., Oxford Studies in Agency and Responsibility 2, pp. 15–​32. 10  For defenses of this skeptical position, see Martha Nussbaum, Anger and Forgiveness: Resentment, Generosity, and Justice (New York: Oxford University Press, 2016), and Owen Flanagan, The Geography of Morals: Varieties of Moral Personality (New York: Oxford University Press, 2016), chaps. 10–​12.

Responsibility and Reactive Attitudes    297 characteristically hold each other to moral requirements, in the ways paradigmatically associated with moral blame, though they may still avoid wrongdoers and withdraw from trust-​based relationships with them, feel sorrow at episodes of wrongdoing, and so on. Second, those who rail against the angrier forms of reactive sentiment do not imagine that we should opt out of the participant stance of involvement in interpersonal relationships with people. On the contrary, they think we should continue to participate in such relationships, only with an orientation of generosity and love toward other individuals rather than a tendency to stand in judgment over them. If these twin commitments are coherent and plausible, however, they tell against the Strawsonian suggestion that participation in interpersonal relationships goes together with a disposition to hold people morally accountable for their actions. There is a further and distinct reservation about the naturalistic and pragmatic versions of Strawsonian optimism. Even if the participant stance does constitutively commit us to holding people morally responsible, it is not clear that this will really suffice to protect the practice of responsibility from skeptical undermining. Incompatibilist skeptics traditionally worry that it wouldn’t ultimately make sense to hold people fully accountable for their actions in a deterministic world, or under conditions in which human agency is simply part of the natural causal order. One way of articulating this thought would be to say that it wouldn’t really be fair or reasonable to hold people to account under the conditions just described. But this thought might be true even if Strawson’s naturalistic and pragmatic conclusions are also granted. That is, a susceptibility to reactive attitudes might well be endemic to human social life, and bound up with forms of interpersonal relation that are superlatively valuable. Consistently with those conclusions, however, it might also be the case that the reactive attitudes, considered just on their own, are always unreasonable or unfair in a world in which human actions are the result of deterministic or natural causal processes. This would be enough, I think, to support the incompatibilist thought that our responsibility practices are not fundamentally in good order. Those practices might belong to forms of social relation that are unavoidable for us, or invaluable to us, or both; but they might also represent forms of reaction to wrongdoing that do not ultimately make sense.11 A different strand in Strawson’s optimistic reflections about responsibility speaks directly to this last concern. This line of thought does not emphasize our natural commitment to the reactive attitudes through which we hold people accountable, or the advantages to human life of the participant stance that goes together with that commitment. Instead, it focuses on the internal conditions to which the reactive attitudes themselves are directly responsive.12 As noted earlier, Strawson calls the emotions at the center of his account “reactive,” because they are reactions to the attitudes that others 11 

See my Responsibility and the Moral Sentiments for a defense of this way of understanding the traditional debate about the conditions of responsibility. 12  Cf. Gary Watson, “Responsibility and the Limits of Evil: Variations on a Strawsonian Theme,” as reprinted in his Agency and Answerability: Selected Essays (Oxford: Clarendon Press, 2004), pp. 219–​59, p. 223 (on the role of “internal criteria” for reactive emotions in Strawson’s account).

298   R. Jay Wallace hold toward us. Specifically, they are responses to the “quality of will” that is reflected in the actions others take toward us, and they depend on whether our expectations for consideration or regard have been met. We resent it, for instance, when others treat us with hostility or contempt; these forms of ill will are the occasions of our resentment, and what it is properly about. So long as disrespectful attitudes of these kinds are present, it seems, resentment would appear to have a normative basis. But this conclusion has potentially significant consequences for the traditional debate about the conditions of responsibility. If a susceptibility to reactive attitudes is the paradigmatic way we hold people accountable for what they do, and if these attitudes in turn are warranted responses to actions that reflect hostility or disregard, then it starts to seem irrelevant to our responsibility practices whether those actions were the result of deterministic causal forces. That is, so long as the actions for which we blame agents are ones that fail to satisfy our expectations for consideration or regard, it does not matter what their causal history might have looked like. We can perhaps imagine a possible social world in which people never have occasion to blame each other for the things they do. As Strawson writes, however: “[t]‌he prevalence of this happy state of affairs would not be a consequence of the reign of universal determinism, but of the reign of universal goodwill” (80). We might refer to this as Strawson’s internal argument about the conditions of moral responsibility, since it focuses on the internal objects of the reactive attitudes, arguing that the traditional metaphysical threats to responsibility would not in fact deprive those attitudes of their ordinary warrant. This internal line of argument can be applied to each of the two Strawsonian interpretations of responsibility distinguished in section 2 above. According to the more relaxed interpretation, you blame someone for their actions when those actions impair their relationship to you and you adjust your own behavior toward them in response. Paradigmatic of moral blame, on this way of developing the Strawsonian approach, are reactions such as avoidance and withdrawal from trust-​based interactions with the person who has treated you badly (e.g., by showing disregard or hostility). The internal objects of these responses are the attitudes of other parties that impair their relationships to you, which give you reasons for the responses characteristic of moral blame. You have good reason to withdraw from trust-​based relationships with people who have treated you with contempt and acted with the intention of harming or humiliating you. These reasons, moreover, remain in place regardless of the metaphysical and causal constitution of the actions and attitudes in which they consist. The reactive account of responsibility, by contrast, interprets moral accountability in terms of the narrower class of reactive attitudes that are constitutively connected to normative expectations or demands. These angrier emotions include, centrally, the core cases of resentment, indignation, and guilt; as we have seen, they are essentially reactions to actions of others that violate our basic interpersonal demands for consideration or regard. Actions that fail to meet these expectations, on account of the quality of will reflected in them, are paradigmatically actions that can be understood to wrong the person at whom the ill will or disregard is directed. The constitutive thought of

Responsibility and Reactive Attitudes    299 resentment, for instance, is that one has been wronged, and behavior that wrongs another person in this way provides the targeted person with warrant for this reactive attitude, and hence for blame.13 Again, however, whether this normative basis for reactive blame is in place is completely independent of facts about the metaphysical character and causal history of the attitudes and actions that constitute the moral wrong. Given the presence in another agent of a negative quality of will that violates our expectations of minimal consideration, there is warrant for us to feel resentment, which is tantamount to blaming the agent for treating us as they have done.

4.  The Ethics of Reactive Blame There are circumstances, however, in which reactive blame seems potentially out of place, despite the fact that the person at whom it is targeted has acted with contempt of or disregard for another agent. Consider, for instance, young children, or those suffering from severe forms of trauma or mental illness, or the figure of the psychopath. People who suffer from these conditions might exhibit the qualities of will that ordinarily provide warrant for the reactive attitudes (including their angrier variants). But it is commonly felt to be problematic to blame them straightforwardly for their actions. Strawson himself distinguished between two different circumstances that can inhibit the reactions through which moral responsibility is characteristically attributed. There are localized factors that inhibit blame for an action by showing that the action did not reflect the kind of quality of will that warrants the reactive attitudes in the first place. And there are less localized factors that bear on the agent’s standing as someone toward whom reactive attitudes are aptly directed, even when the agent shows contempt or disregard or a general lack of consideration for other persons. These less localized factors seem to exempt individuals from entering into responsibility relations with others, despite the fact that they are capable of the actions and attitudes that typically render reactive blame appropriate.14 And if exemptions in this sense are coherent, it seems at least possible that they might generalize, so that it would always be problematic to enter into responsibility relations with others, regardless of whether they fall short of reasonable interpersonal expectations for consideration and regard. Some philosophers are skeptical about the whole category of exemptions from accountability.15 This form of skepticism can be understood as a development of the

13 

Compare Pamela Hieronymi, “The Force and Fairness of Blame,” Philosophical Perspectives 18 (2004), pp. 115–​48, also her “I’ll Bet You Think This Blame is About You,” in D. Justin Coates and Neal A. Tognazzini, eds., Oxford Studies in Agency and Responsibility 5 (2019), pp. 60–​87. 14  Watson distinguishes between excuses and exemptions in this connection, in “Responsibility and the Limits of Evil,” pp. 223–​225. 15  See Scanlon, Moral Dimensions, chap. 4; also his “Learning from Psychopaths,” in Coates and Tognazzini, eds., Oxford Studies in Agency and Responsibility 5, pp. 127–​141. See also Angela Smith,

300   R. Jay Wallace internal line of argument sketched at the end of the preceding section. What gives us reason for the reactions characteristic of blame are the attitudes that others take toward us, and these attitudes are unquestionably present in people, such as the psychopath, whose moral capacities might otherwise be impaired or underdeveloped. If someone in this condition sets out to harm or denigrate you, they will count as having wronged you and impaired their relationship to you.16 This is sufficient to give you reason for reactive emotions and for the broader adjustments in your behavior toward the wrongdoer that are characteristic of blame. (Children may constitute a special case, but not because they are subject to a condition that generally exempts them from responsibility reactions, but because we adults have special fiduciary and tutelary responsibilities to them that interact with, and potentially outweigh, our ordinary reasons to blame them for their lack of consideration and regard.) Against this, other philosophers continue to feel that the conditions mentioned earlier really do function to exempt people from ordinary accountability relations.17 The underlying thought here is that there is something misguided about treating psychopaths, the mentally ill, and young children like ordinary targets of reactive blame, even when they act in ways that otherwise warrant such reactions. As just noted, this is an idea to which Strawson was himself attracted, insofar as he acknowledged that there are conditions that exempt people from ordinary accountability relations. He treated these as conditions that modify our conception, not of the actions that agents perform, but of the agents themselves; they are conditions that “invite us to suspend our ordinary reactive attitudes” toward the agents to whom they apply (78), either temporarily (as in cases of stress or alcohol-​induced impairment), or more permanently (as with the psychopath or those subject to other forms of serious and continuing mental or emotional deficiency). But how exactly do these exempting conditions “invite” us to modify or suspend the reactive attitudes toward those who are subject to them? The suggestion, I take it, is not merely that reactive attitudes are inhibited, causally, by the presence of these conditions, but that their presence somehow renders reactive attitudes inappropriate or rationally questionable. There might well be, in any given case, some default warrant for attitudes

“Control, Responsibility, and Moral Assessment,” Philosophical Studies 138 (2008), pp. 367–​392, and Pamela Hieronymi, “Reflection and Responsibility,” Philosophy & Public Affairs 42 (2014), pp. 3–​41. 16  Though psychopaths can clearly act with hostility toward others, there is an interesting question whether they are capable of such attitudes as disrespect or contempt; on this point, see Dana Nelkin, “Psychopaths, Incorrigible Racists, and the Faces of Responsibility,” Ethics 125 (2015), pp. 357–​390. It is not entirely clear that one can really wrong another agent if one is not capable of showing something like disrespect for them, as bearers of moral claims; see my The Moral Nexus, pp. 153–​154. 17  See, for instance, my Responsibility and the Moral Sentiments, chap. 6. For defenses of the idea that psychopathy might exempt from reactive blame, see Watson, “Responsibility and the Limits of Evil,” and his “The Trouble with Psychopaths,” in Wallace, et. al., Reasons and Recognition, pp. 307–​331. (An important theme in Watson’s discussion of psychopathy and other potential exempting conditions is the idea that reactive blame is incipiently a form of moral address, and so appropriately directed only at agents who are capable of understanding the demands that are so addressed.)

Responsibility and Reactive Attitudes    301 such as resentment or indignation, insofar as the agent subject to the exempting conditions has treated another with contempt or ill will. But those conditions give us reason to suspend the reactive emotions altogether toward agents who are affected by them. This thought is perhaps most naturally applicable to the narrow interpretation of the reactive attitudes, which ties them constitutively to normative expectations or demands. In reference to these attitudes, the idea that the exemptions give us reason to suspend the reactive attitudes altogether, even when qualities of hostile will are displayed that would seem to warrant them, can be understood as the thought that there would be an objection to holding people to normative expectations or demands at all when they are subject to the exempting conditions. But what, more specifically, is the nature of this objection? The classic exempting conditions appear to involve temporary or more permanent impairments of our moral capacities, including our capacity to understand and apply basic moral expectations, and to comport ourselves in accordance with them. The idea is that, when these forms of incapacity affect people, it is somehow problematic to address moral demands or expectations to them. This idea, in turn, is naturally developed in moral terms. The intuitive objection to addressing moral demands to agents who are not able to grasp or comply with them is that it is morally unreasonable, or unfair, to do so. It would be wrong to target reactive attitudes toward people who lack the capacity to understand the moral requirements expressed in them, or to comply with them as they decide what to do; that is the way in which the exempting conditions function to undermine moral accountability.18 But this thought itself can be questioned. Considerations of fairness are normatively significant for us insofar as they provide us with reasons for action; the fact that it would be unfair to treat a person in a certain way is thus something that counts against that way of comporting oneself. But blame, according to the reactive account of it, is primarily a matter of our attitudes. The considerations that provide reasons either to blame agents or to withhold blame from them are considerations that warrant such reactive attitudes as resentment and indignation. These reasons in turn, as we have seen, are naturally understood in terms of the quality of will that agents show toward others, including qualities such as contempt and hostility and disregard, which (as noted above) are sometimes displayed even by those who are subject to the alleged exempting conditions. To appeal to fairness in this context, as a consideration that might warrant the suspension of blame, for example, toward psychopaths and young children, would therefore be to cite the wrong kind of reason; it would represent an attempt to mobilize reasons for action in the critical assessment of attitudes that are responsive to normative reasons of a completely different kind.19 To fend off this skepticism about the exempting conditions would require a story about why those conditions give us reason to modify our reactions to the individuals 18 

See my Responsibility and the Moral Sentiments, chap. 6. See (again) Hieronymi, “The Force and Fairness of Blame” and “I’ll Bet You Think This Blame is About You.” 19 

302   R. Jay Wallace who suffer from them. One way to develop such a story, consistently with the reactive approach to responsibility, would start by noting that it is open to us to manage the emotional attitudes we are subject to in various ways. Even if angry disapprobation might be warranted in virtue of the attitudes of another person to us, for instance, we could resolve to disavow or overcome it, treating it as an insignificant element in our ongoing relations with the person who is its target. Adopting this essentially managerial stance toward ourselves is a matter of action, involving decisions about how we are to treat our own emotions that are properly responsive to practical rather than attitudinal reasons. These practical reasons potentially include the moral considerations of fairness or reasonableness sketched above. Thus, if it seems unfair to expose those individuals to the reactive attitudes who lack the capacity to grasp and comply with the demands implicit in them, this might be a compelling reason for overcoming these emotions, or setting them to the side interpersonally to the extent they resist our efforts to overcome them. This strikes me as a promising strategy for making sense of the normative significance of exempting conditions for our practices of holding people accountable for what they do.20 The distinction between reasons for reactive attitudes and reasons for managing them in certain ways can help us to make sense of other questions about the ethics of blame. Forgiveness, for instance, has long seemed puzzling to philosophers, insofar as it involves a double attitude toward the reactive blame that a person might be subject to in a situation in which they have been wronged.21 On the one hand, blame seems warranted under these circumstances, and it is understood as such by the victim of wrongdoing (otherwise there would be nothing to forgive); and yet, on the other, that person somehow moves beyond blame in offering forgiveness. A clue to what is happening here is perhaps provided by the language we typically use to discuss forgiveness, which is shot through with verbs of agency: In forgiving another person, we “forswear” the resentment to which we are entitled, or relinquish our entitlement to it, somehow. These moves could be understood as exercises of our managerial capacities, whereby we resolve to overcome reactive attitudes in our interpersonal relations with other individuals, while acknowledging that circumstances remain in place that would make them fitting or apt. Something similar might be going on in at least one class of cases that have the structure of hypocrisy, where individual A blames B for doing things to A that A herself routinely does to others. If B has in fact acted wrongfully toward A, then some reactive blame on B’s part might well be warranted. But it would perhaps be morally problematic for A to accede to this reactive blame, and to accord it significance in her relations with B, so long as she fails to acknowledge that her own conduct has been objectionable in the same way, to offer apologies for her wrongdoing, and to take steps to make amends. 20  The line of thought sketched in this paragraph is developed in my “Trust, Anger, Resentment, Forgiveness: On Blame and Its Reasons,” in European Journal of Philosophy 27 (2019), pp. 537–​551. 21  See, for example, Pamela Hieronymi, “Articulating and Uncompromising Forgiveness,” Philosophy and Phenomenological Research 62 (2001), pp. 529–​555.

Responsibility and Reactive Attitudes    303 A’s lack of standing to blame B, in other words, could be understood in terms of a moral objection to A’s treating her own reactive attitudes as interpersonally significant, which may obtain consistently with the fact that B’s behavior provides some warrant for those same attitudes.22 Once these possibilities for managerial control of reactive blame are acknowledged, more extensive critiques of blame begin to make sense. Strawson’s “pessimist,” for instance, can be understood as making the case that reactive blame, even when it is warranted by the attitudes of the agents at which it is targeted, is not something that we should ever fully accede to or treat as interpersonally significant. To manage blame in these affirming ways is to become complicit in a system of social pressures that it is ultimately unreasonable or unfair to subject people to in a world in which their behavior is part of the deterministic order of nature. The incompatibilist’s concern for the fairness of blaming those who lack ultimate freedom of will might thus be brought to bear on the managerial stance we adopt toward warranted resentment when we affirm it rather than overcoming it or trying to set it aside. I do not myself find the pessimistic critique of blame fully persuasive, at the end of the day. But it articulates a challenge that needs to be taken seriously and assessed on its own terms. It cannot be dismissed solely on grounds that it appeals to moral reasons that are of the wrong kind for criticizing the reactive attitudes in which blame plausibly consists.

22 

See my “Hypocrisy, Moral Address, and the Equal Standing of Persons,” Philosophy & Public Affairs 38 (2010), pp. 307–​341.

CHAPTER 15

Resp onse-​D e pe nde nt Theori e s of Resp onsi bi l i t y David Shoemaker

1.  A Brief History of Response-​Dependence While the label “response-​dependent theory” did not appear in the philosophical literature until 1989 (from Mark Johnston, in Smith, Lewis, and Johnston 1989), some response-​dependent-​like theories did exist much earlier than that. These were advanced by some of the original crew of moral sentimentalists: Locke, Hume, Hutcheson, and Smith (see Kauppinen 2017 for discussion). Their basic thought was that one cannot understand certain features of our moral lives—​virtue, vice, rightness, and wrongness—​ without reference to our moral sentiments. Of these, Hume and Smith put forward theories that most clearly resemble response-​ dependent theories of responsibility, although it is not entirely clear at the end of the day just how close the resemblance is. For Hume, holding someone responsible for an action was a matter of feeling—​of sentiment—​not judgment, and such feelings would be psychologically impossible did they not trace the action to that agent’s character. The things that arouse our moral sentiments “become the objects of praise and blame” (Russell 1995: 66). Reason cannot justify—​and indeed is irrelevant to—​these responsibility responses; they are instead utterly natural human responses. And so we simply cannot understand the nature of responsibility without essential reference to the role played by our moral sentiments (Russell 1995: 81). Nevertheless, as Hume notes, even though the sentiments naturally arise in us, “[I]‌n order to pave the way for [them], and give [them] proper discernment . . ., it is

Response-Dependent Theories of Responsibility    305 often necessary, we find, that much reasoning should precede, that nice distinctions be made, just conclusions drawn, distant comparisons formed, complicated relations examined, and general facts fixed and ascertained” (Hume 1983: 15). This is because we cannot, on pain of absurdity, make praiseworthiness and blameworthiness a function of the actual sentiments of praise and blame people happen to have. After all, children and the insane are not responsible, even if some of us respond to them with praising and blaming sentiments. What we need, then, is an account of “proper sentiment,” capable of being “corrected by argument and reflection” (Hume 1983: 15; emphasis mine). Unfortunately, as Paul Russell notes, “Hume provides us with no adequate account of the nature of moral capacities required for a person (or creature) to be deemed an appropriate object of moral sentiment” (Russell 1995: 91). Furthermore, if we were able to articulate the reason-​based conditions for the development of various sentimental responses, it is unclear the extent to which we would still have a response-​dependent understanding of responsibility on our hands, as it could be that the “argument and reflection” giving rise to the discernment necessary for proper sentiment is focused on the response-​independent nature of responsibility, that is, the objective qualities of agents that truly make them praiseworthy or blameworthy. So while people naturally have sentimental responsibility-​responses to one another, argument and reflection may focus their attention on the actual responsibility-​making qualities of agency, and their responses may thus be made correct, or “proper,” by the presence of those objective qualities. In other words, it may be that, while we cannot understand what responsible agency is—​it’s just the thing to which we have a range of sentimental praising and blaming responses—​without essential reference to our holding one another responsible, the propriety conditions of our holding people responsible could still be a function of more metaphysically basic response-​ independent properties.1 Consequently, it is just unclear how response-​dependent Hume’s view really is. Adam Smith’s view is superficially similar to Hume’s, although Smith focuses explicitly on the types of responsibility responses that will resonate with many contemporary theorists—​resentment and gratitude—​and he also fills in the sorts of crucial details about their proper deployment that Hume elides. Gratitude is the sentiment most immediately and directly prompting us to reward someone, according to Smith, whereas resentment is the sentiment most immediately and directly prompting us to punish someone (Smith, TMS: II.1.1). The proper objects of these sentiments will appear to us as meriting, or deserving, these responses. But what are the proper objects of these sentiments? These are the details Hume left out. What Smith says in response

1  As we will see later, the real debate between response-​dependent and response-​independent theorists has to take place at the level of the propriety-​makers for proper responses. As Hume does not dig as deeply as that, I have to reserve judgment on whether he counts as a response-​dependent theorist.

306   David Shoemaker is that “[a]‌ll it can mean to say” that someone is the proper object of one of these sentiments is that “he is an object of gratitude (or resentment) that naturally seems proper and is approved of ” (TMS: II.1.2). And what does this mean? That “the heart of every impartial spectator entirely sympathizes with the passion in question, i.e., that every unbiased bystander entirely enters into the passion and goes along with it” (TMS: II.1.2).2 This seems a pretty thoroughly response-​dependent view: Responsibility is a function of our practices of holding people responsible—​paradigmatically via the sentiments of resentment and gratitude (Hankins 2016: 741)—​and what makes such responses proper is just that they are the sentiments that would be had by any unbiased observer (including oneself).3 So someone is made blameworthy for an action, say, just in case it would be proper for the offended party to respond to her with resentment, and where any reasonable bystander would sympathize with that resentful response. Interestingly, this could allow for a rather permissive account of responsibility, as the targets of people’s proper sentimental responsibility-​responses may not necessarily be restricted to what are often taken to be the necessary agential expressions of intentions or quality of will (see Hankins 2016 for a nice discussion).4 Nevertheless, one might worry that this view would be too permissive. After all, what if everyone—​including all reasonable bystanders—​were to resent those with profound intellectual disabilities, or they didn’t include as the proper objects of gratitude or resentment women and certain minorities (Fischer and Ravizza 1993: 18)? In this case, it would seem, Smith’s view yields the wrong answer.5 It is thus not clear how we might appeal to people’s actual responses to determine the conditions of responsibility without always running into this sort of problem (see also Todd 2016 for a nice presentation of this problem).

2 

While Smith’s talk of an impartial spectator might conjure up thoughts of an “ideal observer,” that is not what he means. Rather, the term simply refers to “an internal mechanism for checking the partiality of our judgments” (Hankins 2016: 741). Thanks to Keith Hankins for discussion on this point. 3  Although one might doubt the ultimate response-​dependence of the account given some passages such as the following: “[N]‌ature has in this way stamped on the human heart, in the strongest and most indelible characters, an immediate and instinctive approval of the sacred and necessary law of retaliation, this being something that comes into play before any thoughts about the utility of punishment” (TMS: II.1.2). This sounds like something a realist/​response-​independent theorist would be very happy with. (Thanks to David Brink for the pointer and comment.) 4  Of course, as both Dana Nelkin and Derk Pereboom rightly note, it could also issue in a fairly restrictive account, depending on how “impartiality” is cashed out. 5  Although as Keith Hankins has pointed out to me (in discussion), Smith may have a way to sidestep this worry, as he distinguishes (in Part 6 of TMS) between perfect propriety and approximation propriety (the degree to which our sentiments approximate perfect propriety), and so provides a way to correct widespread sentimental agreement. This is actually akin to the view I adopt later in the paper. Such overly wide permissibility could also be ruled out, of course, by adopting Hume’s view of the impartiality, which adopts “the steady and general point of view.” (Thanks to Derk Pereboom for this point.)

Response-Dependent Theories of Responsibility    307

2.  Strawson and the Response-​Dependent Turn The contemporary template for response-​ dependent theories of responsibility is found in P. F. Strawson’s “Freedom and Resentment” (Strawson 1962/​2003). But while this essay has purportedly influenced generations of responsibility theorists, Gary Watson correctly notes that never have so many claimed to be Strawsonians without actually embracing Strawson’s “fundamental program” (Watson 2014: 15), namely, its response-​dependent understanding of being responsible. What most contemporary “Strawsonians” embrace instead is just the general strategy of appealing to the reactive attitudes as revealing something interesting about the nature of responsibility. But anyone—​even libertarians—​can embrace that strategy (Watson 2014: 16). After all, the reactive attitudes can be perfectly “informative” about the nature of responsibility even if they are merely our epistemic guides to it, that is, what most reliably enables us to track the true response-​independent conditions of responsibility. What, then, is his response-​ dependent story? He begins by speaking in “commonplaces” about us. First, and most centrally, he emphasizes “the very great importance that we attach to the attitudes and intentions towards us of other human beings, and the great extent to which our personal feelings and reactions depend upon, or involve, our beliefs about these attitudes and intentions” (Strawson 1962/​2003: 75). Manifesting this “basic concern” is the “basic demand” (Watson 2014: 17), our demand for “some degree of goodwill or regard on the part of those who stand in these [human interpersonal] relationships to us . . .” (Strawson 1962/​2003: 76). When those whose attitudes toward us matter injure us or others we care about, the emotional dispositions in which our caring consists tend to be triggered, and we respond with a range of reactive attitudes, including resentment, indignation, and hurt feelings. These attitudes are plausibly construed as expressions of the basic demand (Watson 2004: 225). It is only with the emergence of the basic demand that we get the emergence of responsibility, as it is only “where the demand is felt to be disregarded” (Strawson 1962/​ 2003: 90) that our responsibility responses—​the reactive attitudes—​are naturally triggered. What tends to suspend these responses in ordinary circumstances are cases in which the injuring agent—​despite initial appearances—​wasn’t actually expressing the poor quality of will that is the subject of the basic concern. These suspensions are typically generated by either pleas of excuse (“It was an accident!”) or pleas of justification (“I had no other choice!”). What tends to suspend these responses to agents more generally are extraordinary circumstances, wherein the agent is “incapacitated in some or all respects for ordinary inter-​personal relationships” (Strawson 1962/​2003: 82; emphasis in original). Indeed, this story about our natural suspensions of the reactive attitudes is what delivers their fittingness conditions. We have a reason of fit to regard agents as responsible (i.e., to be prone to hold them responsible via the reactive attitudes) just in case they are fellow members of the moral community, that is, capacitated with respect

308   David Shoemaker to interpersonal relationships (Strawson 1962/​2003: 90).6 And we have a reason of fit to respond to fellow members of the moral community with one or more of the reactive attitudes to their various expressions of quality of will, for example, gratitude in response to their expressions of excellent quality of will, and resentment or indignation in response to their expressions of poor (or insufficiently good) quality of will. So what precisely is the response-​dependent aspect of the theory and where does it come from? Holding responsible, for Strawson, consists most fundamentally in experiencing and/​or expressing the reactive attitudes in accord with various practices (or at least viewing doing so as fitting). To be prone to doing so just is to make the basic demand of one’s moral fellows (Strawson 1962/​2003: 90). But the basic demand is a function of the basic concern, and the basic concern is just a function of our ground-​level non-​rational human sentiments: Given the way we humans have been constructed, the attitudes of our moral fellows toward us simply matter to us, which translates into our being disposed to respond emotionally (with the reactive attitudes) when this concern is betrayed. The facts of responsible agency are facts about agents’ (non-​)compliance with the basic demand. But then this means that the facts of responsibility—​of being responsible—​are ultimately beholden to and a function of the facts about the basic concern and the holding-​responsible-​responses to which the basic concern disposes us. Responsible agency “is to be understood by the practice [of holding people responsible], which itself is not a matter of holding some propositions to be true, but of expressing our concerns and demands about our treatment of one another . . ., stances and responses [which] are expressions of certain rudimentary needs and aversions” (Watson 2004: 222).7 Responsible agency, in other words, can only be understood as response-​ dependent, a function of our human reactive attitudes. The reason ignoring the response-​dependent feature of Strawson’s view is problematic is that without it one simply cannot make sense of his arguments about the (ir) relevance of determinism. Strawson’s aim was to dissolve the standoff between (consequentialist) optimists and (libertarian) pessimists regarding the grounds of our responsibility practices. Both appealed to justifications external to the practices of holding people responsible. The former took the practices to be justified by the facts as we know them, albeit exclusively external forward-​looking impersonal facts about social efficacy and regulation. The latter took them to be justified by facts about desert and libertarian agency, which, while more personal and backward-​looking (and so more faithful to the 6 

This is merely a pro tanto reason; it might be outweighed by other types of reasons (e.g., prudential, moral, aesthetic) in a determination of what one ought, all-​things-​considered, to feel in response to an agent. 7  Technically speaking, there are surely some aspects of our holding-​responsible practices that are a matter of “holding some propositions to be true,” such as, for instance, propositions that some excuse or exemption obtains. (Thanks to Derk Pereboom for this point.) A more charitable way to read this view is by adopting the machinery I am about to suggest in the text, namely, locating Strawson’s response-​ dependence not at the level of individual determinations of responsible agency but rather at the level of fitness-​makers for responsible agency, that is, at the level of making true the relevant propositions about responsibility.

Response-Dependent Theories of Responsibility    309 spirit of our practices), nevertheless went beyond the facts as we know them into metaphysical flights of fancy. Strawson’s (dis)solution was to account for all we mean by “responsibility” without going beyond the facts as we know them (the optimist’s insight), but to remain faithful in so doing to the spirit of our practices, which are largely personal and backward-​looking (the pessimist’s insight). The key to this move is Strawson’s establishing that there is no threat to our practices from external considerations, from demands for justification for the practices as a whole from outside of the framework of those practices. But any theory of our holding-​ responsible-​practices is going to be vulnerable to demands for external justification unless the theorist takes responsibility itself to be ultimately and exclusively response-​ dependent. That’s because response-​independent views of responsibility have our holding-​responsible-​practices ultimately responding to (and so presupposing) the antecedent fact that their targets are responsible, and so the justification for the practices will always have to be grounded in some external fact(s) about agential responsibility. Of course, even if one goes response-​dependent about being responsible, it remains possible that the entire system of reactive attitudes might still have or require an external justification (e.g., consequentialist or contractualist justifications, say), and so it would still be vulnerable to wholesale replacement if warranted. To eliminate such a possibility, Strawson offers two “inescapable framework” arguments (Watson 2014: 21). First, because the relation between holding and being responsible is one of “fittingness”—​adhering to “distinctive norms for attitudes of that kind” (Darwall 2006: 66), all demands for external justification of the practices would demand “wrong kinds of reasons” (this is what Watson [2014: 21] calls the “normative argument”), that is, reasons that don’t bear on the value being appraised by resentment, say. Second, if “ought implies can,” then there are no external justifying reasons relevant to our situation, given that we can’t give up the attitudes and practices manifesting our basic concern.8 They are simply a “given with the fact of human society” (Strawson 1962/​2003: 91). Consequently, the “existence of the general framework of attitudes itself . . . neither calls for, nor permits, an external ‘rational’ justification” (Strawson 1962/​2003: 91). If he is right, then the rest of his argument flows easily: Given that all the “right kinds of reasons”—​reasons of fit—​for suspension of reactive attitudes have neither the content (appealing to causal antecedents) nor the form (as universally applicable) of determinism-​based reasons, those latter sorts of reasons are just irrelevant to the nature and possibility of responsibility. In other words, the “problem” of determinism is just not a problem. This allows Strawson to be an optimist, albeit one that takes seriously the relationships and backward-​looking features of responsibility attributions that the consequentialist optimist ignores.

8  We could give up or alter individual types of responses, but we could not give up the reactive attitudes as a whole, for without them, we wouldn’t have anything “we could find intelligible as a system of human relationships, as human society” (Strawson 1962/​2003: 93).

310   David Shoemaker

3.  Moving Beyond Strawson It’s crucial to point out that Strawson doesn’t really give any argument for response-​ dependence about responsibility. Rather, he merely offers us a series of descriptive claims about human nature that are along the lines of “we fundamentally care about goodwill,” or “these attitudes form the very background framework of our lives,” or “interpersonal relationships are all constituted by vulnerability to the reactive attitudes,” or “we simply can’t give these attitudes up,” or “none of the reasons for excuse internal to our practices are of the deterministic form/​content.” Now if we grant him these and other airy empirical claims, then his response-​dependence does seem to follow. But there are plenty of reasons to doubt at least some of them. First, several people have pointed out that among the excusing reasons Strawson gives are plausibly deterministic reasons like “he had to do it,” or “there was no alternative.” These, it has been thought, might well be grounded in principles (e.g., of fairness) that have universal application akin to the reasons of determinism (see the “generalization strategy” described by Wallace 1994: 114–​116). Others have pointed out that built into the notion of “quality of will” may be a notion of “control,” such that we don’t attribute poor quality of will to someone who had no control over her actions. If so, then questions about quality of will carry over into questions about free will after all (see, e.g., McKenna 2012: 61, n. 3). Second, the psychological inescapability argument is dubitable. Indeed, successful Stoics and Buddhists (granting that they exist), as well as perhaps some free will skeptics, have done precisely what Strawson claims we can’t do, overcoming or eliminating their set of reactive attitudes as grounded in some sort of prudential or metaphysical mistake. Furthermore, Strawson himself allows that we can suspend the participant framework to get relief, occasionally, from the stresses of life and relationships (Strawson 1962/​2003: 79–​80), so why think it couldn’t be done in a more sustained fashion? Third, we may well have good reason to do so. Thomas Nagel argues that, actually, it’s quite natural for us to take up a kind of Strawsonian objective standpoint: It occurs when we think about ourselves and our actions as part of a long causal chain, leading straight back to our heredity and early educational environment. And when we take up this perspective, the purely subjective participant standpoint comes to seem illusory. At the very least, we are left with a kind of irresolvable ambivalence about reasons for preferring either stance (Nagel 1986: 110–​137; Watson 2014: 24–​25). Fourth, one might doubt that we most fundamentally care about goodwill, or goodwill from all. Many us do, certainly. But some of us have a very limited range of people whose attitudes toward them matter (La Famiglia), and still others—​psychopaths and bad neighbors—​seem not to care at all about goodwill in any direction. Perhaps, more importantly, the “we” here is indeed quite circumscribed, as what matters to many in the world may be just the avoidance of shame, which isn’t about goodwill but rather about living up to various kinds of ideals of honor or other communal norms (see, e.g.,

Response-Dependent Theories of Responsibility    311 Graham, Haidt, and Nosek 2009). That is to say, the Strawsonian project looks too focused on what Bernard Williams called “the morality system,” in which mutual regard gives rise to rights, duties, justice, and guilt (Williams 1985: ch. 1). But there are and have been plenty of shame-​based cultures whose members may not be seen to share in the Strawsonian “basic concern” (Russell 2017: ch. 6). Finally, the claim that all interpersonal relationships are shot through with vulnerability to the reactive attitudes seems false. As the case is often put, Gandhi and King held their persecutors to account without the retributive attitudes of resentment or indignation (Stern 1974; Watson 2004: 257–​258). Many clinical psychologists attempt to do the same with the antisocial prisoners they treat (see Pickard and Ward 2011; Pickard 2014). And one might certainly construct relationships grounded in love, where resentment, say, has no place (Pereboom 2014: 178–​186; Wallace 2014). For my part, I think all of these objections are rebuttable. But they remain potent to many theorists, and the reason is, again, simple: Strawson gives no argument for the claims that give rise to his response-​dependent theory of responsibility. Consequently, the position has come to seem as if it begs all the important questions at issue. My aim in what follows is to offer an argument on Strawson’s behalf.

4.  Amusement and the Funny The argument draws from a surprisingly tight analogy between the blameworthy and the funny.9 Start by considering the following strange collection of events: slapstick, wordplay, satire, clever insults, certain sounds (e.g., farts), kids laughing giddily (e.g., at farts), stupidity, pratfalls, and more. They are all things we tend to judge as funny and/​ or respond to with amusement. But how should we characterize the relation between what we judge funny and our amused responses? Call this the characterization question. Further, what precisely unifies these events (on what I will call the “funny list”) as funny and/​or the object of our amusement? Call this the unity question. Start with the characterization question. There are three ways of answering it. First, we might characterize the relation in crudely response-​dependent terms, as follows: X is funny if and only if, and in virtue of the fact that, people are disposed to respond to X with amusement in standard conditions.

Call this Dispositional Response-​Dependence. It is quite implausible, for three reasons. First, what exactly are the standard conditions for amusement? It is very hard to specify what these might be (cf. D’Arms and Jacobson 2006: 201). Second, this analysis just doesn’t at all reflect how we think and talk with one another about the funny. If I don’t

9 

In some of what follows, I draw from Shoemaker 2017.

312   David Shoemaker find the comedian Carrot Top funny, it would be quite strange for you to attempt to convince me otherwise by saying, “But lots of people find him funny!” That’s not the type of reason relevant to engaging my comic sensibilities. Third, and most importantly, we think that lots of people’s amused responses can be wrong. That is to say, we have a normative view of the funny: Some things merit amusement, regardless of people’s actual dispositions to be or not be amused. The dispositional analysis fails to incorporate that normative feature. Dispositional Response-​Dependence gives a bad answer to the characterization question, so there is no need to see how it might address the unity question. How, then, might we capture the normative aspect of the funny via the types of normative reasons that do engage our comic sensibilities? Many theorists have thought we can best do so with a Normative Response-​Independent characterization of the funny, as follows: X merits amusement if and only if, and in virtue of the fact that, it is funny.

Here the funniness is metaphysically prior to the amusement, that is, the fact that something is funny—​independently of our responses—​is what in turn makes amusement merited (or fitting or apt—​I will use these terms interchangeably). With this answer to the characterization question in hand, we can move on to the unity question, for it should be obvious that the adherent of this view now owes us an account of the specific response-​independent properties that make something funny (and so make our judgments of the funny true and our amused responses to the funny merited). There have been many such theories over the years. Most recently, the leading contenders have been the Incongruity Theory (Clark 1987) and the Benign Norm Violation theory (McGraw and Williams 2010). As you would expect, the former identifies the response-​independent property making things funny as a certain kind of incongruity, and the latter identifies that property as a benign norm violation. The problem with both, however, is that they are rife with counterexamples. While some incongruous events are funny, not all are, as illustrated by a Salvador Dali painting. And not all funny things are incongruous, as illustrated by how funny it is to watch a couple of six-​year-​old boys giddy with laughter at hearing a fart sound (that’s just what six-​year-​ old boys do; there’s nothing incongruous about it). While some benign norm violations are funny, not all are, as illustrated by forgetting to floss your teeth at night or wearing white after Labor Day. And not all funny things are benign norm violations, as illustrated by Laurel and Hardy losing their grip on a piano halfway up a set of stairs (cf. Olen 2016). Any response-​independent theory of humor will run into such counterexamples (D’Arms and Jacobson 2006). Of course, one might instead offer a long disjunctive theory to unify the disjointed list of things we find amusing, for example, “What makes something funny is its having properties P, Q, R, . . . and/​or Z.” But a long disjunctive theory fails to be explanatory: We want to know what explains the unity of the list, and a disjunctive theory basically just repeats the list without illumination. Answering the characterization question with a response-​independent theory leaves us without a good answer to the unity question.

Response-Dependent Theories of Responsibility    313 The third answer to the characterization question, then, is the Normative Response-​ Dependent theory of the funny: X is funny if and only if, and in virtue of the fact that, it merits amusement.

What makes the events on the funny list funny is simply that they are all things that merit amusement. On this view, amusement is more metaphysically basic than the funny, as it is the former—​in its merited form—​that determines the contours of the latter. As with the Response-​Independent theorist, the Normative Response-​Dependent theorist owes us an answer to the unity question, albeit now pressed in terms of merit: What do all of the items on the funny list have in common such that they all merit amusement? There must be properties in place that generate reasons for amusement, call them comic properties. When someone tells a joke, for instance, reasons for amusement at it depend on things like its tone, content, timing, execution, and context (and their careful admixture). But that fact is perfectly compatible with, and actually seems to reintroduce, a response-​independent characterization of the funny, as those comic properties might well seem to be the objective properties in virtue of which amusement is merited. Genuine response-​dependence about the funny, therefore, must answer a deeper inquiry: “What makes those objective properties (e.g., tone, content, etc.) the fitting-​ amusement-​makers and not others?” And the answer is this: They are the properties to which refined human comic sensibilities would respond with amusement. Sometimes incongruity is funny and sometimes it’s not, and the same is true of benign norm violations. But why are those properties the fitting amusement-​makers, when they are? It’s because they have been combined with other properties (e.g., of context, tone, etc.) that, together, simply are the properties to which refined comic sensibilities would respond with amusement, other things equal. That’s it. There are various general objections to this move, but I will save them for the analogue discussion about responsibility (for discussion, see Shoemaker 2017: 490–​493). For now, consider this at least to be a plausible burden-​shifting argument in favor of response-​ dependence about the funny. What I will now show is that this burden-​shifting move can be made just as plausibly in a significant subdomain of responsibility, given a very tight analogy to the funny.

5.  Responsibility, Blameworthiness, and Anger Unlike with the funny, where the only relevant response is amusement, there are many responsibility responses, including admiration, disdain, pride, regret, contempt, approval, disapproval, gratitude, hurt feelings, forgiveness, and more (see Shoemaker 2015 for discussion). So what should the object of our response-​dependent investigations be?

314   David Shoemaker Many responsibility theorists are in it to explore the nature and justification of blame (see McKenna 2012 and Pereboom 2014), so we are on well-​trodden ground by starting with a blaming emotion. In the blame domain, theorists zero in on resentment as the paradigm blaming response. However, we cannot do the same without begging the very question at issue. To explain, resentment is almost universally taken to be what D’Arms and Jacobson call a “cognitively sharpened” emotion, namely, anger plus a judgment, for example, that the to-​be-​resented agent culpably wronged you (D’Arms and Jacobson 2003: 143; see also Wallace 1994: 245; Darwall 206: 67; McKenna 2012: 66; and Sommers 2012: 176). But if that’s the correct characterization of our paradigm responsibility emotion, then the game has been given away to the response-​independent theorist, for resentment presupposes the responsibility of the resented agent. If you deliberately step on my foot, and my resentment includes the judgment that you culpably wronged me, then what makes my response apt is just that that constitutive judgment is true, and your judgment will be rendered true by your antecedent responsible blameworthiness, as that’s just what a judgment of culpable wronging amounts to. Cognitive theories of blame beg the question in favor of response-​independence. To be fair to the possibility of response-​dependent responsibility, then, we need a blaming emotional response that is neutral between various approaches, so I propose that we investigate resentment’s emotional core, anger, at least as it is a response to certain types of agential activity.10 So what does this response consist in? Here I will simply stipulate a characterization of blaming anger that I have defended at length elsewhere (Shoemaker 2015: ch. 3; Shoemaker 2017, 2018). Blaming anger’s emotional syndrome—​its various identifying features—​includes a certain affect, namely feelings of heat and aggression, a certain appraisal, namely an evaluation of someone’s agential activity as a slight, and a certain action tendency, namely confrontation (of the slighter).11 Confrontation may include retaliation, but not always; there are plenty of cases in which we want to communicate our anger to someone with no trace whatsoever of a retaliatory impulse, for example, writing an angry letter to our senator, or simply turning off our phone in response to a series of drunk and offensive texts by an ex-​lover. Let us then consider what our response would be to the following types of actions or attitudes: a punch in the face; a broken promise; condescension; ridicule; trespassing; peeping; hacked emails; theft; amusement at our daughter’s bone-​breaking fall; a drunk 10 

There are interesting complications involved in anger, as we also get angry at non-​agential things. I discuss multiple types of anger and why agential anger is just one of them in Shoemaker 2015: ch. 3, and in Shoemaker 2018. It is also crucial for me to note here, that while I will be talking about “blaming anger” and “the blameworthy,” I fully appreciate that there are other forms of blaming that don’t involve anger at all, such as disappointment or disdain. I did not make this point sufficiently clear in Shoemaker 2017, but as just noted, I’m simply talking about a paradigm form of blame and blameworthiness, the form that centrally consists in anger at agents. In the next section, I explicitly discuss whether response-​ dependent approach might also apply to other sorts of responsibility and blaming responses. 11  For introduction and defense of the motivational theory of emotions, see Frijda 1986 and 2007. For illuminating philosophical discussion, see Scarantino 2014, and D’Arms and Jacobson forthcoming.

Response-Dependent Theories of Responsibility    315 driver hitting our dog; a worker tossing slate off a building as we pass below. These are all events to which we are quite likely to respond with blaming anger. But they are also things we judge people to be blameworthy for, that is, responsible for. So again we have to address the characterization and unity questions: (1) How should we characterize the relation between our judgments of blameworthiness and our blaming anger? and (2) What precisely unifies these events (on what I will call the “blameworthy list”) as blameworthy and/​or the object of our blaming anger? In answering the first question, we have the same three options as we had with respect to the funny. The first, Dispositional Response-​Dependence, is a view that many have attributed to Strawson (see, e.g., Bennett 1980; Fischer and Ravizza 1998; and Todd 2016): X is blameworthy (and thus responsible) for some action or attitude A if and only if, and in virtue of the fact that, people are disposed to respond to X with blaming anger for A in certain standard conditions.

This view is implausible for the same three reasons it was implausible with respect to the funny: (1) It is entirely obscure what the “standard conditions” could be; (2) We don’t at all offer this type of reason to one another for our anger, for example, “I’m angry at you because most people would be”; and (3) We think some angry responses are just wrong, so we are committed to a normative, not descriptive, understanding of blaming anger. Indeed, this normative commitment is what generates the popular objection to what people take to be Strawsonian response-​dependence cited at the beginning: Even if a community refused to respond with resentment to women or minorities, or if they did respond with resentment to those with intellectual disabilities, it wouldn’t make the former non-​responsible and the latter responsible. True enough. A plausible response-​ dependent theory must therefore capture the normativity needed to respond to that objection, that is, the relevant responses must be merited. Insofar as Dispositional Response-​Dependence is missing normativity, it gives an implausible answer to the characterization question, which means there is no need to investigate how it might respond to the unity question. The second and third answers to the characterization question seek to explicate the relation between the blameworthy and the angerworthy, that is, what merits blaming anger. The most familiar answer is Normative Response-​Independence: X merits blaming anger for some action or attitude A if and only if, and in virtue of the fact that, X is antecedently blameworthy (and thus responsible) for A.

When you deliberately injure me, you manifest certain properties, properties that make you blameworthy for injuring me. On Normative Response-​Independence, that higher-​ order property—​blameworthiness—​is what renders my blaming anger merited. Advocates of this view—​ and these include most theorists of responsibility nowadays—​thus next owe us an answer to the unity question, providing an account of

316   David Shoemaker the response-​independent property or properties common to the events listed above that make(s) them all count as blameworthy. But just as with response-​independent theories of the funny, every theory of the blameworthy, I contend, is vulnerable to counterexamples. Instead of illustrating this point in terms of individual theories of responsibility, though, I will briefly articulate the blameworthy-​making properties highlighted and shared by many such theories, and I will show that for each of them there are cases that (prima facie) escape their net. For many theorists working on responsibility, there is a triumvirate of action properties that, when those actions are bad or wrong, are individually necessary and jointly sufficient to make people blameworthy for them. These properties are voluntariness, control, and knowledge. The first property is about the agent’s actual will, about whether she did what she wanted to do. The second property is in some respect modal, and is about whether she could have done, been, or responded differently. The third property is about the agent’s epistemic access to various relevant reasons, and is about whether the agent knew what she was doing or knew that what she was doing was wrong.12 Individually, though, none of these properties seems necessary, and together they may be insufficient, for blameworthiness. Consider first voluntariness. Many theorists have come to agree that people may be responsible—​and blameworthy—​not just for their actions but also for many of their attitudes (Adams 1985; Scanlon 1998; Smith 2005; McKenna 2012; Shoemaker 2015: chs. 1–​2; Portmore 2016; Pereboom 2017). The racist who embraces his own white superiority is, we may think, blameworthy for his racist attitudes, even if he never acts on them. So too the person who is amused by moral tragedies may be blameworthy, as may be the inattentive or uncaring spouse. But attitudes such as these are just not a function of the voluntary. Amusement at the horrific is not willed. There is an obvious sense in which these examples of non-​voluntary blameworthiness undercut the necessity of the property of control for blameworthiness as well. After all, if something is not governed by your will, then you lack a certain type of control over it. But one might want to maintain the importance of control for responsibility, and so introduce a different type of control over non-​voluntary attitudes, namely, rational or evaluative control (Smith 2005: 265; McKenna 2012: 194–​195; McHugh 2012; Portmore 2016: 4). This is the control we have when our attitudes are sensitive to our judgments, that is, our evaluations about the worth of the reasons for having them (Scanlon 1998: 18–​22; Smith 2005). In the ideally rational, judging that racism and racist attitudes are wrong would eliminate them. Of course, we are imperfectly rational, and sometimes these attitudes aren’t extinguished. But sometimes they aren’t extinguished in virtue of the fact that we still judge them worth having (e.g., Smith 2004: 341–​342). Consequently, they may still reflect our evaluative stance, and so we may be answerable—​blameworthy

12 

There are too many theorists to cite. Just about every theorist of responsibility has adopted one or more of these conditions. For a few representative citations, see Shoemaker 2017.

Response-Dependent Theories of Responsibility    317 and criticizable—​for having them, even though they are not subject to our voluntary control. Nevertheless, even rational control seems unnecessary for blameworthiness, as illustrated by certain types of tragic negligence. Consider the real-​life case of the loving but tragically negligent father.13 As he was driving his son to a new day care, he starting thinking about the day’s lecture (he was a professor at UC Irvine), and so drove by habit to his university. He left the child in the hot car all day and the child died. There is no question about his love for the child—​everyone who knew him testified to that—​so no implication that his evaluative stance was flawed. Rather, he just forgot. But that forgetting was blameworthy. If one is hesitant to sign on to this reaction, note that at least it seems quite apt for his wife to blame him.14 Regarding the epistemic condition, there seem to be plenty of cases of blameworthy agents who don’t meet it. For example, a worker might toss heavy slates off of a roof, and in so doing endanger those walking below, without knowing what he is doing, or even knowing that what he is doing is wrong (Clarke 2014: 160). He is nevertheless blameworthy, even if we cannot trace his negligence to any prior blameworthy act (such as deliberately setting aside his earlier concerns about pedestrian safety). Second, we assign blameworthiness to plenty of ignorant people, people who are not culpable for their ignorance. Suppose someone grew up in the isolated bayous of Louisiana in a racist household, and so embraced racist values. When he has racist attitudes or performs racist actions, we will find him blameworthy, regardless of whether he knows what he is doing is wrong (see, e.g., Faraci and Shoemaker 2014). Indeed, we tend to find what Huck Finn does in failing to turn in the slave Jim praiseworthy, even though he explicitly thinks he’s doing the wrong thing (and so is ignorant of what’s right, at least under that description; see Arpaly 2003: 75–​78). There are, therefore, serious prima facie counterexamples to each of the purported necessary conditions for blameworthiness. What about their joint sufficiency? Even together there are prima facie false positives for this grouping of properties. A mother might favor preventing an injury to her son over preventing the deaths of five people, and so voluntarily and with control do something she knows to be immoral, and yet not be blameworthy (McKenna 2012: 19). Or we might think that Sophie, in making her choice to allow her daughter to be killed by the Nazis over her son, is not blameworthy, despite knowingly, voluntarily, and with control doing something immoral (McKenna

13  See “UCI Professor Avoids Charges,” Los Angeles Times, 4 October 2003. URL: http://​ articles.latimes.com/​2003/​oct/​04/​local/​me-​hotcar4 14  Gary Watson discussed this case as a prima facie challenge to quality of will theories generally in a 2016 talk, “Second Thoughts.” As with all of these prima facie counterexamples, there are attempted replies one might give to buttress one’s theory against them. Here one might say, for instance, that we blame him insofar as he could have been more vigilant, which implicates control (see Pereboom 2015). Perhaps. But perhaps even when all such opportunities for vigilance are removed, blame remains. At any rate, it’s a very hard case. My only point here is that some fancy dancing is needed to account for the case in any of the standard theories of responsibility. See more on this method in the next footnote.

318   David Shoemaker 2012: 20). What’s missing for blameworthiness, in each case, seems to be something like poor quality of will. Suppose, then, that we add poor quality of will to the list of properties rendering bad agents blameworthy. Will the list now be jointly sufficient? No. As many have argued, we need in addition a history condition, for example, that one’s quality of will must come about as a result of non-​objectionable historical factors (i.e., no manipulation or brainwashing; see, e.g., Fischer and Ravizza 1998: ch. 7; Mele 2006: chs. 6–​7; Vargas 2013: ch. 9; and Pereboom 2014: ch. 4). Nevertheless, even adding a history condition is insufficient to account for all cases. Consider the famous example of moral luck: Two drivers, both of whom are texting while driving, pass through a quiet neighborhood. Nothing happens as the first driver goes through, but as the second driver goes through, a dog runs out into the street and is hit. While we may agree that both drivers exhibit every one of the preceding properties to exactly the same degree, many people feel that the second driver is more blameworthy, especially the dog’s loving owners. If so, then there is some additional property that needs to be added to the list. But what might that property be? And even if we can articulate it, why should we think that that will be the end of it, that there won’t be additional false positives we cannot fully account for?15 Again, we could try for a long disjunctive theory (blameworthiness requires voluntariness, control, knowledge, poor quality of will, and/​or historical properties), but then the theory will lack any real explanatory value. The more plausible approach, then, as with the funny, abandons response-​ independence for response-​dependence about the blameworthy, that is, it starts by answering the characterization question by admitting that it is our responses themselves that determine the contours of the blameworthy. Here is the official formulation of Normative Response-​Dependence: X is blameworthy for some action or attitude A if and only if, and in virtue of the fact that, X merits blaming anger for A.

And then in turning to the unity question, the advocate of Normative Response-​ Dependence will say that, just as one cannot adequately unify all of the items on the

15  I appreciate that response-​independent theorists have attempted to respond to many of these counterexamples. They are merely prima facie counterexamples, after all, just part of the mix of data points in deploying the method of reflective equilibrium. My point has only been to show that these cases generate fancy dancing on the part of all of these theorists, that is, they do not sit at all easily with the theories as given. Part of my burden-​shifting argument, then, is that a response-​dependent view can accommodate all the cases without the need for any fancy dancing. That is one reason it is the more plausible default view. (For this argumentative move regarding blame, see Shoemaker and Vargas, 2021.) But I understand that in order to have a fully convincing argument against response-​independent theories, I need to show that there are counterexamples even the method of reflective equilibrium can’t account for. Think of what I’m doing here as prolegomena to that more ambitious project. (Thanks to Dana Nelkin for pressing this point.)

Response-Dependent Theories of Responsibility    319 funny list without constitutive reference to our sense of humor, so too one cannot adequately unify all of the items on the blameworthy list without constitutive reference to our sense of blaming anger. Of course, once again, we now need an account of the merit, or fittingness, relation. After all, there seem to be quite objective properties to which our blaming anger is a merited response, for example, cruelty, insensitivity, insults, and more. That is why a genuinely response-​dependent theory has to ask and answer the deeper question of why some objective properties constitute the fitting anger-​makers but not others. Let us take the just-​mentioned list to be glossed by “slights.” If slights are objective properties of human interactions, and they are what make blaming anger fitting, then the Normative Response-​Dependent theorist’s position is that what makes slighting someone—​and not, say, farming, kneeling, or nose-​blowing—​the fitting anger-​maker is, constitutively, our sensibilities: These are the sorts of properties to which refined human anger sensibilities tend to respond with blaming anger. There is an immediate objection, though: “Isn’t this just a descriptive dispositional view after all, just at a higher level of abstraction? You’re just describing how a more refined sensibility is disposed to respond. But then don’t the earlier objections to the dispositional view apply here too? Couldn’t even a refined sensibility get things wrong? Where is the normativity?”16 There are two replies a response-​dependent theorist might give to this challenge, which I can only survey very briefly here. First, normativity has to bottom out somewhere, and for the response-​dependent theorist, it would most plausibly bottom out at the level of human sensibility. This means that, when the person with the refined sensibility has a clear-​eyed view of the matter, isn’t tired or depressed, and isn’t under some other distorting influence, her responses are simply what constitute the fitmakers, and so determine the relevant oughts, the reasons we have to be amused (in the case of the funny) or to be angry (in the case of blaming anger).17 A second reply is that the relevant normativity may be found in the refined sensibility. Interestingly, a response-​dependent theorist can be neutral about what determines refinement in a sensibility. For instance, it may well be that the correct story about it is response-​independent, perhaps teleological. One might think, for instance, that there is a distinctly human good, generating reasons independent of our attitudes and particular sensibilities to develop ourselves in order to promote or achieve it. Among these are likely to be reasons to develop our sensibilities in ways called for by that perfectionist good, as their refinement contributes to our human perfection. Once refined, what the sensibilities would then tend to respond to would determine the contours of some of our more specific values, such as amusement and blameworthiness. So the story about refinement would be response-​independent, and its normativity would transfer to what

16  17 

This was an objection first pressed on me, albeit without the attitude, by Angela Smith. I am grateful to Chandra Sripada for suggesting a line like this.

320   David Shoemaker remains a response-​dependent account of the more specific (sensibility-​grounded) values.18 There are of course several other objections one might have to a response-​dependent theory of the blameworthy. Unfortunately, I lack space to survey them here, but for some discussion and response, see Shoemaker 2017.

6.  Beyond Anger I have argued for a response-​dependent theory exclusively about what merits only one specific responsibility response, namely, blaming anger. Of course, anger covers a big part of what people take to be the responsibility terrain, but, as noted earlier, we have many other responsibility responses beyond anger, namely, admiration, disdain, gratitude, pride, regret, hurt feelings, and more. Does a response-​dependent treatment apply to their targets as well? And what about responsibility for neutral actions, those for which we simply have no responses to survey?19 Surely, for example, I’m responsible for taking a walk this morning, despite there being no obviously apt emotional responses to my failure or success in doing so. These are good questions, and they point to some possible future directions in theorizing about responsibility. My hunch is that a response-​dependent account will be quite promising for those responsibility-​responses that are pan-​cultural, emotions from which it would be plausible to draw from in articulating universal features of human responsibility. The relevant methodology would involve investigating their fittingness conditions (perhaps in virtue of exploring the emotions’ motivational components; see Frijda 1986 and Scarantino 2014). Given their presence in most or all human societies, therefore, admiration, contempt, pride, and regret are among the emotions having the best chance of playing a role in that humanity-​wide response-​dependent story, as they are the best candidates for what are issued by distinctively human sensibilities. But the more “cognitively sharpened” responsibility responses—​resentment, indignation, moral (dis)approval, and, perhaps, gratitude—​might require a response-​ independent story, at least to the extent that built into their cognitive sharpenings are judgments about the antecedent responsibility of their targets. It is not entirely clear how to read these results, if true. One beckoning possibility returns us to Strawson, however: Perhaps there are several “natural” human emotional responsibility responses that constitute a response-​dependent story of natural responsibility, whereas the more cognitively sharpened emotional responses are part of a response-​independent story of moral responsibility.20 One diagnostic explanation, then, for the ongoing disputes 18 

I’ve profited here from discussion with Shawn Wang. Thanks to Carolina Sartorio for pressing this point on me. 20  See Strawson’s discussion of the “sympathetic or vicarious or impersonal or disinterested or generalized analogues” to the reactive attitudes in 1962/​2003: 83ff. See also Nick Sars’s discussion of this 19 

Response-Dependent Theories of Responsibility    321 in theorizing about responsibility could be that theorists are talking at cross-​purposes, some theorizing about the natural responses, others theorizing about the moral responses. Consequently, adopting the strategy suggested herein may at least help dissolve such disputes. What it will say more generally about the nature of responsibility remains an open question.21

References Adams, Robert Merrihew. 1985. “Involuntary Sins.” The Philosophical Review: 3–​31. Arpaly, Nomy. 2003. Unprincipled Virtue. Oxford: Oxford University Press. Bennett, Jonathan. 1980. “Accountability.” In Zak van Straaten, ed., Philosophical Subjects. Oxford: Clarendon Press, pp. 14–​47. Clark, Michael. 1987. “Humor and Incongruity.” In John Morreall, ed., The Philosophy of Laughter and Humor. Albany: State University of New York, pp. 139–​155. Clarke, Randolph. 2014. Omissions: Agency, Metaphysics, and Responsibility. Oxford: Oxford University Press. D’Arms, Justin, and Jacobson, Daniel. 2003. VIII. “The Significance of Recalcitrant Emotions: Justin D'Arms and Daniel Jacobson. Royal Institute of Philosophy Supplement 52:127–​145. D’Arms, Justin, and Jacobson, Daniel. 2006. “Sensibility Theory and Projectivism.” In David Copp, ed., The Oxford Handbook of Ethical Theory. Oxford: Oxford University Press, pp. 186–​218. D’Arms, Justin, and Jacobson, Daniel. Forthcoming. Rational Sentimentalism. Oxford: Oxford University Press. Darwall, Stephen. 2006. The Second-​ Person Standpoint. Cambridge, MA: Harvard University Press. Faraci, David, and Shoemaker, David. 2014. “Huck vs. JoJo: Moral Ignorance and the (A) Symmetry of Praise and Blame.” Oxford Studies in Experimental Philosophy 1: 7–​27. Fischer, John Martin, and Ravizza, Mark, eds. 1993. Perspectives on Moral Responsibility. Ithaca, NY: Cornell University Press. Fischer, John Martin, and Ravizza, Mark. 1998. Responsibility and Control. Cambridge: Cambridge University Press. Frijda, Nico. 1986. The Emotions. Cambridge: Cambridge University Press. Frijda, Nico. 2007. The Laws of Emotion. Mahwah, NJ: Lawrence Erlbaum Associates Publishers.

“argument by analogy” in his dissertation ms., Tulane University. This view is also compatible, I think, with the view laid out by Vargas 2013. 21  I’m grateful to Dana Nelkin and Derk Pereboom, both for the invitation to write an entry in this volume and for their terrific comments on my initial draft. Since publishing earlier work on response-​ dependence and responsibility, I’ve had a chance to discuss it with—​and learn from—​some of the best in the business, including most especially Randy Clarke, Justin D’Arms, Andrew Eshleman, Dan Jacobson, Michael McKenna, Angie Smith, Chandra Sripada, and Manuel Vargas. I also profited from an email exchange with Shawn Wang.

322   David Shoemaker Graham, Jesse, Haidt, Jonathan, and Nosek, Brian A. 2009. “Liberals and Conservatives Rely on Different Sets of Moral Foundations.” Journal of Personality and Social Psychology 96: 1029–​1046. Hankins, Keith. 2016. “Adam Smith’s Intriguing Solution to the Problem of Moral Luck.” Ethics 126: 711–​746. Hume, David. 1983. Enquiry Concerning the Principles of Morals. Indianapolis, IN: Hackett Publishing Company. Kauppinen, Antti. 2017. “Moral Sentimentalism.” Stanford Encyclopedia of Philosophy (Spring Edition). URL: https://​plato.stanford.edu/​archives/​spr2017/​entries/​moral-​sentimentalism/​ McGraw, Peter A., and Warren, Caleb. 2010. “Benign Violations: Making Immoral Behavior Funny.” Psychological Science 21: 1141–​1149. McHugh, Conor. 2012. “Control of Belief and Intention.” Thought: A Journal of Philosophy 1: 337–​346. McKenna, Michael. 2012. Conversation and Responsibility. New York: Oxford University Press. Mele, Alfred. 2006. Free Will and Luck. Oxford: Oxford University Press. Nagel, Thomas (1986). The View From Nowhere. Oxford University Press. Olin, Lauren. 2016. “Questions for a Theory of Humor.” Philosophy Compass 11: 338–​350. Pereboom, Derk. 2014. Free Will, Agency, and Meaning in Life. Oxford: Oxford University Press. Pereboom, Derk. 2015. “Omissions and Different Sense of Responsibility.” In Dana Nelkin and Sam Rickless, eds., Agency, Freedom, and Moral Responsibility. London: Palgrave Macmillan, pp. 179–​191. Pereboom, Derk. 2017. “Responsibility, Regret, and Protest.” Oxford Studies in Agency and Responsibility 4: 121–​140. Pickard, Hanna. 2014. “Responsibility Without Blame: Therapy, Philosophy, Law.” Prison Service Journal 2013: 10–​16. Pickard, Hanna, and Ward, Lisa. 2011. “Responsibility Without Blame.” In KWM Fulford, Martin Davies, Richard Gipps, George Graham, John Sadler, Giovanni Stanghellini, and Tim Thornton, eds., The Oxford Handbook of Philosophy & Psychiatry. Oxford: Oxford University Press, pp. 1134–​1154. Portmore, Douglas. 2016. “Maximalism and Moral Harmony.” Philosophy & Phenomenological Research, doi: 10.1111/​phpr.12304. Russell, Paul. 1995. Freedom and Moral Sentiment: Hume’s Way of Naturalizing Responsibility. New York: Oxford University Press. Russell, Paul. 2017. The Limits of Free Will. New York: Oxford University Press. Scanlon, T.M. 1998. What We Owe to Each Other. Cambridge, MA: The Belknap Press of Harvard University Press. Scarantino, Andrea. 2014. “The Motivational Theory of Emotions.” In Justin D’Arms and Daniel Jacobson, eds., Moral Psychology and Human Agency. Oxford: Oxford University Press, pp. 156–​185. Shoemaker, David. 2015. Responsibility from the Margins. Oxford: Oxford University Press. Shoemaker, David. 2017. “Response-​Dependent Responsibility; or, a Funny Thing Happened on the Way to Blame.” The Philosophical Review 126: 481–​527. Shoemaker, David. 2018. “You Oughta Know! Defending Angry Blame.” In Myisha Cherry and Owen Flanagan, eds., The Moral Psychology of Anger. London: Rowman & Littlefield, pp. 67–​88. Shoemaker, David, and Vargas, Manuel. 2021. “Moral Torch Fishing: A Signaling Theory of Blame.” Noûs 55: 581–602.

Response-Dependent Theories of Responsibility    323 Smith, Adam. 1976. The Theory of Moral Sentiments: The Glasgow Edition of the Works and Correspondence of Adam Smith. Oxford University Press UK. Smith, Angela M. 2004. “Conflicting Attitudes, Moral Agency, and Conceptions of the Self.” Philosophical Topics 32: 331–​352. Smith, Angela M. 2005. “Responsibility for Attitudes: Activity and Passivity in Mental Life.” Ethics 115: 236–​271. Smith, Michael, Lewis, David, and Johnston, Mark. 1989. “Dispositional Theories of Value.” Proceedings of the Aristotelian Society, Supplemental Volume 63: 89–​174. Sommers, Tamler. 2012. Relative Justice. Princeton, NJ: Princeton University Press. Stern, Lawrence. 1974. “Freedom, Blame, and Moral Community.” Journal of Philosophy 71: 72–​84. Strawson, P. F. 1962. “Freedom and Resentment.” Proceedings of the British Academy 48: 1–​25. All citations from its reprint in Gary Watson, ed., Free Will, 2nd ed. Oxford: Oxford University Press (2003), pp. 72–​93. Todd, Patrick. 2016. “Strawson, Moral Responsibility, and the ‘Order of Explanation’: An Intervention.” Ethics 127: 208–​240. Vargas, Manuel. 2013. Building Better Beings. Oxford: Oxford University Press. Wallace, R. Jay. 1994. Responsibility and the Moral Sentiments. Cambridge, MA: Harvard University Press. Wallace, R. Jay. 2014. “Emotions and Relationships: On a Theme from Strawson.” Oxford Studies in Agency and Responsibility 2: 119–​142. Watson, Gary. 2004. Agency and Answerability. Oxford: Oxford University Press. Watson, Gary. 2014. “Peter Strawson on Responsibility and Sociality.” Oxford Studies in Agency and Responsibility 2: 15–​32. Williams, Bernard. 1985. Ethics and the Limits of Philosophy. London: Fontana.

Pa rt V I I

R E SP ON SI B I L I T Y, K N OW L E D G E , A N D C AU S AT ION

CHAPTER 16

Et hics Is Hard ! W hat F oll ows? On Mora l Ignorance a nd Bl a me Elizabeth Harman

1.  Ethics Is Hard Ethics is hard. It’s hard to know what our moral obligations are. It’s hard to know whether what we are inclined to do is morally permissible. It’s hard to know what we owe to those we love, and how much we ought to do for strangers. It’s hard to know what we owe to animals. It’s hard to know how we should treat others when they behave badly. Smart, thoughtful people go wrong on moral questions. People who are trying to live morally go wrong. Sometimes people are trying to act as they should, but they are wrong about how they should act, and they do morally wrong things. What follows from the fact that ethics is hard? There are many different implications we could examine. In this chapter, I will focus on the implications for blameworthiness. When someone acts morally wrongly because she is caught in the grip of a false moral view, although she has thought a reasonable amount about morality, is she thereby blameless for so acting? Recently, a number of philosophers have embraced the view that moral ignorance does exculpate in such cases.1 In this chapter, I will outline an attractive line of 1  The claim that moral ignorance exculpates is made or suggested in Susan Wolf ’s “Moral Saints,” Journal of Philosophy 79 (1982), 419–​439; Sarah Buss’s “Justified Wrongdoing,” Noûs 31 (1997), 337–​369; Michael Zimmerman’s “Moral Responsibility and Ignorance,” Ethics 107 (1997), 410–​426; Gideon Rosen’s “Culpability and Ignorance,” Proceedings of the Aristotelian Society 103 (2003), 61–84; Gideon Rosen’s “Skepticism about Moral Responsibility,” Philosophical Perspectives 18 (2004), 295–313); Michael Smith’s “Moore on the Right, the Good, and Uncertainty,” in Metaethics After Moore, Terence Hogan and Mark Timmons, eds., Oxford: Oxford University Press (2006), 133–​148; and Michael Zimmerman’s Living with Uncertainty: The Moral Significance of Ignorance, Cambridge: Cambridge University Press (2010). The claim that moral ignorance exculpates is assumed (but not focused on) in Jeff McMahan’s Killing in

328   Elizabeth Harman thought according to which moral ignorance exculpates. I will then argue that this line of thought is mistaken: being caught in the grip of a false moral view is not exculpatory. Ethics is hard, that’s true. But this doesn’t have the implications for blameworthiness that some people think it does.

2.  Our Question It is clear that ignorance is sometimes exculpatory. In particular, nonmoral ignorance can be exculpatory. Consider Anne, whose husband Bert is ill. She gives Bert what she has every reason to believe is the cure to his illness; in fact, it is poison. Anne does something morally wrong: She poisons her husband. Is she blameworthy? It seems not. At least, she is not blameworthy if she was responsible in managing her beliefs. The following seems true: Anne is blameworthy for poisoning Bert only if she has violated a procedural moral obligation regarding the management of her beliefs (only if she has failed to adequately investigate or to take evidence seriously).2

Anne’s case is one example of the following phenomenon: nonmoral ignorance can exculpate. Our question is whether moral ignorance can similarly exculpate. Let’s consider two cases of agents who do not know the moral truth about their situations. These agents act morally wrongly while caught in the grip of false moral

War, Oxford: Oxford University Press (2009); and Holly Smith’s “The Subjective Moral Duty to Inform Oneself Before Acting,” Ethics 125 (2014), 11–​38. 2  Perhaps the consequent of this conditional should be amended to include other ways one can be morally responsible for one’s nonmoral ignorance, and thereby blameworthy for a morally wrong action. For example, perhaps it should read “only if: (she has violated a procedural moral obligation regarding the management of her beliefs [that is, she has failed to adequately investigate or to take evidence seriously]) or (her belief is a result of motivated ignorance).” A person’s belief is a result of motivated ignorance when the true psychological explanation of why she has the belief is that she wanted it to be true, it was better for her that it be true, or she didn’t want to face its falsity. Michele Moody-​Adams has argued compellingly that motivated ignorance can be a source of blameworthiness for morally wrong action. See Michelle Moody-​Adams’s “Culture, Responsibility, and Affected Ignorance,” Ethics 104 (1994), 291–​309. If the consequent of this conditional should be amended in this way (or in other ways) to include additional ways that one can be morally responsible for one’s nonmoral ignorance, then the corresponding revisions should be made to the three principles discussed on pages 330, 331, and 333, and further stipulations need to be made regarding the central cases of moral ignorance I discuss in the paper: The moral ignorance must be stipulated to not be the result of motivated ignorance. “Our question,” stated on page 329, also has to be revised to include this further stipulation. See footnotes 3, 6, 11, and 13. These revisions would not affect the arguments of the paper. We can take what I have in the main text here, and throughout, as a placeholder for a complete theory of moral responsibility for nonmoral ignorance, which is an interesting topic in its own right. (For discussion of that topic, see Holly Smith, “Culpable Ignorance,” Philosophical Review 92 (1983), 543–​571; and others.)

Ethics Is Hard! What Follows? On Moral Ignorance and Blame    329 views. I stipulate that each agent genuinely believes that what he is doing is morally permissible, that each agent has fulfilled all his procedural moral obligations regarding the management of his beliefs, and that neither agent has any false nonmoral beliefs about his situation. An ancient slaveholder keeps some slaves. He does not believe slaveholding is morally wrong; it has never occurred to him that it might be wrong. A man advocates against the legalization of gay marriage.

These are actual cases. There were ancient slaveholders who never doubted the moral permissibility of slavery; and there are opponents of gay marriage who are sure they are acting morally rightly in trying to prevent it. Ethics is hard; even people who have thought a reasonable amount about morality, and who want to figure out what their moral obligations are, can get it wrong. One might ask: How could these agents, who know the nonmoral truths regarding their situations, really have thought a reasonable amount about morality, and reasoned about morality in good faith? But while their mistakes may be obvious to us, they are not obvious to them. Ancient slavery was based on who conquered whom, not on racial prejudice. The slaveholder knows it is horrible to be a slave. He thinks he is lucky to not be a slave, but that if he had been captured, it would have been morally permissible for his captors to keep him as a slave. The advocate against gay marriage knows that many people disagree with him; but he believes he understands what basic moral mistakes they are making which lead them astray. Nonmoral ignorance exculpates. Our question will be this: Does moral ignorance exculpate too? More precisely, this is our question: If someone does something morally wrong while caught in the grip of a false moral view (according to which what she does is morally permissible or even morally good), and if she has not violated any of her procedural moral obligations regarding the management of her beliefs,3 is she thereby blameless?

Procedural moral obligations regarding the management of one’s beliefs are obligations to do certain things a person can do intentionally, such as thinking about a moral question or considering a particular objection or argument.4 (I will speak of those who hold false moral views as “caught in the grip” of those views, but I do not mean to suggest that 3 

Perhaps we should add “and her belief is not a result of motivated ignorance”; see footnote 2. The obligation to believe only those moral claims that are warranted by one’s evidence is not a procedural obligation but is rather a substantive obligation, so it is not among one’s procedural moral obligations regarding the management of one’s beliefs. A person may fulfill these procedural moral obligations because she tries to get a moral question right, while ending up with a belief that is not warranted by her evidence. (A separate issue is whether this obligation [to believe only those moral claims that are warranted by one’s evidence] is a moral obligation at all, or merely an epistemic obligation.) 4 

330   Elizabeth Harman they could not believe differently; rather, they are “caught” in that it is a bad state to be in, and they are “in the grip” in that they genuinely believe their false moral views.) There are two methods we might use to answer our question. First, we might ask whether the fact that nonmoral ignorance exculpates shows that moral ignorance must exculpate too. Second, we might look directly at cases of wrongdoing by agents caught in the grip of false moral views and ask whether the agents are blameworthy. Let’s pursue the first method and see whether it is fruitful.

3.  What Do We Really Learn from Cases of Nonmoral Ignorance? The idea that if nonmoral ignorance exculpates, then moral ignorance must also exculpate, has been advanced by several philosophers.5 But I will argue that it is misguided. Recall Anne, who poisons Bert thinking she is giving him the cure to his illness. Anne is blameless. What explains this? Here are two possible explanations:

Anne didn’t know she was poisoning Bert. Anne didn’t know she was doing something morally wrong.

These are two different things that someone might say by way of explanation of Anne’s blamelessness. Which one offers the correct explanation? The two possible explanations of Anne’s blamelessness correspond to two general principles: Blameworthiness Requires Some Psychological Ground: A person is blameworthy for behaving in a certain way only if either there is a way of behaving such that (a) she believed she was behaving in that way and (b) behaving

5  See the citations in footnote 1. I discuss the view that moral ignorance exculpates, focusing on Rosen’s 2004 argument, in my “Does Moral Ignorance Exculpate?” (Ratio 24 (2011), 443–468). Rosen 2004 argues for the conclusion that an agent who acts wrongfully while ignorant that she is acting morally wrongly is blameworthy only if her action results from an earlier knowing (akratic) violation of her procedural moral duties in the management of her moral beliefs. In my 2011 sections 1–​3, I discuss a number of objections to his view, some of which appear in or are inspired by points in Moody-​Adams (1994); Alexander Guerrero’s “Don’t Know, Don’t Kill: Moral Ignorance, Culpability, and Caution,” Philosophical Studies 136 (2007), 59–​97; and William FitzPatrick’s “Moral Responsibility and Normative Ignorance: Answering a New Skeptical Challenge,” Ethics 118 (2008), 589–​613. I argue that these objections show that it is not ignorance but false belief that has any chance of exculpating and that even non-​akratic violations of these procedural moral duties may be a source of blameworthiness. I conclude that these objections do not touch a narrower thesis Rosen might hold: that false moral belief exculpates in cases in which an agent has not violated her procedural moral duties in the management of her beliefs. It is that narrower thesis that concerns me in this paper.

Ethics Is Hard! What Follows? On Moral Ignorance and Blame    331 in that way is morally wrong, or she violated some procedural moral obligations regarding the management of her beliefs.

Blameworthiness Requires Moral Knowledge: A person is blameworthy for behaving in a certain way only if either there is some way of behaving such that (c) she believed she was behaving in that way and (d) she knew that behaving in that way is morally wrong, or she violated some procedural moral obligations regarding the management of her beliefs.6 The first principle implies that Anne is blameless because, while she knew she was giving Bert what she believed to be the cure to his illness, it is not morally wrong to give someone what you believe to be the cure to his illness; and while it is morally wrong to poison someone, she didn’t believe she was doing that.7 The second principle thus also implies that Anne is blameless; it simply strengthens the conditions required for blame by including not only that the behavior one believed oneself to be engaged in is wrong but that one knows it is wrong. Note that ignorance does not always exculpate, and neither of the principles above gets this case wrong: Carla spoons something into Daniel’s coffee. She does not know whether it is poison or sugar. She is 50% confident it is poison, and 50% confident it is sugar. In fact it is poison, and it kills Daniel.

Carla is blameworthy for poisoning Daniel. One might worry that a view according to which ignorance exculpates might hold that she is blameless: After all, she didn’t know she was poisoning Daniel. But Carla did know that she was taking a 50% chance of poisoning Daniel, and that this is wrong. So neither principle above implies that Carla is blameless—​she knows she is taking that 50% risk; it is wrong to take that risk, and she knows this is wrong. Which of the two general principles does a better job of explaining Anne’s blamelessness? The first principle does a better job. The first principle, which is a weaker principle, is all we need to explain blamelessness due to nonmoral false belief like Anne’s. Once we see that she did not know about certain nonmoral features of her action, which make it wrong (and that she did not mismanage her beliefs), we see that she is blameless. We need not add the further condition that she did not know she was doing something wrong. Cases of nonmoral ignorance like Anne’s provide no motivation for the second principle. It might have seemed that the fact that nonmoral ignorance exculpates could support the claim that moral ignorance exculpates. But it turns out that cases of nonmoral 6 

Perhaps we should add at the end of each of these principles “or her belief is a result of motivated ignorance”; see footnote 2. 7  Note that we must restrict ourselves to quantifying over certain kinds of knowledge about what one is doing: knowledge of what one is doing, qualitatively described. Anne knows that she is giving this stuff to Bert, and it is in fact morally wrong to give this stuff to Bert.

332   Elizabeth Harman ignorance do not support the second principle. The second principle does imply that moral ignorance exculpates. But to embrace that principle, we would need independent reasons; consideration of exculpation due to nonmoral ignorance does not support it.8

4.  Subjective Wrongness and Blameworthiness Anne’s case is one in which a person caught in the grip of false nonmoral views is blameless for a wrongful action she did not know she was performing. But other credal states besides false belief can also be exculpatory: An agent’s particular state of uncertainty might be exculpatory. Consider the following case.9 Evan is a doctor whose patient is suffering from a life-​threatening condition. Evan knows that medicine A will cure the patient with an unpleasant side effect, and that one of B and C will completely cure the patient with no side effects, while the other will kill the patient. Evan does not know which of B and C is which. Evan gives his patient medicine A.

It is clear that Evan does what he should do. But in saying this, we are making a subjective normative claim: Given his epistemic state, Evan should prescribe medicine A. Similarly, given Anne’s mental state, she should give her husband what she gives him. When discussing Anne, we noted that she poisoned her husband, which one should not do. Anne did something objectively morally wrong, though subjectively morally permissible. The same is true of Evan. He does something objectively morally wrong: He gives his patient a cure with a nasty side effect rather than a cure without any side effect. Evan does something objectively morally wrong, but subjectively morally permissible. (Indeed, what Evan does is subjectively morally required.) It is natural to think that this is why Evan is blameless for acting morally wrongly. Notice that neither of the two general principles we discussed above give the result that Evan is blameless. Anne doesn’t know she is doing something objectively morally wrong; but Evan does! He knows he is giving a less good medicine rather than a better 8 

My point here does not depend on how the second principle is worded. We might instead consider the view that a person is blameworthy for acting in a particular way only if she knew she was acting wrongly. This principle is more concise than the second principle I state; but my objection to the second principle is not that it is too complicated, or that it adds a needless further condition, but simply that it is unmotivated by consideration of nonmoral ignorance exculpating. (Paulina Sliwa, in “Moral Worth and Moral Knowledge,” Philosophy and Phenomenological Research 93 (2016), 393–418, holds that moral knowledge is necessary for praiseworthiness; but I don’t think she would endorse the claim that moral knowledge is necessary for blameworthiness.) 9  Frank Jackson famously introduced a case like this in “Deontic-​Theoretic Consequentialism and the Nearest-​Dearest Objection,” Ethics 101 (1991), 461–​482.

Ethics Is Hard! What Follows? On Moral Ignorance and Blame    333 medicine; he knows it’s objectively wrong to do that. (The two principles don’t imply that Evan is blameworthy; they are simply silent on his case.) Still Evan is blameless. The following seems to provide the right explanation: Blameworthy Only If Subjectively Wrong: An agent is blameworthy for behaving in a particular way only if she behaves subjectively morally wrongly—​thus, only if she behaves as she should not behave, given her whole epistemic state10—​or if she has violated some procedural moral obligations regarding the management of her beliefs.11

This principle is more general than the two principles discussed in the last section. It explains cases such as Evan’s, in which an agent is blameless due to his particular state of uncertainty, although he knows he is doing something objectively morally wrong. And this principle seems to offer an even better explanation of Anne’s blamelessness than the principles in that section: Anne is blameless simply because she does not act subjectively wrongly. She does not act in a way she should not act, given her whole epistemic state (and she has not mismanaged her beliefs). I claim that this principle gives the true explanation of Anne’s blamelessness. Someone might now offer a new argument for the claim that moral ignorance exculpates: 1. Blameworthy Only If Subjectively Wrong: An agent is blameworthy for behaving in a particular way only if she behaves subjectively morally wrongly—​thus, only if she behaves as she should not behave, given her whole epistemic state—​or if she has violated some procedural moral obligations regarding the management of her beliefs. 2. Behaving in a way one believes to be morally permissible is never subjectively morally wrong. Therefore: 3. One is never blameworthy for doing what one believes to be morally permissible, if one has not violated any procedural moral obligations regarding the management of one’s beliefs.

10 

Here I am emphasizing that whether a person behaves subjectively morally wrongly depends on her whole epistemic state. To clarify, whether a person behaves subjectively wrongly is not the same question as whether she behaves as she should not behave, given her whole epistemic state. I am assuming that moral requirement is overriding, and thus that if some behavior is morally wrong, then one should not engage in it, all things considered. But the converse does not follow. Sometimes one should not do something, all things considered, but it is not morally wrong to do it. 11  Perhaps we should add, at the end of this principle, “or her belief is a result of motivated ignorance”; see footnote 2.

334   Elizabeth Harman This argument is, in my view, unsound. But at this point what I want to point out is that it is a question-​begging argument in a discussion of whether its conclusion is true. Taking claim 1 to be true (and to clarify the sense of “subjectively wrong” at use in claim 2), claim 2 is simply what is at issue between those who endorse claim 3 and those (like me) who deny it. Note that what is subjectively morally wrong depends on one’s epistemic state. The question is what role various parts of one’s epistemic state play in determining what is subjectively morally wrong. Do one’s moral beliefs (and credences) alone determine what is subjectively morally wrong for one to do? If they do, then it may seem that it is never subjectively morally wrong to do what one is sure is morally permissible.12 It is clear that one’s moral beliefs are sometimes relevant to what is subjectively morally wrong for one to do. Consider this case: Fred doesn’t know what the red button does. His trusted friend Georgia says, “Be careful! Pushing the red button is deeply morally wrong.” He believes her, yet he pushes the button anyway.

Fred does something subjectively morally wrong in this case. It is his moral belief, acquired on the basis of testimony, that makes what he does subjectively morally wrong. This case shows that sometimes one’s moral beliefs are relevant to what is subjectively morally wrong for one to do. But it doesn’t show that, in general, one’s moral beliefs alone determine what is subjectively morally wrong for one to do. In fact, it doesn’t show that one’s moral beliefs play any role in determining what is subjectively morally wrong for one to do over and above providing evidence for nonmoral beliefs. In this case, Fred would be reasonable in inferring that pushing the red button may well hurt someone. This nonmoral belief would be sufficient to make it subjectively morally wrong to push the button. In the next section I will develop a view on which one’s nonmoral beliefs determine what it is subjectively morally wrong for one to do. (On the view I will develop, one’s moral beliefs are irrelevant to what it is subjectively wrong for one to do—​unless they provide grounds for further nonmoral beliefs.) 12 

It might sometimes be subjectively morally wrong to do something one believes is morally permissible, even if one’s moral beliefs (and credences) alone determine what is subjectively morally wrong for one to do, if one is not sure that it is morally permissible. One might believe that φing is morally permissible while having a small credence that φing is deeply morally wrong. It has been argued that in such a case, one should not φ. (See Jacob Ross’s “Rejecting Ethical Deflationism” Ethics 116 (2006), 742–​768; Alexander Guerrero’s 2007; Andrew Sepielli’s “What to Do When You Don’t Know What to Do” Oxford Studies in Metaethics 4 (2009), 5–28; Dan Moller’s “Abortion and Moral Risk” Philosophy 86 (2011), 425–​443; and William MacAskill’s “Moral Recklessness and Moral Caution” (manuscript).) If these arguments are correct, it is subjectively morally wrong to φ in such a case. I will set such cases aside. Assume that all the agents I discuss who are caught in the grip of false moral views do not just believe those moral views but are sure those moral views are true. See section 10 below, and my paper “The Irrelevance of Moral Uncertainty,” Oxford Studies in Metaethics,” 10 (2015), 53–​79) for my argument that these authors’ views of moral uncertainty are false.

Ethics Is Hard! What Follows? On Moral Ignorance and Blame    335

5.  A View on Which Being Caught in the Grip of a False Moral View Is Not Exculpatory Let’s take stock. The central question we are investigating in this chapter is this: If someone does something morally wrong while caught in the grip of a false moral view (according to which what she does is morally permissible or even morally good), and if she has not violated any of her procedural moral obligations regarding the management of her beliefs, is she thereby blameless?

As I said earlier, there are two methods we might use to answer our question. First, we might ask whether the fact that nonmoral ignorance exculpates shows that moral ignorance must exculpate too. Second, we might look at the cases directly and ask whether the agents are blameworthy. We looked into the first method, and it did not prove fruitful. It turns out that the fact that nonmoral ignorance exculpates gives us no reason to think that moral ignorance exculpates. Let’s turn now to the second method. We already considered two cases of agents who act morally wrongly while caught in the grip of false moral views: the ancient slaveholder and the opponent of gay marriage. Let’s consider two more cases now. In each case, I stipulate that the agents have thought a reasonable amount about morality; have fulfilled all procedural moral duties regarding the management of their moral beliefs; genuinely believe that what they are doing is not just morally permissible, but is morally required; and have no false nonmoral beliefs about their actions:13 A gang member kills a member of an opposing gang out of revenge. Someone in the opposing gang previously killed his friend (and fellow gang member), but he kills someone who didn’t participate in the murder. He believes he is doing the right thing. A Mafia “family” member kills a shop owner who refuses to pay a weekly extortion fee. (The Mafia demands a weekly payment for “protection”; otherwise they threaten violence.) He believes he is doing the right thing.

I claim that in these cases, the agents are blameworthy for their wrongful actions. They know they are killing innocent people (in one case, out of revenge over something the person did not do; in the other case, for financial gain). Indeed, I claim that these are clear cases of blameworthy agents.

13 

Perhaps I should add the stipulation that their false moral beliefs are not the result of motivated ignorance; see footnote 2.

336   Elizabeth Harman One might ask: How could these agents really have thought a reasonable amount about morality, and reasoned about morality in good faith? To see how they could have, let’s spell out each agent’s moral view a bit more. The gang member understands that he himself might be killed one day by a rival gang member; he does not think that the gang member would be acting morally wrongly. The gang member and the Mafia family member, unlike the ancient slaveholder, know that there are people who think their behavior is morally wrong. But they think that others have been suckered into a false touchy-​feely moral view of loving everyone, and that others do not adequately appreciate each person’s moral duties to take care of her own. I claim that these agents are blameworthy for their killings, and that what they do is subjectively morally wrong. What makes their actions something they should not do, in light of their whole epistemic states? It is their knowledge of what they are doing that makes their actions morally wrong: It is subjectively morally wrong to kill someone for one’s own group’s financial gain, or out of revenge at another person’s action. Their moral beliefs, in the absence of any nonmoral knowledge (such as in the button-​pushing case) might have made their actions subjectively morally permissible (by supporting certain further nonmoral beliefs); but given that they know what they are doing, their false moral beliefs cannot turn a subjectively morally wrong action into a subjectively morally permissible action. I will now outline a view on which being caught in the grip of false moral views is not exculpatory. My opponent offers the following two plausible claims: A. Someone who does something morally wrong while caught in the grip of a false moral view (according to which the behavior is morally permissible) is blameworthy for her behavior only if and only because she is blameworthy for her false moral beliefs. B. Someone is blameworthy for her false moral beliefs only if she violated some procedural moral obligations regarding the management of her beliefs. These two plausible claims together imply that being caught in the grip of a false moral view is exculpatory; they imply my opponent’s answer to the central question we are discussing. I deny both of these plausible claims, but I endorse both of the following variants of claim A:

A*. Someone who does something objectively morally wrong but subjectively morally permissible is blameworthy for her behavior only if and only because she is blameworthy for being in the epistemic state that makes her behavior subjectively permissible. A**. Someone who does something morally wrong while caught in the grip of a false moral view (according to which what she does is morally permissible) is blameworthy for her behavior only if she is blameworthy for her false moral beliefs. I grant claim A*, and it is a claim on which my opponent and I can agree. We disagree about the implications of claim A*. My opponent thinks that if one is caught in the grip

Ethics Is Hard! What Follows? On Moral Ignorance and Blame    337 of a false moral view, then the actions one believes to be morally permissible are thereby not subjectively morally wrong; I disagree. In granting claim A**, I grant the only if part of claim A, while disputing the and only because part. In my view, agents who act morally wrongly while caught in the grip of false moral views are indeed blameworthy for their false moral views. But their blameworthiness for their moral beliefs does not explain their blameworthiness for their actions; they are not blameworthy for their actions merely because they are blameworthy for their beliefs; rather, their actions and their beliefs are blameworthy for similar reasons. I adopt this view of blameworthiness for behavior: Behaving in a certain way is blameworthy just in case (and to the degree that14) the behavior results from the agent’s caring inadequately about what is morally significant—​where this is not a matter of de dicto caring about morality but de re caring about what is in fact morally significant.15

An agent cares de dicto about morality if the agent wants to act morally. An agent cares de re about what is in fact morally significant if the agent cares about the features of her actions that actually do matter morally. For example, an agent who wants to keep her promises and to avoid hurting others, thereby cares de re about some things that are in fact morally significant.16 Here is my view about the blameworthiness of beliefs17: Beliefs (and failures to believe) are blameworthy if they involve inadequately caring about what is morally significant. Believing a certain kind of behavior is wrong on the basis of a certain consideration is a way of caring about that consideration. Some failures to believe moral truths relevant to one’s behavior are not blameworthy. For example, if one blamelessly falsely believes a nonmoral claim, and this leads to one’s false moral belief, then one’s false moral belief does not involve inadequately caring about what is morally significant.

Let’s consider the Mafia family member to illustrate my view. The Mafia family member knows he is killing an innocent person in order to secure financial gain for his family. He is inadequately moved by the fact that his action kills an innocent person. He acts as he does because he does not have a strong enough desire not to kill innocent persons—​he 14  A person’s degree of blameworthiness depends on several factors, including both the extent to which she cares inadequately about what is morally significant and the extent to which what she is doing is a morally wrong or bad thing to do. (For example, killing someone is more blameworthy than telling a small wrongful lie, even if both actions result from equally complete indifference to what is morally significant in each case.) 15  This view has been argued for by Nomy Arpaly, in Unprincipled Virtue (2003); see also Julia Markovits, “Acting for the Right Reasons,” Philosophical Review (2010) 119(2), 201–​242, which develops a view similar to Arpaly’s, but focuses on the moral worth of morally good actions. 16  A consequentialist would deny that promise keeping actually matters morally. 17  This part of my view goes beyond the views in Arpaly 2003 and Markovits 2010.

338   Elizabeth Harman does not care enough about not killing innocent persons. This is why he is blameworthy for the killing. But on my view, the Mafia family member is also blameworthy for his moral beliefs. His belief that it is morally right to kill the shop owner is a way of having a morally objectionable attitude to the shop owner—​holding it is to hold that the shop owner’s life is cheap and can permissibly be sacrificed to his family’s own goals. This attitude itself is blameworthy, on my view. Another interesting case to consider is the much-​discussed case of Huckleberry Finn, who travels with Jim, an escaped slave.18 Huck believes that he is morally required to turn Jim in, but at a crucial moment when he could easily do so, Huck does not. He “resolves to be bad” instead. One version of the case is this: Huck does genuinely believe that it is morally required to turn Jim in, but despite this he is moved by Jim’s humanity, and this is why he refrains from turning Jim in. It has been asked: is Huck praiseworthy for acting? I have not offered an account of praiseworthiness, but the following sits nicely with the view I have proposed: An agent is praiseworthy for a morally good action just in case the agent’s action resulted from caring about the features of the situation that make the action a morally good action.19

On this view, Huck is praiseworthy for refraining from turning Jim in. This is a conclusion that I can happily embrace. But on my view, Huck is also blameworthy for something: his moral belief that he should turn Jim in. Huck’s psychology, in my view, involves his both caring about Jim’s humanity—​it moves him to refrain from turning Jim in—​yet also not adequately caring about Jim’s humanity—​it does not move him to believe that Jim deserves to not be a slave; it does not prevent his false moral belief. There are two ways we could understand Huck’s psychology, in both of which he is somewhat blameworthy on my view. One possibility is that Huck does care about Jim’s humanity, but not fully. The other possibility is that Huck simultaneously has two conflicting attitudes, two conflicting levels of care toward Jim’s humanity: He cares about it fully, but he also cares very little about it. (Compare the way that a person might have two conflicting beliefs: I believe I will be off campus on Tuesday, and I also believe I’ll have lunch in the cafeteria with Adam on Tuesday.) On my view, Huck is both praiseworthy for refraining from turning Jim in and blameworthy for his moral view about Jim. I think this is the right thing to say about Huck, but

18  Huckleberry Finn and Jim are characters in the novel Huckleberry Finn by Mark Twain, originally published in 1884. Discussions of Huck Finn and whether he is praiseworthy, blameworthy, or neither occur in Jonathan Bennett’s “The Conscience of Huckleberry Finn,” Philosophy 187 (1974), 123–​134; Arpaly 2003, Markovits 2010; Sliwa 2016; and others. 19  This is the view in Arpaly 2003; a similar view is developed in Markovits 2010 (Markovits focuses on an agent’s motivating reasons, rather than what the agent cares about). I raise a worry about how Arpaly’s view should be developed in my “Discussion of Nomy Arpaly’s Unprincipled Virtue,” Philosophical Studies 2007.

Ethics Is Hard! What Follows? On Moral Ignorance and Blame    339 I realize that those who have been at pains to point out that Huck is praiseworthy might want to say that his praiseworthiness for acting is the whole story. Note that “inadequately caring” about what matters morally may be a matter of having a certain caring attitude (or set of attitudes); it may not merely be the absence of a certain caring attitude (or set of attitudes). We see this in the version of Huck in which he both fully cares about Jim but also inadequately cares; he has two conflicting attitudes of care. The view I have offered in this section has three aspects that we might distinguish. There is (a) my view that being caught in the grip of a false moral view is not exculpatory; and there is my account of why this is true, which is (b) my view about blameworthiness for action; I have also offered (c) my view about blameworthiness for belief. In the next three sections, I will explore each of these three aspects of my view, clarifying my view and responding to objections.

6.  Objections to My View That Being Caught in the Grip of a False Moral View Is Not Exculpatory In this section, I will consider two objections to my view that being caught in the grip of a false moral view is not exculpatory. Someone might object as follows: It’s a matter of luck what kind of moral community one is born into, and what kind of moral views one is taught. As a matter of luck, one might end up caught in the grip of false moral views. But one can’t be blameworthy simply because one experienced bad luck, so being caught in the grip of a false moral view must exculpate.

This objection fails because one can be blameworthy as a result of experiencing bad luck. For example, one might experience bad luck in the situations one faces. Two people may have very similar dispositions. One may be faced with a difficult moral choice between doing the right thing at great cost to himself and doing the selfish, wrong thing. If he chooses the wrong thing, he is morally unlucky compared to the other person, who might well have also chosen the wrong thing in this situation but never faced this choice and so is not blameworthy in this way. This kind of moral luck exists; but there are also other kinds of moral luck. There is moral luck in the consequences of one’s action: A person who has one glass of wine too many, drives, and kills a child as a result is more blameworthy than a person who acts similarly but is lucky not to encounter a child in his path. The first person is blameworthy for killing this child and that family’s resentment of him is appropriate. These are ways that he is blameworthy that have no correlate for the other luckier drunk driver. Similarly, I claim, there is moral luck in one’s constitution:

340   Elizabeth Harman One may be unlucky to have become a bad person, and people caught in the grip of false moral views are unlucky to have been constituted as bad people. They are blameworthy for their actions, and unlucky to be blameworthy in this way.20 Another objector might say the following: Perhaps people caught in the grip of false moral views are blameworthy for their wrongful actions when they know the features of the actions that in fact make them wrong. But surely someone who acts wrongly while knowing she is acting wrongly is more blameworthy than someone who does not know she is acting wrongly.

This is a natural thought. But we can see that it is mistaken by considering the following case: Kurt believes that homosexuality is morally wrong and that the government should not endorse these morally wrongful unions by allowing same sex couples to legally marry. He works hard to prevent same sex marriage from being legal in his state; he is a prominent public intellectual whose advocacy has a real effect. (Kurt is a sophisticated contemporary opponent of same-​sex marriage who has no false nonmoral beliefs about homosexuals and homosexuality.21) Louis believes that homosexual love is no different from heterosexual love and that same sex couples should be legally entitled to marriage. But Louis is a talented political operative who sells his services to the highest bidder. An anti-​same-​sex-​marriage group hires him, and he devotes himself tirelessly to preventing same sex marriage from being legal. He knows it’s morally wrong to do this kind of work just for a lot of money, but he does it anyway. The efforts of Kurt, Louis, and others prevent same sex marriage from becoming legal in Mark’s state of residence. This places a variety of burdens on Mark and his same-​sex partner.

What attitudes should Mark have toward Kurt and Louis? In particular, what kinds of resentment is it appropriate for Mark to feel toward them? Both agents behave morally wrongly in a way that is bad for Mark. Both contribute to the denial of basic rights to Mark. If the objector were correct, then Mark would be licensed in being more resentful of Louis than of Kurt. Unlike Kurt, Louis knows that what he is doing is morally wrong. But Mark may appropriately feel more resentful of Kurt than Louis; indeed, there are two kinds of resentment that Mark could reasonably feel toward Kurt, though he could reasonably feel only one kind toward Louis. Mark could appropriately and reasonably resent both Kurt and Louis for working hard to deny his rights, knowing that they were 20  See Thomas Nagel, “Moral Luck,” in Mortal Questions, Cambridge: Cambridge University Press (1979), which is more ambivalent about the phenomenon of moral luck that I am. In my view, moral luck is simply a real phenomenon. (Note that one might endorse my view that false moral belief is not exculpatory, and thus endorse the existence of moral luck in one’s constitution, without endorsing the independent claim that luck in the consequences of one’s actions can affect how blameworthy one is.) 21  Some of these do exist.

Ethics Is Hard! What Follows? On Moral Ignorance and Blame    341 working hard to deny certain rights to him simply because his life partner is a man and not a woman: Mark can resent each of them for failing to take the impact on him (and others like him) as sufficient reason to refrain from acting. Mark could also appropriately resent Kurt for taking Mark’s relationship to be less morally valuable than a heterosexual relationship: Mark could appropriately resent the moral attitude Kurt takes toward Mark’s relationship in Kurt’s moral beliefs about Mark and his relationship. But this second kind of resentment is not one that Mark could appropriately have toward Louis. Louis does not take this further morally objectionable attitude toward Mark. Because there is a close connection between how much resentment would be appropriate from the victim of a wrongful action, and how blameworthy the agent of the wrongful action is, I conclude that Kurt is more blameworthy than Louis, and that the objector’s assertion is false.22

7.  My View of Blameworthiness for Behavior In this section, I will say a bit more about the part of my view that concerns blameworthiness for behavior. When a person acts wrongly, this is either because she cares inadequately about the features of her situation that make her action morally wrong, or because while she cares adequately about those features, something has interfered with her action’s manifesting what she cares about. If she has false nonmoral views about her situation, then her action may not manifest inadequate caring about the features that make her action wrong, because she may not know about these features. This is why nonmoral ignorance can exculpate. If a person has a psychological condition that involves delusion, her psychological condition may exculpate; this is an example of nonmoral ignorance exculpating. But even a person who knows her nonmoral situation may still act in a way that does not reflect what she really cares about; a psychological condition such as depression may interfere with her action’s manifesting what she really cares about. The view I am proposing allows that psychological conditions can exculpate (or can mitigate blame) when they interfere with an action’s manifesting what an agent really cares about.23 Note that this view takes there to be a psychologically real phenomenon of caring about certain features of one’s action, which is not simply identical to believing that those features matter morally: Sometimes caring about a feature and believing it matters go

22 

Note that I don’t need to rely on the claim that two kinds of resentment must constitute more resentment than one kind. Rather, I want to point out that Kurt’s moral belief is the appropriate basis for a great deal of resentment, in support of the claim that it is appropriate to feel more resentment of Kurt than of Louis. 23  My view agrees with Arpaly 2003 on this point.

342   Elizabeth Harman together; sometimes they come apart.24 Huck, for example, cares about Jim’s humanity but does not believe that Jim’s humanity matters morally in the same way that a white person’s does. One psychological condition is a bit different: psychopathy. Let’s consider an idealized version of the condition in which the agent knows what his nonmoral situation is, knows what morality requires, but is simply unmoved by any concern for other agents. This is an agent who genuinely does not care about others; it is not the case that his psychological condition interferes with his actions’ expressing his caring. The view I am developing might hold that psychopaths are morally responsible for their morally wrong actions. Or it might hold that there is an additional condition necessary for moral responsibility—​the ability to be moved by any moral considerations at all. While it might seem that psychopaths are clearly not responsible for their actions, I am not so sure. What’s clear is that psychopaths are very different from other people, and that feeling and expressing the reactive attitudes toward them has a different meaning and efficacy than with ordinary persons. Psychopaths are beyond hope. That may affect how it is reasonable to interact with them, even while it may be true that they are fully blameworthy for their actions. Compare two cases in which an insane person murders someone, and the murdered person’s loved ones feel real resentment toward the insane person. In the first case, the insane person was suffering from a full-​blown delusion in which he thought his victim was actively trying to kill him and that killing her was his only hope of surviving. In the second case, the insane person was a psychopath who simply does not care at all about other people, though knows that they can suffer like he does. In this pair of cases, the resentment of the first family is inappropriate while the resentment of the family in the second case is not. While it is understandable that the first family feels resentment toward someone who murdered their loved one, in fact the agent did not show any disrespect toward the victim, nor any failure to value her life; the agent simply had a false view of what was happening in the situation. By contrast, the psychopath understood his situation perfectly, and did fail to care about the life he was ending. That the second family’s resentment is appropriate suggests that psychopaths really are blameworthy.25

24 

In my view, believing a feature matters morally is one way of caring about a feature, but not the only way. 25  The question of whether psychopaths are moral responsibility for their actions is discussed in Matthew Talbert, “Blame and Responsiveness to Moral Reasons: Are Psychopaths Blameworthy?,” Pacific Philosophical Quarterly 89 (2008), 516–​535; Gary Watson, “The Trouble with Psychopaths,” in Reasons and Recognition: Essays in Honor of T. M. Scanlon, sed. R. J. Wallace, R. Kumar, and S. Freeman (Oxford: Oxford University Press, 2011), 307–​331; David Shoemaker, “Attributability, Answerability, and Accountability: Toward a Wider Theory of Moral Responsibility,” Ethics 121 (2011), 602–​632; and Dana Kay Nelkin, “Psychopaths, Incorrigible Racists, and the Faces of Responsibility,” Ethics 125 (2015), 357–​390.

Ethics Is Hard! What Follows? On Moral Ignorance and Blame    343

8.  My View of Blameworthiness for Beliefs In this section, I will consider three objections to my view that agents are blameworthy for false moral beliefs that are relevant to their actions. Here is the first objection: Someone who has fulfilled her procedural moral obligations regarding the management of her beliefs and is caught in the grip of a false moral view is epistemically justified in believing the false moral view and would be epistemically unjustified in believing the true moral view. But one can’t be blameworthy for having beliefs that are justified, and one can’t be blameworthy for failing to believe in an unjustified way.

I will make three points in response to this objection. First, it may well be that some epistemically justified beliefs are nevertheless blameworthy. Consider the view that we owe our friends the benefit of the doubt. We may well owe it to our friends to refrain from believing ill of them, even in the face of sufficient evidence to epistemically justify such beliefs.26 In such a case, an epistemically justified belief would be blameworthy. (This view is compatible with the claim that we don’t owe it to our friends to refrain from thinking ill of them in the face of overwhelming evidence.) For my second point in response, I will grant that some people caught in the grip of false moral beliefs are epistemically justified in these beliefs, on the basis of testimony. Nevertheless, it does not follow that they are not in a position to be epistemically justified in believing the moral truth. Consider the following case. My college friend Nora and I studied math together. She likes to talk to me about math, though it has been years since I studied it. She tells me that a certain mathematical claim is true; I believe her. In fact, that claim is false, and if I thought about it, I might figure that out—​I remember enough math to figure it out. In this case, I am epistemically justified in believing the claim, though I am also capable of becoming epistemically justified in believing it is false. Similarly, a person could be epistemically justified in holding a false moral view, while she is nevertheless in a position to become epistemically justified in believing the moral truth. Third and finally, it is not at all clear to me that those caught in the grip of false moral views ever have epistemically justified beliefs. One’s total evidence is relevant to whether one’s beliefs are justified. Mere testimonial evidence cannot make one justified in believing something when one has adequate evidence to the contrary. Every ordinary

26  Sarah Stroud and Simon Keller have both argued for such a view; Jennifer Lackey has argued they are wrong. See Keller’s “Friendship and Belief,” Philosophical Papers 33 (2004), 329–​351; Stroud’s “Epistemic Partiality in Friendship,” Ethics 116 (2006), 498–​524; and Lackey, “Why There is No Epistemic Partiality in Friendship” (manuscript).

344   Elizabeth Harman person’s life experience gives her a great deal of evidence that is relevant to what is morally required and permissible.27 Here is another objection: If someone couldn’t have done something, she is not blameworthy for failing to do it. People who have fulfilled their procedural epistemic moral obligations but are caught in the grip of false moral beliefs could not have believed otherwise.

My response to this objection echoes my third point in response to the last objection. In fact, each of us has a great deal of moral evidence which makes it possible for us to realize the moral truth. That we do not all succeed when we try does not mean that for some of us, realizing the moral truth is impossible.28,29 Finally, consider this objection: An ordinary person fails to believe lots of moral truths, because she hasn’t thought about those issues at all. Is she thereby blameworthy?

This objection invites me to clarify my view. Inadequately caring about what is morally significant occurs if one forms a specific belief about the issue (even if one merely has an implicit belief) or if the issue is relevant to one’s behavior. But caring adequately about what is morally significant does not require believing all moral truths (not even implicitly).

9.  A Moderate Position? My opponent thinks that false moral belief is exculpatory if a wrongdoer has tried to figure out the moral truth, but then gotten it wrong. I hold that false moral belief is never 27 

I spell out the suggestion in this paragraph in my “Moral Testimony Goes Only So Far,” Oxford Studies in Agency and Responsibility 6 (2019): 165–185. To clarify, my view is that we all have a lot of evidence that makes realizing the moral truth possible for us. My view is not that we have so much evidence that if we try to realize the moral truth, we will do so; obviously this is not true, as many people do try, and get it horribly wrong. 28  As I mentioned in section 1, my locution “caught in the grip” may suggest that a person is stuck and cannot escape her false moral view, but I do not intend to say that about people who are caught in the grip of false moral views. (It is not true of them.) Rather, they are “caught” in that this is a bad rather than a good situation to be in, and they are “in the grip” in that they genuinely believe their false moral views. 29  Note that I do not hold that a person is blameworthy for her false moral views because she could have believed differently. Rather, she is blameworthy for her false moral views simply in virtue of holding them; they are themselves attitudes of inadequately caring about what matters morally. While I do claim that each person could realize the moral truth, that claim is not crucial for my view. We can be blameworthy for features of ourselves that are beyond our control; for a recent defense of this claim, see Pamela Hieronymi, “Reflection and Responsibility,” Philosophy and Public Affairs (2014) 42(1), 3–​41.

Ethics Is Hard! What Follows? On Moral Ignorance and Blame    345 exculpatory. We might consider an intermediate, moderate position according to which false moral belief is exculpatory just in case a wrongdoer has tried to a reasonable degree to figure out the moral truth (has met her procedural obligations) and she has formed an epistemically justified false moral belief. On this view, unjustified false moral belief is never exculpatory, but justified false moral belief can be exculpatory.30 Let’s consider the version of this view on which it is nontrivial: According to the moderate view, some people caught in the grip of false moral views are epistemically justified. I want to make two points about this view. First, it is not clear that this view is intuitively supported by an analogy with the exculpatory power of nonmoral false belief. If I make an honest mistake of nonmoral reasoning, end up with an unjustified false nonmoral belief, and then hurt someone though I think I’m helping her, my honest mistake does tend to exculpate: I am blameless unless blameworthy for having ended up with this unjustified false belief. Some honest mistakes of nonmoral reasoning are blameless, and in such cases I will be blameless for hurting the person I hurt. So this moderate view may be unmotivated: Unjustified false nonmoral beliefs can exculpate, so if some false moral beliefs can exculpate, why would it only be justified false moral beliefs that exculpate? Second, and more importantly, my argument in this chapter can be adapted to show that this view is false. A proponent of this view holds that people can be epistemically justified in holding false moral views. How would they become so justified? A proponent of this view must hold that it is possible to become justified in believing false moral views through testimony, through being convinced by compelling but unsound arguments, or through either route. But then we can consider the versions of the Mafia family member and the gang member who are justified in their false moral views via one or both of these routes. These murderers are nevertheless blameworthy. A proponent of the moderate view might try to block my argument by denying that the Mafia family member or the gang member could be justified in their false moral views. Indeed, there is a tendency among some philosophers to divide moral questions into easy questions and hard questions, supposing that one can be epistemically justified when getting the hard questions wrong, but one cannot be epistemically justified when getting the easy questions wrong.31 But what is easy and what is hard varies greatly depending on where a person sits, in culture and in space. Some of what looks easy to us looked hard to our ancestors; and some of what looks hard to us will no doubt look easy to our descendants. No moral questions are simply easy or hard. If we instead hold that whether a question is easy or hard can vary with the person who is considering it, then we can consider versions of the Mafia family member and the gang member for whom realizing the moral truth would be hard. Nevertheless, these agents know what they are doing (killing innocent people), and are blameworthy. 30  Views along these lines are offered or suggested by Smith (2006), Guerrero (2007), MacAskill (manuscript), and Chelsea Rosenthal, “Ethics for Fallible People” (2019 dissertation, New York University). 31  For two examples, see Fitzpatrick (2008) and MacAskill (manuscript).

346   Elizabeth Harman

10.  The Irrelevance of Moral Uncertainty I have offered a view on which being caught in the grip of false moral views is not exculpatory. On my view, someone who knows her nonmoral situation, but has a false moral belief, subjectively should act as morality actually requires, and not as she believes she should act. A person’s moral beliefs are often irrelevant to how she subjectively should act; they are relevant only when they provide information about her nonmoral situation that she otherwise lacks. Fred, who is told by Georgia not to push the red button, thereby learns some nonmoral information—​that pushing the red button does something bad such as hurting someone and does not do something good such as making someone happy; this is why his moral belief is relevant to how he subjectively should act. This view has implications for recent discussions of moral uncertainty. Some authors have argued that an agent should be morally cautious, refraining from doing something she believes is morally permissible if she has some credence that it might be morally wrong.32 Other authors have argued for the stronger conclusion that agents should sometimes do what they believe is morally wrong if they have some credence that failing to do it might be much more morally wrong.33 These views imply that agents caught in the grip of false moral views subjectively should act as their moral views dictate. These views thus have the underappreciated implication that being caught in the grip of false moral views is exculpatory. One implication of the view I have developed here is that this recent work on moral uncertainty, and the debates internal to it, are misguided. That work assumes that what an agent subjectively should do—​in light of her whole epistemic state—​depends solely on the agent’s normative credences; it does not recognize the role that an agent’s nonnormative credences can play in determining how she subjectively should act.34

11.  Conclusion Because ethics is hard, some people who have been responsible in the management of their moral beliefs have nevertheless ended up with false moral views. 32 

Views along these lines are offered by Guerrero 2007, Moller 2011, and MacAskill (manuscript); see footnote 12. 33  Views along these lines are offered by Ross 2006 and Sepielli 2009; see footnote 12. 34  I offer this argument in “The Irrelevance of Moral Uncertainty,” Oxford Studies in Metaethics 10 (2015), 53–​79. Other arguments that also challenge recent work on moral uncertainty are given in Brian Weatherson, Normative Externalism, Oxford University Press (2019); and in Brian Heddon, “Does MITE Make Right? On Decision Making Under Normative Uncertainty,” Oxford Studies in Metaethics 11 (2016), 102–​128.

Ethics Is Hard! What Follows? On Moral Ignorance and Blame    347 This chapter has discussed the question: When these agents act morally wrongly, believing they are acting morally permissibly, are they blameworthy? I have argued that while nonmoral ignorance does exculpate, moral ignorance does not, and I have developed a view of blameworthiness on which moral ignorance does not exculpate.35

35 

For helpful comments on drafts of this paper, I thank Avery Archer, Peter Graham, Alexander Guerrero, Errol Lord, Thomas Scanlon, Seana Shiffrin, Angela Smith, and audiences at the Australasian Association of Philosophy; the Bellingham Summer Philosophy Conference; Fordham University; Harvard University; Princeton University; Reed College; University of California, Los Angeles; University of Pennsylvania School of Law; University of Southampton; and Vrije University.

CHAPTER 17

Resp onsibi l i t y a nd Causat i on Carolina Sartorio *

1.  Introduction Moral responsibility and causation are connected in important ways. Most obviously, we have the power to causally affect the world, and, arguably, it is thanks to those causal powers that we can be morally responsible (praiseworthy or blameworthy) for outcomes that we help bring about. Causation is, on this picture, our “bridge to the world.” Thus causation seems central to the project of understanding the conditions under which we can be morally responsible for outcomes. This is a derivative form of responsibility. That is to say, our responsibility in these cases is not basic but derived or inherited from other things for which we are responsible, such as the choices that we made and that led to those outcomes. Also, and a bit less obviously, causation is relevant to our moral responsibility in cases of basic responsibility, that is to say, in cases where our responsibility is not inherited from our responsibility for other things. Consider certain choices for which we may be basically responsible. At the very least, causation matters in these cases in that the causal history of those choices (how we made those choices, or what led to our making them) is relevant to whether we made them freely. And freedom is a requirement for moral responsibility—​at least for the form of responsibility that I will be concerned with here.1 *  Thanks to Randy Clarke, Dana Nelkin, Derk Pereboom, and the members of a graduate seminar at the University of Arizona in the fall of 2017 for helpful comments on an earlier draft. 1  This form of responsibility, which is sometimes known as “accountability,” involves being eligible to be held to account for what you do, which requires acting freely or being in control of what you do (see, e.g., Fischer and Ravizza 1998).

Responsibility and Causation    349 In this chapter I examine different aspects of the connection between causation and moral responsibility. I discuss derivative responsibility in section 2 and basic responsibility in section 3.

2.  Derivative Responsibility and Causation It is natural to think of causation as the vehicle of transmission of responsibility (Sartorio 2007). That is to say, assuming that we have figured out what kinds of things we are basically or non-​derivatively responsible for, it is natural to think of causation as the relation that allows for responsibility to be transmitted or passed on to other things, things for which we can then be derivatively responsible (assuming we also meet the relevant epistemic conditions for responsibility; more on this later). In this section I discuss this picture in more detail. As we will see, the connection between causation and responsibility turns out to be quite complex, even when we are dealing with derivative responsibility (where, as I noted before, the link between the two concepts is most obvious). In what follows I discuss some of the main issues that arise when thinking about derivative responsibility and the connection with causation. The role that causation plays in transmitting moral responsibility can be captured by means of principles of derivative responsibility. In general terms, these are principles that give conditions (necessary and/​or sufficient conditions) for the transmission of responsibility from more basic things to outcomes, where one of the conditions is a causal condition. Here are a couple of examples of principles of that kind that seem at least initially plausible: N (for “Necessary”): Responsibility for an outcome requires responsibility for behaving in a way that caused the outcome. S (for “Sufficient”): If an agent is responsible for an act, the act caused an outcome, and the agent foresaw that the act would cause (or was likely to cause) that outcome, then the agent is also responsible for the outcome.

Note that these principles make different kinds of claims about derivative responsibility: N identifies a potentially necessary condition for that kind of responsibility and S identifies a potentially sufficient condition for that kind of responsibility. But both principles involve a causal condition: In the case of N, the causal condition itself is claimed to be a necessary condition for responsibility, and, in the case of S, the causal condition is conceived as one in a larger set of conditions that are only jointly sufficient for responsibility.

350   Carolina Sartorio Note that N and S could be combined into a stronger principle stating conditions that are both necessary and sufficient for derivative responsibility. Here is a possible formulation of such a principle: N&S: An agent is responsible for an outcome if and only if there is some behavior of hers such that: she is responsible for that behavior, the behavior caused the outcome, and she foresaw that the behavior would cause (or was likely to cause) that outcome.

Given that the stronger principle N&S entails both N and S, it inherits any problems that N or S might have. So here I will limit the discussion to an analysis of potential problems that arise for each of the individual principles, N and S. I’ll start with S. Among other things, S attempts to capture the well-​known fact that there are epistemic conditions that an agent needs to meet in order to be morally responsible for anything. In the case of responsibility for outcomes, it is common to understand the epistemic condition in terms of the concept of foreseeability. The lack of foreseeability typically excuses. Imagine that an evil mastermind has secretly rigged things in such a way that eating the last apple in your refrigerator would trigger a series of events that would result in a remote village being wiped out. If you eat the apple and the outcome ensues, you fail to be responsible for the outcome because you didn’t foresee that it would happen. Thus in this case S wouldn’t entail that you are responsible for the outcome. In contrast, S would entail that you are responsible for the outcome in a case where the outcome was a foreseen result of what you did. Imagine, for example, that you drink heavily at a party knowing that you will have to drive back home intoxicated, and foreseeing the likely consequences of your decision. Imagine that on your way back you are involved in an accident where somebody is hurt as a result of your driving under the influence of alcohol. In this case S correctly entails that you are responsible for the harm that you caused. So S seems to capture the intuitive idea that behaving in a way that causes harm is enough for you to be derivatively responsible for the harm, in cases where you are responsible for that behavior and you foresaw that the harm was likely to result from it. Now, note that, as it stands, S won’t be able to capture other cases where we want to hold agents morally responsible for outcomes, such as some cases of negligence. Imagine that you act carelessly, in a way that a reasonable person wouldn’t have, but without actually foreseeing that you could cause any harm. In some such cases we probably would want to hold you responsible for the outcome, even if you didn’t foresee the harm, for (we want to say) you should have foreseen it. So, if we want S to capture these cases, this means that the epistemic condition would have to be relaxed a bit. Recall that S only gives a sufficient condition for responsibility, so negligence cases are not strictly speaking counterexamples to S (they are just cases where S is silent on the agent’s responsibility). But if we are looking for a single unifying principle that covers both kinds of cases, cases of foreseen harms and cases of negligent harms, then the current formulation of S won’t do.

Responsibility and Causation    351 Note that accepting that agents can be responsible in some cases of unforeseen harms lowers the bar considerably for the epistemic conditions that agents would have to meet in order to be responsible. As a result, some argue that it is only appropriate to hold agents responsible in cases of this kind if the agents’ lack of foresight can be traced back to some earlier decision for which they are responsible (see, e.g., Zimmermann 1986). This is sometimes referred to as a “tracing” condition. In turn, others disagree, arguing that agents can be responsible in some of these cases even where their lack of foresight cannot be traced to an earlier event for which they are responsible (see, e.g., Sher 2009 and Clarke 2014). (For a more comprehensive discussion of these and related issues, see the collection of papers in Nelkin and Rickless 2017.) Another potential problem with S concerns cases of causal deviance. These are cases where the agent’s behavior results in the harm, but in a way that radically departs from the way in which it was expected or could have reasonably been expected to do so. Imagine that an evil husband gives his wife a glass of water to drink, mistakenly thinking that it contains poison. Imagine that when the wife drinks the water she unexpectedly chokes on the water and dies. Here it seems that we can’t hold the evil husband responsible for his wife’s death, even if the death was the foreseen result of an act for which he is responsible. The reason is that the causal chain is too “deviant”: there is a significant departure from the way in which the act was expected to issue in the death (for a discussion of the deviance condition, see, e.g., Feinberg 1970). In order to deal with this problem, one could add an extra condition requiring that the causal chain linking the behavior to the outcome not be deviant, or one could tweak one of the existing conditions in such a way that deviant causal chains are ruled out. For example, one could revise the epistemic condition to say that the agent foresaw (or should have foreseen) that the harm would come about in roughly the way it did. As usual, however, it is hard to give a precise account of what a deviant causal chain is, or of how we should understand “roughly” in this revised formulation of the epistemic condition. There are other issues with S that concern the two appearances of the word “responsible.” Recall S’s formulation: S: If an agent is responsible for an act, the act caused an outcome, and the agent foresaw that the act would cause (or was likely to cause) that outcome, then the agent is also responsible for the outcome.

Imagine that the outcome is bad. Then the question arises: Should we just understand “responsible” as “blameworthy”? If so, is this for one or both occurrences of the word? In order to answer these questions, consider the following scenario. You need to leave a room or you’ll asphyxiate. You know that opening the door would result in an explosion in an adjacent room and cause some damage. You freely choose to open the door and walk out of the room. As predicted, the explosion occurs and the harm is done. In this case we don’t want to hold you blameworthy for the harm, because you had a good excuse: Your life was at stake. However, note that there is clearly a sense (a more morally neutral sense) in which you are responsible for your act of opening the door. After

352   Carolina Sartorio all, you did it freely, you were aware of what you were doing, and of the consequences it would have. Plus, your act in fact resulted in the expected outcome (in the expected way). Thus, if we understand the first occurrence of the word “responsible” in principle S in this more morally neutral sense, and the second occurrence as “blameworthy,” S wrongly entails that you are blameworthy for the outcome in this case. One way to address this issue is to understand both occurrences of the word “responsible” in the same way, for example, as meaning “blameworthy.” Here is a possible formulation of the revised principle: S (Blameworthiness): If an agent is blameworthy for an act, the act causally resulted in a bad outcome, and in roughly the expected way, then the agent is also blameworthy for the outcome.2

Note that this principle doesn’t entail that you are blameworthy for the harm in the case where you open the door. For, given the special circumstances you were in, you are clearly not blameworthy for the act of opening the door in that case. Again, you are arguably responsible for that act (and for the outcome) in a more morally neutral sense of the word “responsible,” but you are not responsible for that act (or for the outcome) in the sense of being blameworthy.3 Note that we can formulate a similar principle about praiseworthiness: S (Praiseworthiness): If an agent is praiseworthy for an act, the act causally resulted in a good outcome, and in roughly the expected way, then the agent is also praiseworthy for the outcome.

Arguably, this is also a better formulation than a principle that mixed praiseworthiness with a more neutral sense of responsibility. To illustrate why, consider the following case discussed by Knobe (2003). The CEO of a company implements a policy only moved by the desire to increase the company’s profits. As it happens, a foreseeable side effect of the policy is that it will help the

2 

Feinberg (1970) and Sartorio (2016: 77) propose different versions of this principle. The principle might still need further refinement. This time imagine that there are two doors, A and B, both of which would issue in the explosion, but you wrongly believe that only opening door A would cause the harm. Imagine that you open that door wanting to cause the harm, and the harm ensues. Some would want to say that in this case you are blameworthy for opening door A. (This assumes that you can be blameworthy for acts that aren’t wrong, since opening door A in these circumstances arguably isn’t wrong. For a defense of this view, see, e.g., Zimmerman 1997, Haji 1998, Graham 2011, and Capes 2012.) It seems to me, however, that you are clearly not blameworthy for the harm that you caused. But, assuming that you are blameworthy for opening door A, S (Blameworthiness) entails that you are also blameworthy for the harm. So this would call for further revision of the principle. What seems to be going on in this case is that, even assuming that the agent is blameworthy for her act, and even if the act leads to the harm in the expected way, there is still a mismatch between the agent’s beliefs (such as the belief that she is acting wrongly) and what is actually the case. So, if a revision were needed, it would have to require the absence of that kind of mismatch. 3 

Responsibility and Causation    353 environment, but this is not what moves the CEO to act. In that case, the CEO doesn’t seem praiseworthy for helping the environment, despite the fact that he freely and knowingly chose to implement a policy that foreseeably resulted in such a benefit to the environment. But note that we also wouldn’t say that the CEO is praiseworthy for implementing the policy, precisely because he didn’t do it for the right reasons. (If he is responsible for implementing the policy, it is, again, only in a more morally neutral sense of responsibility, one that doesn’t involve praiseworthiness.) Thus, S (Praiseworthiness) doesn’t have the implication that the CEO is praiseworthy for helping the environment, which is the result we wanted.4 So far we have been looking at principle S and different reasons to think that it might need revision. Now let’s switch to N, the principle that gives necessary conditions for derivative responsibility. Here is again principle N: N: Responsibility for an outcome requires responsibility for behaving in a way that caused the outcome.

Like S, N is an intuitive principle: Roughly, it captures the idea that we can’t be morally responsible for an outcome unless we are “part of the reason” it happened, or unless we were somehow involved in its causal history. Imagine that you tried to cause some harm and the harm happened, but it didn’t happen because of what you did. In that case it seems that you cannot be morally responsible for the harm. Wanting or trying to be involved in the causing of harm isn’t enough to make you responsible for the harm; in addition, you actually must have been involved in the causing of the harm. This is the thought that N is supposed to capture. Despite its initial plausibility, N also faces some potential challenges. First, in principle it seems that we could be morally responsible for some non-​causal consequences of our acts. Perhaps causation is our first bridge to the world, but then other non-​causal notions of consequence can take over in such a way that we also end up being responsible for some non-​causal results of our acts. As an example of how this could happen, note that, according to some views of causation, outcomes that are overly “extrinsic” in nature cannot enter in causal relations (see, e.g., Lewis 1986a). So imagine that an assassin kills the parents of a child, thus turning the child into an orphan. The child’s becoming an orphan is a consequence of the assassin’s act, but, on those views of causation, it is not a causal consequence of the assassin’s act. For it is an overly extrinsic event: Its occurrence heavily depends on the obtaining of external circumstances (whether the child is an orphan is determined by what happens, not to him, but to his

4  Thanks to Tim Kearl for help with this paragraph. Knobe compares this case with another case where the CEO implements a policy motivated, again, just by the desire to make a profit, but foreseeing that the policy will have a harmful effect on the environment. In that case, the CEO seems blameworthy for the harm to the environment. So the contrast between these cases suggests that the conditions for blameworthiness and praiseworthiness differ in ways that track the agent’s intentions or motivations for action.

354   Carolina Sartorio parents). On these views, then, when the assassin shoots the parents, this has some causal consequences (the parents’ death) as well as some non-​causal consequences that follow from those (the child’s being an orphan). And, of course, in this case we want to blame the assassin for both consequences of his action. This suggests that we might have to revise N. Ultimately, whether it needs revision is something that will likely depend on what the true theory of causation (or the “causal relata”) is, in particular, it will depend on whether overly extrinsic events can be causal consequences of acts. Also, consider the fact that we can be morally responsible for the outcomes of our omissions, in addition to the outcomes of our positive acts. For example, a parent who fails to attend to her child’s needs can be responsible for the ensuing harm to her child. But some believe that omissions cannot be causes, for they think of omissions as absences (absences of actions of a certain kind), and they think that only (positive) events can be causes and effects (see, e.g., Dowe 2001 and Beebee 2004). If this is right, it follows that causation needn’t even be our first bridge to the world: In the case of omissions, the consequences of our omissions are never causal, and N would have to be revised accordingly to accommodate this fact. Others believe that omissions and absences in general can be causes (see, e.g., Schaffer 2000, Lewis 2004, and McGrath 2005). So, again, N’s truth will hinge on the relevant facts about causation and the causal relata. But it is important to note that, even those who believe that omissions cannot be causes typically tend to assume that there is an alternative relation or concept that can ground our moral responsibility in those cases. For example, Dowe (2001) introduces the concept of “quasi-​causation” that applies to omissions. Quasi-​causation is, essentially, merely possible causation, in that it makes reference to causal relations that obtain in other possible worlds. For instance, a mother’s neglecting her child may be a quasi-​cause of the harm to the child in the sense that, had she attended to her child’s needs, her doing so would have caused the child’s well-​being. With this notion of quasi-​causation in place, then, one could revise N to accommodate the outcomes of our omissions by requiring that the agent’s behavior be a cause or a quasi-​cause of the outcome. It is interesting to note, however, that, at least in the case of theories like Dowe’s, causation still plays an important role in grounding the moral responsibility of agents by omission. Even though actual causation isn’t what plays that role, possible causation is. (Another related proposal is Beebee’s suggestion that omissions can causally explain without being causes of what they explain. Her account also appeals to possible causal relations; see Beebee 2004.) Finally, even setting omissions aside, N may be open to counterexample. For instance, consider cases of (symmetric) overdetermination, such as firing squad cases where several shooters kill a victim and each of them would have been sufficient to bring about the death. According to some views, overdeterminers are not causes (for discussion, see, e.g., Lewis 1986b, Postscript E). But presumably we want to blame the shooters (all of them) for the outcome. Conversely, there are views according to which overdeterminers are causes (see, e.g., Schaffer 2003); in fact, this is probably the majority view nowadays. So how we approach this will, once again, depend on our views about causation. But, here too, it is important to note that even those who assume that overdeterminers aren’t

Responsibility and Causation    355 causes typically think that there is another way to link the agents to the outcome in overdetermination cases. The obvious alternative is to appeal to some form of collective causation. In particular, Lewis (1986b, Postscript E) argues that the mereological sum of the overdetermining events is a cause of the outcome in overdetermination cases. So, in the firing squad case, the mereological sum of all the shootings causes the victim’s death. If so, N could be revised by requiring that the agent be responsible for some cause of the outcome, which could be an individual behavior of hers or perhaps, in some cases, a more complex event involving her behavior and also the behavior of other agents. (Sartorio (2004 and 2015) argues that another form of collective causation, a disjunctive form, is required to deal with the responsibility of agents in other kinds of special scenarios. If so, this calls for a similar revision of N.) Putting this all together, N may have to be replaced by something like the following (much more complex but probably more accurate and less vulnerable to counterexample) principle: N*: Responsibility for an outcome requires responsibility for some cause or quasi-​ cause of the outcome, or for something that had the outcome as a non-​causal consequence.

In sum, we have seen that, although when it comes to derivative responsibility there is clearly an important connection between responsibility and causation, it is hard to say exactly what that connection is, or to express that connection in the form of completely general principles that take into account all the subtleties involved. In particular, as we have seen, in several instances the view that we should embrace on these issues will depend on what we take the relevant facts about the causal relation to be. Still, the conception of causation as, roughly, a ground of transmission of responsibility remains as a powerful idea, and one that seems basically right when it comes to responsibility for outcomes in the world.

3.  Basic Responsibility and Causation Now let us turn to the relation between causation and basic responsibility. As I anticipated in the introduction, one way in which causation matters to basic responsibility is that the causal histories of our acts arguably matter to whether we act freely, and thus to whether we are morally responsible for those acts. For example, at least part of what determines whether an agent made a choice freely is whether it was the causal result of coercion, compulsion, or an ordinary process of deliberation. But are actual causal histories the only thing that matters to freedom? This is less clear. According to the classical model of freedom, this is not all that matters, for something else is required in addition to a certain kind of causal history, namely, having alternative possibilities of action. The thought is intuitive enough: we cannot act freely unless there

356   Carolina Sartorio is more than one way we could have acted. Thus, according to this view, in order to determine whether an act is free, we must look partly outside the actual causal history of events and into facts that pertain to whether the agent had the ability to do otherwise. On some views, this is understood in terms of purely counterfactual possibilities that were still within the agent’s reach. But this classical view of freedom has been forcefully contested in recent years, since the publication of a highly influential paper by Frankfurt (1969). Although Frankfurt’s main goal in that paper was to undermine the view that freedom requires alternative possibilities, some of the ideas put forth there also help motivate a simpler conception of freedom, one according to which freedom is just a function of actual causal histories. According to this view, all it takes for an act to be free is to be caused in a certain kind of way. As a result, it is natural to expect that causation will play a particularly central role in this conception of freedom (more on this later). The view that freedom is exclusively a function of actual causal histories is supported by the same examples that Frankfurt used to undermine the idea that freedom requires alternative possibilities, which have since then been known as “Frankfurt-​style” cases. Here is one of Frankfurt’s own examples: Jones and the Neuroscientist: A neuroscientist wants Jones to perform a certain action. He is prepared to go to considerable lengths to get his way, but he prefers to avoid showing his hand unnecessarily. So he waits until Jones is about to make up his mind what to do, and he does nothing unless it is clear to him (he is an excellent judge of such things) that Jones is going to decide to do something other than what he wants him to do. If it were to become clear that Jones is going to decide to do something else, the neuroscientist would take effective steps to ensure that Jones decides to do what he wants him to do, by directly manipulating the relevant processes in Jones’s brain. As it turns out, he never has to show his hand because Jones, for reasons of his own, decides to perform the very action he wants him to perform. (Frankfurt 1969)

As Frankfurt noted, Jones seems to act freely and be responsible for his act, although, given the presence of the neuroscientist, he lacked alternative possibilities. So examples of this kind can be used to argue against the idea that freedom requires alternative possibilities. (Although whether they are actually successful in undermining the classical conception of freedom is still a matter of controversy. The literature on Frankfurt-​style cases is now huge; a good resource is the collection of papers in Widerker and McKenna 2003.) At the same time, cases of this kind can be used to motivate a different picture of freedom, one according to which all that matters to freedom are the actual causal histories. Briefly, here is why (for a more extended discussion of this point, see Sartorio 2017). As Frankfurt noted, these examples illustrate an important distinction: The distinction between the factors that make it the case that the agent couldn’t have done otherwise, on the one hand, and the factors that explain why the agent acted as he did, on the other. I’ll use the labels “inevitability factors” to refer to factors of the first kind and “explanatory factors” to refer to factors of the second kind. Frankfurt thought that examples like Jones and the Neuroscientist help us see that the only factors that

Responsibility and Causation    357 are relevant to freedom are the explanatory factors. First, note that in a case like Jones and the Neuroscientist the inevitability factors and the explanatory factors come apart. The explanatory factors are the factors involved in Jones’s own process of deliberation that resulted in his choice, and the inevitability factors are the factors concerning the neuroscientist’s existence that make it the case that Jones couldn’t have made any other choice. Second, given that the neuroscientist never intervenes in the actual course of events, many of us tend to regard his presence as simply irrelevant to Jones’s responsibility for his choice. Intuitively, Jones seems just as responsible as he would have been if the neuroscientist hadn’t been present, given that he made his choice completely on his own, on the basis of his own reasons. So, when the inevitability factors and explanatory factors come apart, we see that only the explanatory factors are relevant to the agent’s freedom. In contrast, in typical cases where an agent is unable to do otherwise and is intuitively not responsible for what he does, the inevitability factors coincide with the explanatory factors. For example, in a case where an agent is coerced to act in a certain way, the coercive threat is what makes the act inevitable, but it is also what explains why the agent did what he did. As a result, in these cases it is not immediately obvious what accounts for the agent’s lack of responsibility, if it is the lack of alternatives or just the particular kind of causal history that the act has. For this reason, Frankfurt thinks, cases like Jones and the Neuroscientist put us in a better position to decide this issue (and, on his view, they support the idea that only the explanatory factors are relevant). Of course, this is not the place to try to decide whether Frankfurt is right about this. So my goal in what follows will be a much more modest one. I will look more closely into this alternative picture of freedom (the one inspired by Frankfurt-​style cases and motivated by Frankfurt’s remarks on the distinction between inevitability factors and explanatory factors), and I will examine a number of important questions that arise for that view. Given the topic of this chapter, I am particularly interested in discussing the role that causation plays in a view of this kind. The first question that arises for the view is this: How, exactly, should we understand the claim that only the explanatory factors are relevant to freedom? At first blush, the most natural way to understand it seems to be this: Actual Causes: Only facts about the actual causes of an act (or some subset of those actual causes) are relevant to whether the act was done freely.

That is to say, Actual Causes claims that, when trying to determine whether an act was done freely, all we have to look at are those factors that were part of the causal history of the act. Recall that our focus now is basic responsibility. So imagine that an agent is basically responsible for the act of making a certain choice (a mental act) at a certain time. Then Actual Causes says that what accounts for the fact that the agent made that choice freely is only facts concerning the factors that actually caused the agent’s choice. (Or, more precisely, some relevant subset of those factors. This qualification is warranted because acts, like all other events, have a great number of causes, and many of them

358   Carolina Sartorio probably won’t be relevant to whether the agent acted freely. I’ll leave this qualification implicit in what follows.) Actual Causes is a very simple and elegant view of freedom. Again, it is motivated by a natural way of understanding the thought that only the explanatory factors, the factors that actually explain an agent’s behavior, are relevant to the agent’s freedom. Causation plays a central role in a view of this kind. For note that what Actual Causes says is that freedom is exclusively grounded in certain causal facts. For example, what grounds the fact that an agent made a choice freely is the fact that the choice was caused by a process of a certain kind, say, an ordinary process of deliberation. Notice that this isn’t just to say that there was such a process, but, more precisely, that such a process causally resulted in the agent’s choice. Thus, on this view, causation plays an important role in that only causal facts (facts of the kind “C caused E”) ground the freedom of agents. However, Actual Causes may need some tweaking. First, imagine that one believes, again, that omissions cannot enter in causal relations (because one believes that omissions are absences and that absences in general cannot enter in causal relations, as discussed in the previous section). Imagine that one also believes that we can be basically responsible for some of our omissions, for example, for not making a certain choice at a certain time (the omission of a mental act). If omissions cannot enter in causal relations, then this means that omissions don’t have causal histories, and thus Actual Causes entails that nothing is ever relevant to whether we freely omit to act. So, depending, again, on one’s view about the causal powers of omissions, one might need to revise Actual Causes in order to accommodate basic responsibility for omissions. And the natural solution to this problem would be, again, to broaden the view in whichever way is needed to accommodate omissions, by incorporating concepts that one’s metaphysics of omissions allows for and that can plausibly ground the agent’s moral responsibility in those cases (such as the concept of quasi-​causation discussed in the previous section, or the idea that omissions can be causally explanatory without themselves being causes). But there is a second reason to think that Actual Causes needs tweaking. It is the fact that there could be factors that help determine the constitution of the actual causal history in each case but that are not, themselves, part of the actual causal history (Sartorio 2011 and 2016). Unless one believes that causation is a primitive metaphysical relation, there will always be other more metaphysically basic facts that ground the relevant causal facts, and that are not themselves part of the actual causal history. Still, given that those facts help determine the constitution of the causal history, and given that the causal history is relevant to the agent’s freedom, those facts seem to be indirectly relevant to the agent’s freedom (they are relevant to the agent’s freedom, but only given the role they play in determining the constitution of the causal history). However, as it stands, Actual Causes doesn’t allow for the relevance of any such facts. For example, imagine that Ann is basically responsible for making the choice to lie to Bert at a certain time. Imagine that Ann’s choice was caused by an ordinary process of deliberation, one that included weighing the reasons for and against lying to Bert, etc. On the view that we are considering, the fact that Ann’s choice was caused in that way (let’s call this fact the “causal history fact”) accounts for the fact that it was made freely

Responsibility and Causation    359 (the “freedom fact”). But imagine that one also believes in a reductive theory of causation, say, a counterfactual theory of causation, according to which causation reduces to counterfactual dependencies of some kind—​dependencies of the sort “if event X hadn’t occurred, then event Y also wouldn’t have occurred.”5 On a view of this kind, the causal history fact in question is in turn grounded in a “counterfactual dependence fact” (a fact involving the relevant relations of counterfactual dependence), as represented by the following diagram: Counterfactual dependence fact → Causal history fact → Freedom fact As a result, on this picture, the freedom fact seems to be grounded in both the causal history fact and the counterfactual dependence fact. For it is implausible to suggest that only the causal history fact grounds the freedom fact, if the causal history fact is in turn grounded in the counterfactual dependence fact.6 So this means that Actual Causes might need to be revised to allow for the relevance of the grounding facts. But, once again, we see that whether it needs to be revised, and how we should revise it, ultimately depends on the relevant metaphysical facts about causation. The preceding discussion suggests that the original principle, Actual Causes, might need to be replaced by a more complex version, perhaps one that looks like this: Actual Causes*: Only facts about the actual causes or quasi-​causes of an act, and the grounds of those facts, are relevant to whether the act was done freely.

Now, at this point one might have a different kind of concern. One might worry that the resulting view of freedom will end up collapsing into a much more liberal model where all sorts of facts, not just facts about the actual causal history, play a role in grounding freedom. Perhaps it will even end up collapsing back into the classical model, according to which freedom is grounded in alternative possibilities! This would be an odd and unexpected result, since the two models are supposed to be rival conceptions of freedom. But this worry is unfounded. According to Actual Causes*, the only facts that are relevant to freedom are still those facts that, directly or indirectly, concern the actual causal history. This means that all other facts are irrelevant to freedom, and this is likely to include facts about alternative possibilities. Still, it is interesting to note that it is in principle possible for a view to respect Actual Causes* and to also be an alternative-​ possibilities view (see Sartorio 2016, ch. 1). What would have to happen for this to be the case is that the same facts that ground the causal history facts also make it the case that

5 

It is worth noting that no plausible view of causation will identify causation with simple counterfactual dependence, because the resulting view is clearly open to counterexamples involving overdetermination, preemption, etc. But there are other more sophisticated and more plausible views out there that still reduce causation to facts about counterfactual dependence; for discussion, see Paul and Hall 2013. 6  Although some think that there are exceptions to the transitivity of grounding, in typical cases transitivity holds. For discussion of the grounding relation and its properties, see, e.g., Raven 2015.

360   Carolina Sartorio the agent has or lacks alternative possibilities. But notice that any such view would be an extremely watered-​down version of the alternative-​possibilities view. It would be a view according to which alternative possibilities are relevant to freedom, but only because they help determine what the actual causal history of events is. Most likely, a view of this kind will rely on some implausible assumptions about the metaphysics of causation. At any rate, this is not the kind of view that people typically have in mind when they think that freedom requires alternative possibilities of action: They don’t typically have in mind facts that determine what the actual explanation of action is, but more robust facts that make it the case that the agent has genuine access to alternatives. Finally, one might worry that Actual Causes* is too simple in that facts pertaining to actual causal histories aren’t rich enough, by themselves, to ground an agent’s freedom. For example, it is common to suggest that an agent’s freedom has to do with general abilities that can remain unexercised in the actual scenario, such as certain capacities for responsiveness to reasons. To the extent that those capacities are actually unexercised and are not relevant to the constitution of the actual causal history, Actual Causes* would in fact entail that they are irrelevant to the agent’s freedom. In particular, it is interesting to note that most views that were inspired by Frankfurt’s argument against the principle of alternative possibilities focus on some factors that don’t pertain to the actual causal history. Notably, Fischer and Ravizza (1998) focus on certain modal properties of the actual causal mechanisms (the “responsiveness” of mechanisms), and a central thesis of their view is the claim that these properties are relevant to the agent’s freedom despite their not being exercised in the actual scenario. (See also the view developed by McKenna (2013), which appeals to unexercised dispositions of agents.) Fischer and Ravizza seem to think that without appealing to something like the responsiveness of mechanisms we cannot fully capture everything that is involved in an agent’s acting freely, in particular, we cannot capture the thought that a free agent is someone who is suitably responsive to reasons when he acts. Moreover, they argue that this is still within the framework of a conception of freedom that rejects the relevance of alternative possibilities, because a mechanism being suitably responsive doesn’t require the existence of genuine alternative possibilities of action. Now, although this may be true, the focus on factors that don’t pertain to the actual causal history seems to be in tension with the initial motivation for the model of freedom based on actual causes. Frankfurt’s insight, recall, was that only facts concerning the actually explanatory factors are relevant to an agent’s freedom. Thus, in suggesting that other kinds of factors play a central role in grounding freedom, these views fail to respect that fundamental insight.7

7  Interestingly, Frankfurt’s own positive view (Frankfurt 1971) may end up having the same problem, at least on some interpretations of the view. Roughly, the view is that being morally responsible requires us to have a will with a certain hierarchical structure, one where our higher-​order desires align with our causally efficacious first-​order desires. But nothing in this view seems to require that those higher-​order desires themselves also be causally efficacious. If so, this is in tension with the idea that only causally efficacious factors are relevant to freedom. (For discussion of this point, see Sartorio 2016, ch. 1, sec. 3.2.2.).

Responsibility and Causation    361 Alternatively, Sartorio (2016, ch. 4) argues that it is, in fact, possible to capture the idea that free agents are agents who are sufficiently responsive to reasons within a framework that respects the fundamental insight by Frankfurt. The key lies in realizing that causal histories are much richer than they seem to be at first sight. In particular, when agents act freely, the causal histories of their acts reflect their responsiveness to reasons because they contain, in addition to the reasons that motivate them to act, several absences of reasons that also explain their behavior. Intuitively, this is supposed to capture the fact that when we act we are not just responding to what is the case on the relevant occasion, but also to what is not the case on that occasion. All of these factors are part of the actual causal history or the actual explanation of the agent’s behavior. On this view, even a very simple act such as my deciding to go for a walk one morning has a complex causal history, when it’s done freely. The causal history includes the reasons that moved me to go for a walk (say, the desire to exercise), but also the absence of reasons that would have moved me to stay home (say, the belief in an imminent snowstorm). (Although, here too, if absences cannot be causes, the claim about the role played by absences of reasons would have to be cashed out in different metaphysical terms. Once again, we see the relevance of the metaphysics of causation; for discussion of this point, see Sartorio 2016, ch. 2.)

References Beebee, H. (2004). “Causing and Nothingness.” In J. Collins, N. Hall, and L.A. Paul (eds.), Causation and Counterfactuals. Cambridge, MA: MIT Press, 291–​308. Capes, J. (2012). “Blameworthiness Without Wrongdoing.” Pacific Philosophical Quarterly 93(3): 417–​437. Clarke, R. (2014). Omissions: Agency, Metaphysics, and Responsibility. New York: Oxford University Press. Dowe, P. (2001). “A Counterfactual Theory of Prevention and ‘Causation’ by Omission.” Australasian Journal of Philosophy 79: 216–​226. Feinberg, J. (1970). “Collective Responsibility.” In Doing and Deserving: Essays in the Theory of Responsibility. Princeton, NJ: Princeton University Press, 222–​251. Fischer, J., and M. Ravizza (1998). Responsibility and Control. Cambridge: Cambridge University Press. Frankfurt, H. (1969). “Alternate Possibilities and Moral Responsibility.” Journal of Philosophy 66(23): 829–​839. Frankfurt, H. (1971). “Freedom of the Will and the Concept of a Person.” Journal of Philosophy 68: 5–​20. Graham, P. (2011). “Fischer on Blameworthiness and ‘Ought’ Implies ‘Can.’ ” Social Theory and Practice 37(1): 63–​80. Haji, I. (1998). Moral Appraisability: Puzzles, Proposals and Perplexities. New York: Oxford University Press. Knobe, J. (2003). “Intentional Action and Side Effects in Ordinary Language.” Analysis 63: 190–​193. Lewis, D. (1986a). “Events.” In Philosophical Papers, Volume II. New York: Oxford University Press, 241–​269.

362   Carolina Sartorio Lewis, D. (1986b). “Causation.” In Philosophical Papers, Volume II. New York: Oxford University Press, 159–​213. Lewis, D. (2004). “Void and Object.” In J. Collins, N. Hall, and L.A. Paul (eds.), Causation and Counterfactuals. Cambridge, MA: MIT Press, 277–​290. McGrath, S. (2005). “Causation by Omission: A Dilemma.” Philosophical Studies 123: 125–​148. McKenna, M. (2013). “Reasons-​Responsiveness, Agents and Mechanisms.” In D. Shoemaker (ed.), Oxford Studies in Agency and Responsibility, Volume 1. Oxford: Oxford University Press, 151–​183. Nelkin, D., and S. Rickless (2017). The Ethics and Law of Omissions. New York: Oxford University Press. Paul, L. A., and N. Hall (2013). Causation: A User’s Guide. Oxford: Oxford University Press. Raven, M. (2015). “Ground.” Philosophy Compass 10(5): 322–​333. Sartorio, C. (2004). “How to Be Responsible for Something Without Causing It.” Philosophical Perspectives 18(1): 315–​336. Sartorio, C. (2007). “Causation and Responsibility.” Philosophy Compass 2(5): 749–​765. Sartorio, C. (2011). Actuality and Responsibility. Mind 120 (480):1071–​1097. Sartorio, C. (2015). “Resultant Luck and the Thirsty Traveler.” Methode 4(6): 153–​172. Sartorio, C. (2016). Causation and Free Will. Oxford: Oxford University Press. Sartorio, C. (2017). “Frankfurt-​Style Examples.” In M. Griffin, N. Levy, and K. Timpe (eds.), Routledge Companion to Free Will. London: Routledge, 179–​190. Schaffer, J. (2000). “Causation by Disconnection.” Philosophy of Science 67: 285–​300. Schaffer, J. (2003). “Overdetermining Causes.” Philosophical Studies 114(1–​2): 23–​45. Sher, G. (2009). Who Knew? Responsibility Without Awareness. New York: Oxford University Press. Widerker, D., and M. McKenna, eds. (2003). Moral Responsibility and Alternate Possibilities: Essays on the Importance of Alternative Possibilities. Aldershot: Ashgate. Zimmerman, M. (1986). “Negligence and Moral Responsibility.” Noûs 20: 199–​218. Zimmerman, M. (1997). “A Plea for Accuses.” American Philosophical Quarterly 34(2): 229–​243.

PA RT V I I I

R E SP ON SI B I L I T Y, L AW, A N D J U ST IC E

CHAPTER 18

Resp onsib i l i t y, Pu nishm en t, a nd Pred omi na nt Retribu ti v i sm David O. Brink*

Punishment combines blame and sanction, censuring wrongdoers for failing to comply with authoritative legal norms and attaching penalties of various kinds to especially significant forms of wrongdoing. There are many possible rationales for punishment. Some are forward-​looking, providing a consequentialist justification of punishment in terms of the goods—​such as deterrence, rehabilitation, the expression and enforcement of community norms, or reconciliation—​that punishment can produce. Other rationales are backward-​looking, providing a retributivist justification of punishment in terms of the agent’s desert, which is itself grounded in the degree of her culpable wrongdoing. Blame is an aversive reactive attitude that presupposes that its target is blameworthy—​that the target is culpable or responsible for some kind of wrongdoing. Insofar as punishment presupposes blame and blame presupposes responsibility and desert, punishment also presupposes responsibility and desert. If so, retributivism must be at least part of the truth about punishment. Some form of retributivism is defensible, once we clarify misconceptions about retributivism. This leaves a choice between pure retributivism and a mixed conception that includes retributive elements. That choice is best resolved in favor of a mixed conception in which retributive elements predominate.

*  I would like to thank audiences at the University Texas School of Law and the Chinese University of Hong Kong and especially John Deigh, Hon Lam Li, Dana Nelkin, Derk Pereboom, and Lawrence Sager for helpful feedback on earlier versions.

366   David O. Brink

1.  Blame, Punishment, and Responsibility Punishment combines blame or censure and sanction of some kind. Blame is an aversive reactive attitude that presupposes that its target is blameworthy—​that the target is culpable or responsible for some kind of wrongdoing, whether that is a shortcoming in action or attitude (Brink and Nelkin 2022). Punishment has a condemnatory function in which breach of norms makes one a legitimate target of censure. If so, punishment involves blame and so presupposes that the target of punishment is blameworthy. Blameworthiness consists of some sort of wrongdoing for which the agent is responsible. But punishment involves sanction, as well as censure. Punishment goes beyond blame and attaches sanctions of various kinds to especially significant forms of culpable wrongdoing. Punishment can be informal when sanctions are imposed by private parties, either as individuals or as groups, to perceived culpability. Punishment is official when the state has a system of sanctions that it imposes for culpable violations of its rules. Legal punishment is official punishment in which a legal system systematically imposes sanctions for culpable violations of legal rules by its members. Our focus will be on legal punishment. Why should the state sanction and not merely censure legal wrongdoing? We do well to remember that sanctions can include a diverse array of restrictions on the rights and privileges of citizenship, ranging from fines, probationary status, community service, to incarceration. A reasonably just criminal justice system specifies both crimes mala in se—​wrongs in themselves—​and mala prohibita—​wrongs by virtue of prohibition. Crimes mala in se violate natural duties and include wrongs against the person, such as murder, rape, assault, theft, and fraud. Crimes mala prohibita do not violate natural duties and can include crimes of prostitution, drug use, gambling, tax evasion, traffic violations, loitering, and truancy. Because they involve crimes against the person that violate natural duties, crimes mala in se tend to be more serious offenses than crimes mala prohibita, which serve more to facilitate security, cooperation, and coordination. Sanctions are needed as a supplement to censure to ensure adequate levels of compliance with the law’s prohibition on crimes mala in se and mala prohibita. In a reasonably just legal system citizens have good moral reasons to avoid crimes mala in se and mala prohibita. But moral reasons are not always sufficient to ensure adequate levels of compliance. Some individuals lack moral motivation entirely, for others moral motivation is weak, and those for whom moral motivation is usually strong are nonetheless subject to occasional weakness. A system of legal sanctions provides imperfect moral beings with prudential motivation that supplements and reinforces, rather than replaces, moral reasons for compliance (von Hirsch 1993: ch. 2). In the case of legal punishment, the state claims authority to censure and sanction for culpable violation of duly enacted norms specifying crimes mala in se and mala prohibita. As such, punishment can be contrasted with other forms of social

Responsibility, Punishment, and Predominant Retributivism    367 control that involve detention—​both prospective detention in which an agent is civilly committed or quarantined, prior to wrongdoing, because he is a danger to himself or others, and reactive detention in which the agent has engaged in wrongdoing for which he is not culpable and, hence, is excused, but nonetheless poses an ongoing danger to himself or others. Detention, of either kind, involves neither blame or sanction. Moreover, punishment can also be contrasted with forms of social control that do employ sanctions but not in response to culpable wrongdoing. The state might employ sanctions to control the attitudes and behavior of citizens in legal or extralegal ways, for instance, to encourage loyalty and discourage legitimate dissent. What seems crucial to punishment is that it is a public response to a perceived breach of legally promulgated norms. Theories of punishment answer various questions about the justification of punishment, including (1) Whom should we punish? (2) Why should we punish? (3) How much should we punish? and (4) How should we punish? A satisfactory theory of punishment should provide an answer to all these questions (and perhaps more). How we answer one question may affect how we can answer other questions. Different rationales for punishment approach these issues differently, sometimes conceptualizing crime differently and focusing on different aims of punishment. 1. Deterrence: Punishment aims to promote compliance with legal norms, and it does so by achieving deterrence, both specific and general—​deterring the individual who is punished from further wrongdoing and other potential criminal wrongdoers. 2. Rehabilitation: Criminal conduct is an expression of social dysfunction, and punishment should aim to rehabilitate and re-​socialize criminals, equipping them with psychological and social skills and other resources for being well-​integrated and productive members of society. 3. Norm reinforcement and communication: Because criminal conduct breaches social norms, punishment serves to express, reinforce, and communicate community norms. 4. Restorative justice: Criminal conduct breaches community norms and often wrongs a victim, in which case punishment should provide a mediation among the wrongdoers, the victims, and the larger community with the aim of restoring moral community and securing reconciliation. 5. Retributive justice: Criminal misconduct involves culpable wrongdoing—​that is, wrongdoing for which the agent is responsible—​and punishment should aim to hold criminals accountable based on their desert and in proportion to the magnitude of their culpable wrongdoing. These rationales are not exhaustive, but they are representative. Notice that rationales for punishment can be pure, appealing to a single factor or aim to justify punishment, or mixed, if they appeal to combinations of factors. Notice also that the first four rationales are broadly forward-​looking, appealing to some good consequence of punishment,

368   David O. Brink whereas the last rationale is backward-​looking, appealing to the historical notion of desert. Here too, conceptions of punishment can be purebreds or hybrids.1 We can test rival rationales by assessing their systematic comparative plausibility. In particular, we should see how well they answer these different questions about justified punishment, in particular, by measuring their commitments on these questions against our own considered convictions about these matters, and by comparing their theoretical virtues. Of special concern, when assessing the adequacy of any rationale, is whether its account of whom to punish is under-​inclusive—​failing to justify punishment of those who seem to merit punishment—​or over-​inclusive—​justifying punishment of those who do not seem to merit punishment. A conception or rationale for punishment is defective insofar as it appears either under-​inclusive or over-​inclusive. The English jurist William Blackstone (1723–​1780) famously claimed that it is worse to punish the innocent than to let the guilty go free.2 Blackstone’s asymmetry implies that all else being equal it is worse to over-​punish than to under-​punish and, hence, that all else being equal it is a bigger vice in a theory of punishment to be over-​inclusive than to be under-​inclusive. Most normative conceptions combine accommodation of our pre-​theoretical normative commitments and reform of those commitments. Insofar as a conception subsumes and explains considered normative convictions, it accommodates them. But insofar as a conception conflicts with considered normative convictions, it calls for reform or revision. We should not expect perfect accommodation, because our moral judgments are incomplete, inconsistent, and subject to various kinds of bias and distortion. But complete reform is no more plausible than complete accommodation. Indeed, complete reform threatens to introduce a change in subject matter and should be explored only as a last resort, only if we conclude that our considered normative convictions are hopelessly muddled. Typically, reform is like the hole in a doughnut, made possible by a surrounding substance of accommodation. We accept local reforms as part of a process of global accommodation. This means that reform is always, or at least typically, partial. But to say that reform is partial does not imply that it cannot be significant. How revisionary a conception can be is something we cannot decide in advance of looking at the conception, its degree of accommodation, and the nature of its reforms. So too with blame and punishment. Our pre-​theoretical beliefs about blame and punishment are not sacred; they may be biased, incomplete, or inconsistent and may require revision. So, while it is some evidence against a conception that it is under-​inclusive or 1 

If, as I believe, restorative justice is most plausible when it incorporates, rather than eliminates, accountability, then restorative justice is best understood as a mixed conception, incorporating both a backward-​looking (retributive) focus on desert and a forward-​looking focus on reconciliation. For this sort of treatment of restorative justice, see Duff 2001: 92–​106; Duff 2009; and Allais 2012. 2  Blackstone’s version of this asymmetry involves a 10:1 ratio: “better that ten guilty persons go free than that one innocent party suffer” (Blackstone 1791: Book IV, Chapter 27, page 358). Benjamin Franklin thought that the ratio is 100:1 (Franklin 1785: Vol. 9, page 293). It is the asymmetry itself, rather than any particular ratio, that is important for our purposes.

Responsibility, Punishment, and Predominant Retributivism    369 over-​inclusive, that evidence may not be decisive. The most compelling conception of punishment may be revisionary in some respects if it fits our other beliefs sufficiently well. By these standards, we have reason to be skeptical of purely forward-​looking conceptions of punishment, because they promise to be both under-​inclusive and over-​inclusive. Purely forward-​looking conceptions will be under-​inclusive, because they imply that we have no reason to punish when doing so would not secure the relevant good consequences. No doubt, punishment often does or at least can promote deterrence, rehabilitation, norm reinforcement and communication, and reconciliation. But we can imagine circumstances in which punishment would not secure these goods—​in which rehabilitation was not possible and in which punishment could not be publicized or would be widely, but falsely, regarded as illegitimate. In such circumstances, someone who was fully responsible for a serious wrongdoing would go unpunished. Nonetheless, most of us would think that there was at least some reason to punish culpable wrongdoers in such circumstances. These would be cases of under-​ punishment. Moreover, purely forward-​looking conceptions will be over-​inclusive, because they imply that we have reason to punish in the absence of culpable wrongdoing if doing so would promote the relevant consequential values. Perhaps such cases would be rare or merely hypothetical, but they are problematic because they violate the norm that we should punish only culpable wrongdoers. In such cases, purely forward-​ looking conceptions of punishment are over-​inclusive. If we accept Blackstone’s asymmetry, then, all else being equal, the sin of over-​inclusiveness is more serious than the sin of under-​inclusiveness. P. F. Strawson criticizes instrumentalist conceptions of responsibility that justify ascriptions of responsibility and the reactive attitudes in purely forward-​looking terms, by appeal to deterrence, norm reinforcement, and the like for invoking the wrong sort of reason for thinking someone responsible and applying the reactive attitudes (1962: 61–​62). Adapting Strawson’s point, we might say that purely forward-​ looking conceptions of punishment provide the wrong sort of reason for punishment. We should not punish for purely forward-​looking reasons, in the absence of culpable wrongdoing, because the proper target of punishment is wrongdoing for which the agent is responsible. This is the claim that culpable wrongdoing is a necessary condition of justified punishment. This is an important insight of retributive conceptions, whether pure or mixed. Perhaps we don’t yet have a conclusive refutation of purely forward-​ looking conceptions of punishment. Every conception of punishment might prove revisionary, in which case we would need to determine the least revisionary conception, and until we’ve examined the implications of other conceptions, we don’t know that some purely forward-​looking conception won’t be least revisionary. This is already reason to examine other conceptions, but there is reason to expect that retributive rivals, whether pure or mixed, will be less revisionary and, hence, more plausible.

370   David O. Brink

2.  Skepticism About Responsibility and Punishment Retributive conceptions of punishment, whether pure or mixed, limit punishment to cases of culpable wrongdoing. Culpable wrongdoing is wrongdoing for which the agent is responsible. This means that legitimate punishment is conditional on the agent’s responsibility for her wrongdoing. This, in turn, means that punishment is hostage to realism about responsibility and is threatened by skepticism about responsibility. In this way, punishment is tied to traditional debates about whether we are ever responsible for our actions. That debate is often framed in terms of the compatibility of free will and responsibility with the truth of causal determinism—​the thesis that every event is determined or necessitated by the laws of nature and the prior history of the world. Compatibilists insist that responsibility is compatible with the truth of causal determinism, whereas incompatibilists insist that the two are not compatible. That means that incompatibilists who are also determinists are hard determinists, embracing skepticism about free will and responsibility. However, some incompatibilists are libertarians, rejecting determinism and embracing indeterminism. Most libertarians are realists about free will. But it is an open question whether indeterminism can ground responsibility and whether we could be responsible for uncaused choices. If not, then indeterminism cannot rescue realism about freedom and responsibility. In that case, incompatibilism would be sufficient for skepticism about free will and responsibility. Skeptics about responsibility have reason to be skeptical about punishment. They could still accept purely forward-​looking conceptions of punishment—​defending punishment because it does or can promote various goods, such as deterrence, rehabilitation, norm enforcement and communication, and reconciliation—​because they do not presuppose that the agent has done wrong or is responsible for it. But punishment involves both censure and sanction, which are attitudes and practices that seem to presuppose wrongdoing for which the target is responsible. This is why purely forward-​ looking conceptions seem to appeal to the wrong sort of reason for punishment. If we accept this link between legitimate punishment and responsibility, then skeptics about responsibility should be abolitionists about punishment. Nonetheless, they can seek to replace punishment with forms of social control that do not presuppose censure, sanction, and responsibility. In the criminal law, we acquit wrongdoers who are not responsible for their wrongdoing by reason of insanity. But if they remain a danger to themselves or others, they can be subject to involuntary commitment, during which they should receive psychiatric treatment designed to render them non-​dangerous and make them suitable for reintegration into society. Involuntary commitment is a form of reactive detention. We engage in preemptive detention when we quarantine people who are potential dangers to themselves or others, even in the absence of having caused harm. The skeptic about responsibility might be an abolitionist about punishment but a

Responsibility, Punishment, and Predominant Retributivism    371 proponent of involuntary commitment and quarantine (e.g., Pereboom 2014: chs. 6–​7; Carusso 2018, 2019).

3.  Realism About the Reactive Attitudes and Responsibility Skepticism about free will and responsibility is an intelligible position and, in another context, merits extended engagement. But skepticism about free will and responsibility is a position of last resort, to be embraced only if there is no plausible conception of responsibility. Skepticism about responsibility is extremely revisionary insofar as it denies that we are ever responsible for anything. This makes it an error theory, claiming that our assumptions about responsibility are systematically mistaken. To justify an error theory, the principles that entail skepticism must be more plausible than the assumptions that they require us to forego. Especially when the assumptions in question are widely held and deeply entrenched in our practices, this burden is difficult to discharge. The assumption that normal, mature adults are, for the most part, responsible for their conduct is a pervasive one. As Strawson reminds us, this assumption is implicated in many of our reactive attitudes, especially those involving praise and blame, such as pride and resentment, and it underlies our practices of punishment. Though these assumptions and practices could be systematically mistaken, we should accept this conclusion reluctantly, only if we can identify no viable conceptions of responsibility. One way to identify a viable conception of responsibility is to look to our practices that presuppose responsibility. Strawson proposed that we do so by looking at the reactive attitudes and the way they mediate interpersonal relationships. Some reactive attitudes, such as agent-​regret, need not implicate responsibility. But our reactive attitudes involving praise and blame—​such as pride, resentment, and guilt—​do involve responsibility. Strawson links responsibility and reactive attitudes, such as those of resentment and indignation, in a biconditional fashion. Strawson’s Thesis: Reactive attitudes involving blame and praise are appropriate just in case the targets of these attitudes are responsible.

Strawson’s thesis can be interpreted in two very different ways, depending on which half of the biconditional has explanatory priority. According to a response-​dependent interpretation, there is no external, or response-​ independent justification of our attributions of responsibility. This reading fits with Strawson’s view that our reactive attitudes and ascriptions of responsibility, as a whole, do not admit of external justification. Particular expressions of a reactive attitude might be corrigible as inconsistent with a pattern of response, but the patterns of response are not themselves corrigible in light of any other standard. Similarly, particular ascriptions

372   David O. Brink of responsibility might be corrigible in light of patterns in our ascriptions of responsibility, but the patterns themselves are not corrigible in light of any other standards. Responsibility judgments simply reflect those dispositions to respond to others that are constitutive of various kinds of interpersonal relationships. Response-​dependence might be the right interpretation of Strawson (Watson 1987: 222; Wallace 1994: 19; and Shoemaker 2017). However, it is an open question whether it is the philosophically most plausible interpretation of that thesis. An alternative interpretation of Strawson’s thesis is realist, rather than response-​ dependent. This interpretation stresses the way that the reactive attitudes make sense in light of and so presuppose responsibility. As such, the reactive attitudes are evidence about when to hold people responsible, but not something that constitutes them being responsible. It’s true that the reactive attitudes are appropriate if and only if the targets are responsible, but it’s the responsibility of the targets that makes the reactive attitudes toward them fitting or appropriate. In the biconditional relationship between responsibility and the reactive attitudes, it is responsibility that is explanatorily prior, according to this realist interpretation. This involves a response-​independent conception of responsibility. The crucial question in deciding between these two different interpretations of Strawson’s thesis is whether the study of excuses leads to a conception of responsibility that can be specified independently of our reactive attitudes and that serves to ground or justify those attitudes. If so, this favors a realist conception of responsibility. If not, it favors a response-​dependent conception. The realist interpretation of Strawson’s thesis promises to ground and justify the reactive attitudes in a way in which the response-​dependent reading cannot. If we focus on wrongdoing, the realist can and should appeal to the following working hypothesis about the inverse relation between responsibility and excuse. The Responsibility-​Excuse Biconditional: An agent is responsible for wrongdoing if and only if she is not excused for it.

According to this biconditional, an agent is responsible for that for which she has no excuse, and she is excused for conduct for which she is not responsible. If so, excuse is a window onto responsibility and vice versa.3 Strawson accepts the Responsibility-​ Excuse Biconditional and takes it to support a form of compatibilism. Excusing and exempting conditions depend on specific failings—​sometimes in individual cases, for instance, involving inadvertence or duress, and sometimes of a general nature, involving immaturity or insanity. As such, they don’t presuppose the falsity of determinism, which is a thesis that applies to all cases (1962: 67–​69). If hard determinism were true, we should excuse everyone, because no one is

3 

Michael Moore describes excuse as “the royal road” to responsibility (Moore 1997: 548). But it’s important to recognize that it is a two-​way street.

Responsibility, Punishment, and Predominant Retributivism    373 responsible. But the criminal law and our moralized reactive attitudes are selective about when to excuse. We excuse for particular kinds of incapacity, such as insanity, immaturity, or duress. But then our reactive attitudes and associated practices of excuse presuppose some version of compatibilism.

4.  Realism and Fair Opportunity If we model responsibility by attending to excuses, we would do well to look at the criminal law, because it has a reasonably settled conception of excuses. The criminal law recognizes two main kinds of excuse—​incompetence excuses, such as insanity, and duress excuses. Incompetence involves impairment of an agent’s internal capacities. By contrast, duress involves a situational failing in which wrongful interference by another with the agent’s options deprives her of the fair opportunity to act on her own deliberations about how to behave. In this way, standard excuses recognized by the criminal law reflect impairment of the agent’s capacities or opportunities. If we treat excuse as a window onto responsibility, we might factor responsibility and, hence, culpability into two main conditions—​normative competence and situational control. This conception of responsibility draws on and synthesizes two traditions of thinking about responsibility—​the reasons-​responsive wing of the compatibilist tradition of thinking about free will and responsibility (e.g., Wolf 1990; Wallace 1994; Fischer and Ravizza 1998; Nelkin 2011; Brink and Nelkin 2013; McKenna 2013; Vargas 2013; and Brink 2021) and the fair choice approach to criminal responsibility (e.g., Hart 1968; Moore 1997; and Morse 1994, 2002). If someone is to be culpable or responsible for her wrongdoing, then she must be a responsible agent. Our paradigms of responsible agents are normal mature adults who are normatively competent. They must be able to distinguish between the intensity and authority of their desires and impulses. They must not simply act on their strongest desires, but be capable of stepping back from their desires, evaluating them, and acting for good reasons. If so, normative competence involves reasons-​responsiveness, which itself involves both cognitive capacities to distinguish right from wrong and volitional capacities to conform one’s conduct to that normative knowledge. It is important to frame this approach to responsibility in terms of normative competence and the possession of these capacities for reasons-​responsiveness. In particular, responsibility must be predicated on the possession, rather than the use or exercise, of such capacities. We do excuse for lack of competence, but not for lack of performance. Provided the agent had the relevant cognitive and volitional capacities, we do not excuse the weak-​willed or the willful wrongdoer for failing to recognize or respond appropriately to reasons. If responsibility were predicated on the proper use of these capacities, we could not hold weak-​willed and willful wrongdoers responsible for their wrongdoing. Indeed, the fact of wrongdoing would itself be exculpatory, with the absurd result that we could never hold anyone responsible for wrongdoing. It is a condition

374   David O. Brink of our holding wrongdoers responsible that they possessed the relevant capacities or competence. Normative competence, on this conception, involves two forms of reasons-​ responsiveness: a cognitive ability to recognize reasons for or against conduct and a volitional ability to conform one’s will to this normative understanding. Both dimensions of normative competence involve norm-​responsiveness. As a first approximation, we can distinguish moral and criminal responsibility at least in part based on the kinds of norms to which agents must be responsive. Moral responsibility requires capacities to recognize and conform to moral norms, whereas criminal responsibility requires capacities to recognize and conform to norms of the criminal law. Though there will typically be considerable overlap between moral and criminal norms, especially in morally legitimate criminal systems, there is no reason to expect the coincidence to be perfect. Indeed, in most liberal regimes there will be many moral norms the violation of which the legal system will not criminalize. Normative competence requires the cognitive capacity to make suitable normative discriminations, in particular, to recognize wrongdoing. If so, then we can readily understand one aspect of the criminal law insanity defense. Most plausible versions of the insanity defense include a cognitive dimension, first articulated in the M’Naghten rule that excuses if the agent lacked the capacity to discriminate right from wrong at the time of action.4 However, normative competence requires more than cognitive competence. It also requires volitional capacities to form intentions based on one’s practical judgments about what one ought to do and to execute these intentions over time, despite distraction, temptation, and other forms of interference.5 Volitional impairment might take the form of (a) irresistible urges or paralyzing phobias that are neither conquerable nor circumventable,6 (b) severe depression or listlessness, or (c) damage to the prefrontal cortex of the brain, causing significant difficulty for agents in conforming to their own judgments about what they ought to do, as in the famous case of Phineas Gage.7

4 

M’Naghten’s Case, 10 Cl. & F. 200, 8 Eng. Rep. 718 (1843). For skepticism about the volitional dimension of normative competence, see Morse 2002. For a defense of the volitional dimension of normative competence, against volitional skepticism, see Brink and Nelkin 2013: 96–​302. 6  See Mele 1990. A desire is conquerable when one can resist it and circumventable when one can act so as to make acting on the desire impossible or at least more difficult. The alcoholic who simply resists cravings conquers his impulses, whereas the alcoholic who throws out his liquor and stops associating with former drinking partners or won’t meet them at places where alcohol is served circumvents his impulses. Conquerability is mostly a matter of will power, whereas circumventability is mostly a matter of foresight and strategy. Both are matters of degree. 7  Phineas Gage was a 19th-​century railway worker who was laying tracks in Vermont and accidentally used his tamping iron to tamp down a live explosive charge, which detonated and shot the iron bar up and through his skull, damaging his prefontal cortex. Though he did not lose consciousness, over time his character changed. Whereas he had been described as someone possessing an “iron will” before the accident, afterward he had considerable difficulty conforming his behavior to his own judgments about what he ought to do. See Damasio 1994. 5 

Responsibility, Punishment, and Predominant Retributivism    375 Recognition of a volitional dimension of normative competence argues against purely cognitive conceptions of insanity, such as the M’Naghten test, and in favor of a more inclusive conception, represented in the Model Penal Code, which conceives of insanity as involving significant impairment of either cognitive or volitional competence.8 A person is not responsible for criminal conduct if at the time of such conduct as a result of a mental disease or defect he lacks substantial capacity either to appreciate the criminality [wrongfulness] of his conduct or to conform his conduct to the requirements of law [§4.01].

If so, normative competence involves both cognitive and volitional competence. Both cognitive and volitional competence involve sensitivity to reasons. Sensitivity is a scalar notion. People can be more or less sensitive to reasons, and cognitive or volitional competence can be more or less impaired. If we analyze responsibility and excuse in terms of normative competence, this implies that both responsibility and excuse are, at least in principle, scalar concepts (e.g., Nelkin 2016, Brink 2021: ch.15). There may be important pragmatic reasons for making bright-​line discriminations in certain contexts, privileging certain degrees of normative competence, as necessary and sufficient for responsibility and excuse. These are interesting and complicated issues, beyond the scope of this chapter. But if we must recognize thresholds of normative competence, if only for pragmatic reasons, it seems fair to require, as the Model Penal Code does, substantial, rather than bare, normative capacity as a condition of responsibility and significant, rather than complete, impairment of capacity as grounds for an excuse. An important part of an agent’s being responsible for wrongdoing consists in her being a responsible agent. This we have conceptualized in terms of normative competence and analyzed into cognitive and volitional capacities. But there is more to an agent being culpable or responsible for her wrongdoing than her being responsible and having done wrong. Excuse is not exhausted by denials of normative competence. Among the factors that may interfere with our reactive attitudes, including blame and punishment, are external or situational factors. In particular, duress can induce an agent to engage in wrongdoing that she would not otherwise have performed. The paradigm situational excuse is coercion by another agent, as when one is threatened with physical harm to oneself or a loved one if one doesn’t participate in some kind of wrongdoing, for instance, driving the getaway car in a robbery. The Model Penal Code adopts a reasonable person version of the conditions under which a threat excuses, namely, when a person of reasonable firmness would have been unable to resist, provided the actor was not himself responsible for being subject to duress (§2.09).

8  The Model Penal Code (1985) is a model statutory text of fundamental provisions of the criminal law, first developed by the American Law Institute in 1962 and subsequently updated in 1981. It was intended to serve as a model for local jurisdictions drafting and revising their criminal codes. For further discussion of the Model Penal Code conception of insanity, see Brink 2013.

376   David O. Brink Duress is not an incompetence excuse. It does not compromise the wrongdoer’s normative competence or her status as a responsible agent. But it does challenge whether she is responsible for her wrongdoing. This shows that we need to distinguish between being a responsible agent and being responsible for one’s wrongdoing. Being normatively competent is sufficient for being a responsible agent, but it is necessary but not sufficient for being responsible for one’s wrongdoing. The duressed wrongdoer is a responsible agent, but nonetheless she is not responsible for her wrongdoing. Responsibility for wrongdoing requires both normative competence and situational control. Being responsible is a matter of having the right capacities, but being responsible for one’s conduct requires both capacities and appropriate opportunities. This conception of responsibility factors responsibility into normative competence and situational control and factors normative competence into cognitive and volitional capacities. This conception seems plausible, in significant part because it promises to accommodate our practices of excuse in both moral assessment and the criminal law pretty well. Incapacity excuses deny normative competence, whereas duress excuses deny the opportunity to exercise those capacities free from inappropriate interference by others. But it would be nice if there were some unifying element to its structure. One possible umbrella concept is control. Freedom from coercion and duress, cognitive competence, and volitional competence all seem to be aspects of an agent’s ability to control her actions. But control seems important, at least in part, because it seems unfair to blame agents for outcomes that are outside their control. This suggests that the umbrella concept should be fairness, in particular, the fair opportunity to avoid wrongdoing, because failure of either normative competence or situational control violates the norm that blame and punishment be reserved for those who had a fair opportunity to avoid wrongdoing. If we treat the fair opportunity to avoid wrongdoing as the key to responsibility, we get the following picture of the architecture of responsibility. Fair Opportunity to Avoid Wrongdoing

Situational Control

Normative Competence

Cognitive Competence

Volitional Competence

The fair opportunity conception of responsibility draws on resources familiar from reasons-​responsive conceptions of compatibilism and fair choice approaches to criminal responsibility and makes sense of the criminal law’s conceptualization of excuses

Responsibility, Punishment, and Predominant Retributivism    377 into impairments of capacities (incompetence) and impairment of opportunities (duress).9 The fair opportunity conception of responsibility is a response-​independent conception of responsibility, because it specifies the conditions of responsibility independently of the reactive attitudes. As such, the realist about Strawson’s thesis can draw on the fair opportunity conception to identify the conditions under which the reactive attitudes are fitting and justified.

5.  Retributive Essentials Realism about Strawson’s thesis says that our reactive attitudes are correct or justified insofar as they track the responsibility of their targets. Blame and punishment are reactive attitudes and practices. On the realist interpretation, Strawson’s thesis represents some negative attitudes and practices, including blame and punishment, as fitting responses to wrongdoing for which the agent is responsible or culpable. This is a retributivist thesis about blame and punishment being fitting or deserved insofar as the agent is responsible or culpable for her wrongdoing. More specifically, retributivism says that blame and punishment ought to be proportional to desert, which is itself the product of two independent variables—​wrongdoing and culpability or responsibility.10

P α D (= W x R )

On this view, desert is the product of the magnitude of wrongdoing and the degree of responsibility. In principle, different wrongs can be arranged on a scale of seriousness. Presumably, the magnitude of a wrong is often influenced by the amount of harm it produces or risks, though we probably shouldn’t reduce wrongness to harmfulness, because there are harms that are not wrong (e.g., cases of self-​defense and necessity) and because some wrongs may not be harmful (e.g., unfair free-​riding). Degree of responsibility should be thought of as a percentage of full responsibility, which we

9  Gary Watson distinguishes between two kinds of responsibility: attributability and accountability (Watson 1996). He associates attributive responsibility with action that expresses the agent’s real self, with an aretaic perspective on her action, and with her own evaluative commitments or orientation. By contrast, he thinks that accountability involves liability to sanctions, raises questions about the fairness of imposing sanctions, and presupposes that the agent had control and a fair opportunity to avoid sanctions. We can explain attributability in terms of quality of will and accountability in terms of fair opportunity. I am skeptical that we can explain moral responsibility and blame entirely in terms of attributability and quality of will without invoking accountability and fair opportunity (contrast Scanlon 2008: ch. 4). But, however that issue is resolved, criminal responsibility and punishment clearly require accountability and fair opportunity. 10  Here, I adapt some ideas in Nozick 1981: 363–​366. My formula resolves some ambiguities and inconsistencies in his discussion.

378   David O. Brink could represent on a 0–​1 scale. Thus, an agent would deserve punishment commensurate with the magnitude of the wrong she has committed just in case she is fully responsible for her wrongdoing and fractional punishments corresponding to reduced responsibility. There are at least three claims that are essential to any version of retributivism, whether it is a pure or mixed conception of punishment. 1. Desert in the form of culpable wrongdoing is a necessary condition of blame and punishment. 2. Proportionate justice sets an upper limit on permissible blame and punishment. 3. Blame and punishment are fitting responses to culpable wrongdoing in the sense that there is a strong pro tanto case for blame and punishment that is proportionate to desert. First, to avoid punishing the innocent, even when doing so would have consequentialist value, we must recognize retributivism as at least part of the truth about blame and punishment—​specifically, we must recognize desert as a necessary condition on blame and punishment. Second, proportionate just deserts track culpable wrongdoing and set an upper limit on the amount of punishment that is permissible in individual cases. Third, culpable wrongdoing is the desert basis of punishment in the sense that blame and punishment are pro tanto fitting responses to wrongdoing for which the agent is responsible. Whether proportionate blame and punishment are always all things considered appropriate responses will depend on the balance of reasons and whether our conception of retributivism is pure or mixed. Whether blame and punishment are fitting responses to culpable wrongdoing may depend on whether the wrongdoing in question is moral or legal. Moral retributivism says that blame and punishment are fitting responses to culpable moral wrongdoing. This is plausible insofar as persons are moral agents who are accountable for their conduct; respecting persons means that it is fitting that they be blamed and punished for wrongdoing for which they are responsible and that they merit or deserve blame and punishment. Legal retributivism says that blame and punishment are fitting responses to culpable legal wrongdoing. However, it is less clear that culpable legal wrongdoing always deserves blame and punishment if only because legal norms and moral norms are distinct and legal wrongdoing is not necessarily moral wrongdoing. But where legal wrongdoing is not morally wrong, it is less clear that blame and punishment are deserved, even pro tanto. One might conclude that there is a pro tanto case for punishing legal wrongdoing when and only when legal wrongdoing is also moral wrongdoing. But there can be a pro tanto case for punishing legal wrongdoing even in cases in which the legal norms being transgressed are morally problematic provided that the moral flaws in the law are fairly minor, do not occur too often, and do not have a systematically disparate impact on citizens. On this view, the duty to obey the law is part of a package deal, arising from a sufficiently just social compact. Herbert Morris’s justification of punishment by appeal

Responsibility, Punishment, and Predominant Retributivism    379 to fairness might be understood this way (1968). Morris argues that basic provisions of the criminal law against murder, rape, assault, theft, and trespass are norms the general observance of which is mutually beneficial. Culpable wrongdoers are those who free-​ ride, enjoying the benefits of others’ compliance with these norms without incurring the costs of doing their part to maintain the system from which they benefit. But free-​riding on mutually beneficial social practices is unfair. The culpable wrongdoer claims more than his fair share, and fairness demands that he be punished according to the terms announced in advance as the penalty for noncompliance. The mutually beneficial character of the laws of a reasonably just legal system extends to laws prohibiting crimes mala in se and mala prohibita. In the case of crimes mala in se, laws do not make the conduct wrong but rather declare it be wrong, legally reinforcing the wrongful status of wrongs against persons. In the case of crimes mala prohibita, the laws make the conduct wrong through enactment. In both cases, the prohibitions tend to be mutually beneficial, making noncompliance a public wrong, calling for state censure and sanction. In the case of crimes mala in se but not mala prohibita, the conduct is wrong independently of legal enactment. So in the case of crimes mala in se there are two kinds of reason for censure and sanction—​reasons to enforce natural duties and reasons of fairness—​whereas in the case of crimes mala prohibita there is just one kind of reason for censure and sanction—​reasons of fairness. This appeal to fairness to help anchor desert has an important scope limitation—​it only applies to social practices that are sufficiently just and mutually beneficial. If the social scheme does not meet this condition, then noncompliance may not be unfair. In particular, if one lives in a sufficiently unjust social scheme in which the benefits and burdens of social cooperation are not distributed fairly, then noncompliance with the scheme’s norms—​legal wrongdoing—​need not be morally wrong and, hence, blame and punishment may not be deserved. In such systems, blame and punishment of crimes mala in se will still be fitting, but blame and punishment of some crimes mala prohibita may not be fitting. When the scope condition is satisfied, legal, as well as moral wrongdoing, for which the agent is responsible merits punishment in the sense that there is a strong pro tanto reason to punish culpable wrongdoing. A suitably qualified retributivism not only has considerable intuitive appeal but also fits well with important parts of criminal jurisprudence. For instance, it explains well the two main kinds of affirmative defense a defendant can offer—​justifications and excuses.11 Justifications and excuses each deny one of the two elements in the retributivist desert basis for blame and punishment—​ wrongdoing and responsibility. Justifications, such as self-​ defense or necessity, deny wrongdoing, whereas excuses, such as insanity or duress, deny culpability or responsibility.

11  While justifications and excuses are the two main affirmative defenses available to defendants, there are also pragmatic or policy-​based exemptions, such as prosecutorial immunity for diplomats. See, Robinson 1997: 96–​124, 204–​207 and Berman 2003.

380   David O. Brink

6.  Misconceptions About Retributivism Though some form of retributivism is the dominant conception of punishment in the criminal law, it has a mixed reputation in other quarters. Some of these misgivings may reflect concerns about the pure retributivist thesis that desert is the only factor affecting blame and punishment and that desert is always sufficient to justify punishment whatever the costs or moral perplexities. Mixed conceptions of punishment that recognize desert as one element may avoid worries directed at pure retributivism. However, there are other concerns about retributivism that would seem to apply to all forms of retributivism, whether pure or mixed. 1. Retributivism is implausibly committed to lex talonis, which involves punishing wrongdoers in kind (e.g., an eye for an eye). 2. Retributivism is implausibly committed to harsh punishments, such as those found in mass incarceration—​trying juveniles as adults, three strikes laws, significant mandatory minima, long sentences for nonviolent crime, and post-​ incarceration penalties, such as disenfranchisement, loss of public assistance, and requirements to disclose one’s criminal record to potential employers. 3. Retributivism is implausibly committed to valuing the felt suffering of the wrongdoer. 4. Retributivism is implausibly committed to viewing retribution as an intrinsic good. Some, if not all, of these concerns reflect misunderstandings of retributivist essentials or accretions in the retributivist tradition. Retributivism requires proportionate justice, not punishment in kind. There are some ways that the state ought not to treat its citizens, even if they have committed heinous crimes. For instance, even in the Unites States, where we allow capital punishment and permit various severe punishments, we have a constitutional prohibition on cruel and unusual punishment, which can plausibly be interpreted as a prohibition on severely inhumane and disproportionate punishments. The state ought not to torture torturers or rape rapists; such punishments would arguably constitute cruel and unusual punishment. We can still punish proportionately without punishing in kind. Lex talonis might allow for swift but inhumane punishments in the case of rape or torture; proportional punishments that are humane might take longer. Some writers lay the problems associated with mass incarceration at the door of the retributivist (Kelly 2002, 2009, 2019). Retributivism requires accountability and proportionate punishment, but it’s not obvious why that supports trying juveniles as adults, steep mandatory minima, long prison sentences, brutal prison conditions, or

Responsibility, Punishment, and Predominant Retributivism    381 post-​incarceration penalties—​such as disenfranchisement, the inability to serve on juries, the loss of public assistance, and the duty to report felony convictions to potential employers. These are some of the aspects of the war on drugs and mass incarceration that have led to calls for criminal justice reforms (e.g., Alexander 2010). But there’s no reason to assume that retributivism supports such draconian penalties. Indeed, there is every reason to believe that mass incarceration imprisons people out of proportion to their deserts. What proportionality clearly requires is comparative just deserts. This gives us only ordinal information about the magnitude of just deserts until we have anchors for the ordinal scale. A natural way to provide anchors is to determine the most severe punishment permissible for the most heinous crime (e.g., serial murder and torture) and the mildest form of sanction for the most minor infraction (e.g., jaywalking). There is room for disagreement about what quantum of punishment is appropriate for the anchors, especially at the maximum. Norway sets the maximum at 21 years, even for Anders Breivik, who killed 77 people in terrorist attacks, including 69 participants at the Social Democratic Youth League summer camp. The United States currently sets the maximum at capital punishment. We might wonder whether capital punishment is too severe, especially because it forecloses the possibility of discovering wrongful convictions and because it has proven difficult to administer capital punishment in a way that is not racially unjust. Perhaps the top end of the scale should be life in prison without the possibility of parole (LPWP). But even if LPWP is the top end of the scale, most other crime, including violent crime, deserves considerably less punishment. We also need to distinguish the length and manner of punishment. We tend to think of punishment in terms of incarceration. But incarceration is only one form punishment can take. It can also take the form of diversionary treatment (such as house arrest, counseling, and/​or community service), fines, and restitution. When punishment does involve incarceration, it need not be in overcrowded and inhumane conditions that endanger the health and safety of convicts and that provide little educational or vocational training to equip prisoners to succeed in life outside of prison. It is open to retributivists to think that proportionality requires shorter periods of incarceration and more humane forms of incarceration for many crimes. Indeed, one might think that incarceration should be limited primarily to violent crime and serial offenders and that various kinds of diversionary treatment, such as community service and house arrest, would be appropriate for most forms of nonviolent crime and first offenders. In these ways, one can be a retributivist without endorsing the degree of punitiveness in the current American criminal justice system. Retributivists can and do think that prison conditions, criminalization for nonviolent offenses, sentence length, and post-​ incarceration penalties are disproportionately punitive and violate just deserts.12

12 

See, e.g., the retributivist critique of the juvenile transfer trend in Brink 2004.

382   David O. Brink What is the currency of desert? Both retributivists and their critics often assume that it is felt suffering and claim that retributivism is committed to the proposition that it is intrinsically good for culpable wrongdoers to suffer (Kant 1797–​98: 331; Ross 1930: 135–​138; Hart 1968: 234–​235; Moore 1997: 163; Kolber 2009: 182; Tadros 2011: 63; and Berman 2013: 87). That’s one form that retributivism might take. But it is not essential to retributivism. Retributivism can instead be understood as the idea that desert demands accountability, where it is fitting to hold wrongdoers accountable in ways that reflect the nature and gravity of their wrongs and their degrees of culpability. But the currency of accountability need not be suffering. Instead, it might consist in the deprivation of certain rights and privileges of citizenship, which is temporary for all but the most serious forms of wrongdoing. As we have seen, retributivists endorse the idea that punishment is a fitting response to the culpable wrongdoer free-​riding on the mutually beneficial social compact among citizens involved in the rule of law. The free-​rider’s noncompliance with the rule of law unfairly exploits the compliance of others. So, fairness requires punishing culpable wrongdoers who free-​ride. But if this is part of the rationale for retributive deserts, then there’s a strong case to be made that the appropriate currency of punishment should be some modification in the rights and privileges of citizenship, because it is the social compact among citizens on which the wrongdoer free-​rides. Offenders have reason to regard the loss of rights and privileges of citizenship, even if temporary, as unwelcome and may well experience psychic costs and suffering as a result. If so, suffering may be a common by-​product of punishment, but punishment aims at accountability, measured in terms of the rights and privileges of citizenship, not suffering. Many of the same writers who link retributivism with the value of suffering want to claim that imposing suffering on culpable wrongdoers is an intrinsic or non-​ instrumental good. Accountability does not require suffering, and the currency of retributivism should instead be loss of rights and privileges. If we avoid the mistake about suffering, should we then embrace the idea that retributivism treats accountability in the form of deprivation of rights and privileges for culpable wrongdoers as an intrinsic good? As long as we remember that the state of affairs is the complex one of being a deprivation of rights and privileges that is consequential on and proportionate to culpable wrongdoing, we might claim that retributivism is committed to viewing that as intrinsically or non-​ instrumentally good. But this should be understood as a moral good. For we could equally well say that what makes blame and punishment fitting responses to culpable wrongdoing is that they are demands of justice. This, after all, is why retributive essentials are often described in terms of just deserts or proportionate justice. It seems we can describe retributive outcomes either as a moral good or as a demand of justice. If the right is prior to the good, then the justice claim is prior in order of explanation to the moral goodness claim.

Responsibility, Punishment, and Predominant Retributivism    383

7.  Predominant Retributivism So far, we have defended a conception of punishment with retributive elements, but we have been agnostic between pure retributivism and a mixed conception with retributive elements. Pure retributivism claims that desert, understood in terms of culpable wrongdoing, is the only factor relevant to justifying punishment and that desert is both necessary and sufficient for punishment (e.g., Moore 1997). By contrast, a mixed conception of punishment combines retributive and forward-​looking elements. There are different kinds of mixed conceptions, but most insist that desert is a necessary condition of punishment, because only that will avoid the most serious worries about over-​inclusiveness afflicting pure forward-​looking conceptions. One mixed theory is a disjunction of desert and consequentialist values, claiming that either is sufficient to justify punishment. But disjunctivism inherits the problems of over-​inclusiveness and under-​inclusiveness of pure forward-​looking conceptions, because it allows consequential values alone to justify punishment. A conjunction of desert and consequentialist values that treats desert and good consequences as individually necessary and jointly sufficient for punishment is more plausible, because it avoids the more serious over-​inclusiveness objection to pure forward-​looking conceptions. Nonetheless, conjunctivism is still under-​ inclusive insofar as it doesn’t recognize any reason to punish culpable wrongdoing when doing so does not have good consequences. If these were our only options, then pure retributivism might be the most plausible conception. But these are not the only kinds of mixed conceptions, and pure retributivism doesn’t explain the persisting attractions of forward-​looking considerations, such as deterrence, rehabilitation, norm enforcement and communication, and reconciliation. The most plausible formulation of retributive ideas is a mixed conception of punishment in which retributive ideas dominate, which we might call predominant retributivism (Brink 2012).13 Any form of retributivism insists that culpable wrongdoing serves as a constraint on blame and punishment. This is sometimes called the negative retributivist thesis. Desert is a necessary condition for blame and punishment, because both are impermissible in the absence of culpable wrongdoing, and desert places an 13  Another mixed conception is the sort of rule consequentialist conception of punishment defended by Rawls (Rawls 1955). He distinguishes between the prospective (legislative) perspective on a practice, specifying the reasons for having a practice, and the retrospective (judicial) perspective on that practice, applying the reasons within a practice. This allows him to claim that the reasons for having a practice of punishment are consequentialist but the reasons for deciding how to handle cases arising under the practice are retributive. It’s not clear why Rawls thinks that prospective issues can be neatly sequestered from retrospective ones. Presumably, the prospective perspective asks what’s the optimal set of laws and judicial institutions and practices for enforcing them, and it’s not clear why optimal institutional design wouldn’t allow the judiciary to depart from retributive constraints when it would be clearly optimal to do so. It’s doubtful that consequentialist reasons for having a practice will justify a practice that operates in a strictly retributive fashion.

384   David O. Brink upper limit on permissible punishment, which must not exceed what would be proportional to desert. Both pure and predominant retributivism insist that desert is not just a constraint on blame and punishment but also provides an important rationale for these attitudes and practices. This is sometimes called the positive retributivist thesis.14 Both claim that blame and punishment are fitting responses to culpable wrongdoing in the sense that culpable wrongdoing is sufficient for a strong pro tanto case for blame and punishment. Other things being equal, we should blame and punish culpable wrongdoers proportionate to their desert. However, other things are not always equal. This is where predominant retributivism diverges from pure retributivism. Whereas the pure retributivist thinks that desert is the only factor affecting the decision to punish, predominant retributivism recognizes that there can be non-​desert-​based reasons not to punish proportionately, in particular, to punish less than required by desert. Such non-​desert factors can include the costs of punishment; the prospects for rehabilitation; evidence of remorse and restitution; and the value of mercy, forgiveness, and reconciliation. The strong pro tanto case for blame and punishment based on desert must sometimes compete with these countervailing considerations and may not always carry the day. Because predominant retributivism allows that desert can compete with non-​desert reasons against proportionate punishment, it is a mixed theory of punishment. Predominant retributivism is also a mixed theory of punishment insofar as it allows desert and non-​desert factors to address different questions about punishment. A complete theory of punishment should address whom to punish, why we should punish, how much we should punish, and how we should punish. Retributivism offers a unified answer to the first three questions: We should punish culpable wrongdoers; we should do so, because they deserve punishment for culpable wrongdoing; and we should punish them proportionally to the degree of their culpable wrongdoing. Though desert may constrain how we punish, because it constrains how much we should punish, it leaves the manner of punishment largely underdetermined (e.g., Husak 2018). Provided that we punish all and only the guilty, punish them because they deserve punishment, and punish them in proportion to, or at least not in excess of, their just deserts, we are free to and should punish them in ways likely to promote various forward-​looking values, such as deterrence, rehabilitation, norm enforcement and communication, and reconciliation. Another way to reconcile forward-​ looking rationales for punishment and the retributivist backward-​looking focus on desert is to allow consequentialist considerations to supplement considerations of desert in the determination of precisely how much to punish. Despite the apparent precision of the retributivist formula for punishment, desert may determine only an interval, rather than a precise quantum, of punishment. Indeed, interval sentencing is reflected in various sentencing guidelines, including the

14 

Duff has a useful discussion of the negative and positive retributive desert theses (Duff 2001: 12, 19).

Responsibility, Punishment, and Predominant Retributivism    385 United States Federal Sentencing Guidelines.15 Whether punishment intervals reflect genuine metaphysical indeterminacy in the desert basis of particular crimes or epistemological indeterminacy in our ability to track small but genuine differences in culpable wrongdoing is an interesting question, which we need not settle here. As long as there is limited indeterminacy in the desert basis for punishment, whether metaphysical or epistemological, there is a need to specify further a precise punishment value within the deserved interval. So, it is open to us to appeal to various consequentialist rationales to fine-​tune the quantum of punishment within the space of punishments set by desert. Predominant retributivism is a conception of punishment that mixes a backward-​ looking focus on desert with various forward-​looking rationales for punishment, but in which desert conditions forward-​looking considerations in three ways. First, desert in the form of culpable wrongdoing is a necessary condition of blame and punishment. Second, retributive justice that is proportionate sets an upper limit on permissible blame and punishment. Third, blame and punishment are fitting responses to culpable wrongdoing in the sense that there is a strong pro tanto case for blame and punishment that is proportionate to desert. In some cases, there may be sufficient reason to punish less than just deserts, but this will be in spite of the demands of just deserts.

8.  Concluding Remarks Punishment involves both censure and sanction. As such, it seems to presuppose that the targets of punishment are blameworthy—​that is, that they are culpable or responsible for wrongdoing. This makes purely forward-​looking conceptions of punishment that appeal to deterrence, rehabilitation, norm communication and reinforcement, or reconciliation and permit punishment in the absence of culpable wrongdoing implausible, and it favors conceptions that are retributive in holding that punishment should be proportional to desert, understood as the product of wrongdoing and responsibility. Realism about Strawson’s thesis linking responsibility and the reactive attitudes encourages us to look to our practices of excuse to identify the nature of responsibility and culpability. When we follow this methodological advice, we can avoid skepticism about responsibility and are led to the fair opportunity conception of responsibility, which represents responsibility as consisting in suitable capacities—​normative competence—​and opportunities—​situational control. Fair opportunity is a response-​ independent conception of responsibility of the sort required by realism about Strawson’s thesis. This realist interpretation of Strawson’s thesis supports some form of retributivism insofar as it implies that it is fitting to blame and punish culpable wrongdoing. Retributive justice seeks accountability, not suffering, and it condemns various aspects of mass incarceration as inconsistent with just deserts. However, this case for

15 

United States Sentencing Commission, Guidelines Manual (Nov. 2016).

386   David O. Brink retributivism is agnostic between pure retributivism and mixed conceptions that include important retributive elements. Predominant retributivism is a mixed conception of punishment that can explain the appeal of forward-​looking rationales for punishment while maintaining that they are constrained by desert. On this view, desert is a necessary condition on permissible punishment, sets an upper bound on permissible punishment in the form of proportionate justice, and provides a strong pro tanto case for blame and punishment that is proportionate to desert.

References Alexander, Michelle. 2010. The New Jim Crow (New York: The New Press). Allais, Lucy. 2012. “Restorative Justice, Retributive Justice, and the South African Truth and Reconciliation Commission,” Philosophy & Public Affairs 39: 331–​363. American Law Institute. 1985. Model Penal Code and Commentaries (Philadelphia). Berman, Mitchell. 2003. “Justification and Excuse, Legality and Morality,” Duke Law Journal 53: 1–​26. Berman, Mitchell. 2013. “Rehabilitating Retributivism,” Law & Philosophy 32: 83–​108. Blackstone, William. 1791. Commentaries on the Laws of England, 4 vols. (London: Strahan & Woodfall). Brink, David. 2004. “Immaturity, Normative Competence, and Juvenile Transfer,” Texas Law Review 82: 1555–​1585. Brink, David. 2012. “Retributivism and Legal Moralism,” Ratio Juris 25: 496–​512. Brink, David. 2013. “Responsibility, Incompetence, and Psychopathy,” Lindley Lecture 53: 1–​41. Brink, David. 2021. Fair Opportunity and Responsibility (Oxford: Clarendon Press). Brink, David and Nelkin, Dana. 2013. “Fairness and the Architecture of Responsibility,” Oxford Studies in Agency and Responsibility 1: 284–​313. Brink, David and Nelkin, Dana. 2022. “The Nature and Significance of Blame,” in Oxford Handbook of Moral Psychology, eds. J. Doris and M. Vargas (New York: Oxford University Press). Carusso, Greg. 2018. “Skepticism About Moral Responsibility,” Stanford Encyclopedia of Philosophy . Carusso, Greg. 2019. “The Public Health-​Quarantine Model,” in The Oxford Handbook of Moral Responsibility, eds. D. Nelkin and D. Pereboom (New York: Oxford University Press). Damasio, Antonio. 1994. Descartes’ Error: Emotion, Reason, and the Human Brain (New York: Putnam). Duff, R.A. 2001. Punishment, Communication, and Community (Oxford: Clarendon Press). Duff, R.A. 2009. “Restoration and Retribution,” in Principled Sentencing, 3d ed., eds. A. von Hirsch, A. Ashworth, and J. Roberts (Oxford: Hart Publishing). Fischer, John and Ravizza, Mark. 1998. Responsibility and Control (New York: Cambridge University Press). Franklin, Benjamin. 1785. “Letter to Benjamin Vaughn, March 14, 1785,” in The Writings of Benjamin Franklin, ed. A. Smyth (New York: Macmillan, 1906), vol. 9. Hart, H.L.A. 1968. Punishment and Responsibility (Oxford: Clarendon Press). Husak, Douglas. 2018. “Kinds of Punishment” in Moral Puzzles and Legal Perplexities, ed. H. Hurd (New York: Cambridge University Press).

Responsibility, Punishment, and Predominant Retributivism    387 Kant, Immanuel. 1797–​98. Metaphysics of Morals in Immanuel Kant, Practical Philosophy, trans. M. Gregor (New York: Cambridge University Press, 1996) (Prussian Academy pagination). Kelly, Erin. 2002. “Doing Without Desert,” Pacific Philosophical Quarterly 83: 180–​2015. Kelly, Erin. 2009. “Criminal Justice Without Retribution,” Journal of Philosophy 106: 419–​439. Kelly, Erin. 2019. The Limits of Blame: Rethinking Punishment and Responsibility (Cambridge, MA: Harvard University Press). Kolber, Adam. 2009. “The Subjective Experience of Punishment,” Columbia Law Review 109: 182–​236. McKenna, Michael. 2013. “Reasons-​Responsiveness, Agents, and Mechanisms,” Oxford Studies in Agency and Responsibility 1: 151–​183. Mele, Al. 1990. “Irresistible Desires,” Noûs 24: 455–​472. Moore, Michael. 1997. Placing Blame (Oxford: Clarendon Press). Morris, Herbert. 1968. “Persons and Punishment,” The Monist 52: 475–​501. Morse, Stephen. 1994. “Culpability and Control,” University of Pennsylvania Law Review 142: 1587–​1660. Morse, Stephen. 2002. “Uncontrollable Urges and Irrational People,” Virginia Law Review 88: 1025–​1078. Nelkin, Dana. 2011. Making Sense of Freedom and Responsibility (New York: Oxford University Press). Nelkin, Dana. 2016. “Difficulty and Degrees of Moral Praiseworthiness and Blameworthiness,” Noûs 50: 356–​378. Nozick, Robert. 1981. Philosophical Explanations (Cambridge, MA: Harvard University Press). Pereboom, Derk. 2014. Free Will, Agency, and Meaning in Life (Oxford: Clarendon Press). Rawls, John. 1955. “Two Concepts of Rules,” Philosophical Review 64: 3–​32. Robinson, Paul. 1997. Structure and Function in the Criminal Law (Oxford: Clarendon Press). Ross, W.D. 1930. The Right and the Good (Oxford: Clarendon Press). Scanlon, T.M. 2008. Moral Dimensions: Permissibility, Meaning, and Blame (Cambridge, MA: Harvard University Press). Shoemaker, David. 2017. “Response-​Dependent Responsibility,” Philosophical Review 126: 481–​527. Strawson, Peter. 1962. “Freedom and Resentment,” Proceedings of the British Academy 48: 1–​25. Tadros, Victor. 2011. The Ends of Harm: Moral Foundations of Criminal Law (Oxford: Clarendon Press). Vargas, Manuel. 2013. Building Better Beings (Oxford: Clarendon Press). Wallace, R. Jay. 1994. Responsibility and the Moral Sentiments (Cambridge, MA: Harvard University Press). Watson, Gary. 1987. “Responsibility and the Limits of Evil: Variations on a Strawsonian Theme,” reprinted in Watson 2004. Watson, Gary. 1996. “Two Faces of Responsibility,” reprinted in Watson 2004. Watson, Gary. 2004. Agency and Answerability (Oxford: Clarendon Press, 2004). Wolf, Susan. 1990. Freedom Within Reason (New York: Oxford University Press).

CHAPTER 19

Legal Resp onsi bi l i t y Psychopathy, a Case Study Elizabeth Shaw

1.  Introduction Sentimentalists argue that the capacity to experience certain emotions is necessary for moral understanding.1 This view, if sound, has potentially important practical implications. Firstly, if people who lack the relevant emotional capacities exist, then these people lack moral understanding. Secondly, if (as many moral philosophers and legal theorists believe) moral understanding is a necessary condition for moral and criminal responsibility, then people who lack the relevant emotional capacities are not morally or criminally responsible. Thirdly, to the extent that society subjects emotionally incapacitated individuals, who are not morally or criminally responsible to harm, in the form of blame and punishment, then society is currently inflicting unjustified harm on these individuals, given that responsibility is widely considered a prerequisite for justifiable blame or punishment. Arguably, these considerations give rise to a morally pressing need to investigate whether there are people with the relevant emotional incapacities and whether reliable techniques can be developed for identifying them, to enable them to be exempted from blame and punishment. (Although, even if blaming and punishing emotionally incapacitated individuals were unjustified, it could still be justified to detain/​treat them to prevent them from harming others.) Philosophers and cognitive scientists have increasingly engaged in interdisciplinary research into these questions. A consensus seemed to be emerging (for an overview, see Litton 2010) that psychopaths

1  For discussions of varieties of sentimentalism and rival theories, see Kauppinen (2014) and Gill (2007).

Legal Responsibility   389 were plausible candidates for individuals lacking in the capacities that sentimentalists consider necessary for moral understanding and that the most promising technique for identifying psychopathy was the Psychopathy Checklist-​Revised (PCL-​R) (Hare 2003). Many philosophers who have written about the topic have endorsed the conclusion of Malatesti and McMillan’s (2010) edited collection on psychopathy that “. . . based on a consideration of the empirical literature there are good reasons for not considering psychopaths morally responsible” (p. 319).2 However, increasingly, the empirical evidence regarding psychopaths’ supposed incapacities has yielded seemingly conflicting results that have been very hard to interpret. This empirical uncertainty raises the question of how society should treat individuals who may have responsibility-​undermining emotional incapacities. Some findings seem to indicate significant neurological and behavioral differences between psychopaths and non-​psychopaths. However, other studies have failed to show significant abnormalities among the sample of psychopaths that were tested; most research indicates that psychopaths’ deficits seem to be a matter of degree; and there is some evidence that (at least some of) psychopaths’ abnormalities emerge only under certain, specific conditions (for overviews, see, e.g., Blair 2017; Maibom 2017; Tillem et al. 2019). Based on this equivocal evidence, a growing number of theorists have expressed varying amounts of skepticism about the claim that psychopaths are not morally responsible. For example, Jurjako and Malatesti (2018) have argued that psychopaths’ deficits seem context-​ dependent and that psychopaths should only be relieved from responsibility, if at all, in “unusual” (p. 1020) cases where the situation in which the crime was committed resembled the conditions in those experiments where psychopaths’ deficits were apparent. Maibom (2008, 2018) has argued that the evidence suggests that, at most, psychopaths’ emotional capacities are merely impaired and so they cannot be excused based on a lack of these capacities. Fox et al. (2013) claim their “analysis of the empirical evidence” suggests that psychopaths should be considered partially responsible (p. 1). Jalava and Griffiths (2017) conclude that “in the absence of consistent data and a coherent way of determining their meaning, it is premature to use the data to pronounce on psychopaths’ responsibility . . . Inconsistent but suggestive data do not imply merely reduced responsibility, nor do they allow for tentative conclusions about responsibility” (p. 9). This chapter will argue that moral and legal philosophers should respond to this empirical uncertainty by paying more attention to the issue of who should bear the burden of proof, and to what standard, before an individual can be held responsible. In the context of criminal responsibility, in common law criminal justice systems, the general rule, based on the presumption of innocence, is that the state bears the burden to prove the accused’s guilt beyond reasonable doubt.3 However, there is an

2 

Although some theorists who endorse this conclusion base their arguments on psychopaths’ (supposed) rational rather than emotional incapacities for discussion see Malatesti & McMillan 2010, ch13. 3  Woolmington v. DPP [1935] AC 462, 481.

390   Elizabeth Shaw exception to this rule, known as “the presumption of sanity,” which, in some jurisdictions, places the burden of proof on the accused to prove certain mental incapacity defenses (such as the “insanity” defense) on the balance of probabilities.4 In England and Wales, the Law Commission (2013) has proposed that the accused should only have to bring some evidence to raise the accused’s mental incapacity as a live issue, and then, after presenting this evidence, it would be for the prosecution to prove beyond reasonable doubt that the accused, in fact, possessed the mental capacities required for criminal responsibility.5 This chapter will provide some reasons for thinking that the Law Commission’s proposal is in accordance with the moral rationale underlying the beyond reasonable doubt principle. Applying these conclusions about the burden and standard of proof to the problem of individuals with emotional incapacities, it will be argued that (a) the current evidence is not strong enough to warrant excusing someone on the basis of a psychopathy diagnosis alone, but (b) if there is a plausible interpretation of the evidence about the nature of psychopathy in general that is consistent with the sentimentalist’s basis for denying that such individuals have the capacity for moral understanding, then this evidence together with an assessment of the particular individual’s capacities (which must be determined on a case-​by-​case basis) could provide grounds for excusing particular individuals. Of course, sentimentalism is a contested theory and so there is theoretical uncertainty about this aspect of the case for excusing an individual with psychopathy (not just empirical uncertainty about the scientific evidence concerning the nature of psychopathy). However, I will suggest that the considerations relating to the burden and standard of proof should be applied to this theoretical uncertainty (as well as to empirical uncertainty) and that any doubt about sentimentalism should be dealt with in favor of the accused. The second section of what follows will outline the argument, based on sentimentalism, for excusing individuals with certain emotional incapacities from moral and criminal responsibility. The third section will briefly describe key features of psychopathy as diagnosed by the PCL-​R. The fourth section will discuss the appropriate burden and standard of proof when determining the criminal responsibility of psychopaths. The fifth section will summarize relevant research on psychopathy and will discuss an interpretation of this research that is consistent with psychopaths being entitled to an excuse on sentimentalist grounds. Finally, the sixth section will discuss the need for individualized evidence and will argue that it should be decided on a case-​by-​case basis whether someone diagnosed with psychopathy should be excused from criminal responsibility.

4  M’Naghten’s Case (1843) 10 Cl & F 210. Ramadan (2002, p. 247) provides a useful summary of variations between jurisdictions in terms of the nature of the burden placed on the accused. 5  The Law Commission made this proposal in the context of arguing that the current insanity defense in England and Wales should be abolished and replaced by a “recognised medical condition” defense.

Legal Responsibility   391

2.  Emotional Capacities, Moral Understanding, and Responsibility Moral sentimentalists take different views about the precise role of emotions in moral understanding and about which emotions might be required, and about the nature of emotion (for overviews of these issues, see Kauppinen 2014). One of the earliest proponents of sentimentalism was the 18th-​century philosopher David Hume, who argued that feelings of approval or disapproval were a necessary part of making moral judgements (Hume 1975). Hume thought that these feelings stemmed from sympathy, which involved sharing another person’s feelings. Many contemporary theorists who subscribe to the idea that psychopaths lack moral understanding have based their arguments on the idea that psychopaths lack the capacity for “empathy,” which resembles Hume’s conception of sympathy (Deigh 1995; Fine and Kennett 2004; Levy 2007; Morse 2008; Haji 2009; Glenn et al 2011; Focquaert et al. 2014). There are different conceptions of empathy (for a taxonomy see Maibom 2017b). For the purposes of this chapter, the following account of empathy will be outlined, which seems broadly to capture what many theorists seem to have in mind. Empathy involves the ability to feel an appropriate emotion in response to someone else’s emotion or predicament (e.g., compassion at someone’s distress).6 It implies that one can experience a feeling of caring7 to some extent about the other person and that one can feel some aversion to the idea of other people being harmed. The connection between empathy and moral understanding, on this view, is that one cannot “get the point” of moral norms prohibiting harm to

6 

This capacity to respond to another’s emotion with an appropriate emotion is sometimes called “affective empathy.” In contrast, “cognitive empathy” (or “theory of mind”) is the ability to work out what someone is feeling (“affective” theory of mind) or knows/​believes (“cognitive” theory of mind). (For a useful diagram outlining the connections between these concepts, see Sebastian 2011, p. 814). These deficits may interact. Difficulties working out what someone is feeling could interfere with one’s ability to respond with appropriate emotions. Similarly, if one feels nothing when witnessing another’s emotion, one may attend less to others’ emotions, interfering with the ability to recognize others’ emotions. It is often claimed that psychopathy mainly impairs affective empathy (e.g. Blair 2008; Bollard 2013; Nelkin 2016), whereas certain other conditions, such as mild autism spectrum disorders, mainly affect cognitive empathy (Montgomery et al. 2016 and see references listed in Baron-​Cohen et al. 2019). However, there is some evidence, discussed in section 5.3 of this chapter, suggesting that some psychopaths may suffer both from reduced affective empathy and (to some extent) reduced cognitive empathy (specifically reduced automatic affective theory of mind). It seems that individuals with mild autism spectrum disorder can have intact moral understanding and a strong moral code, despite problems with cognitive empathy, suggesting that reduced cognitive empathy per se is not an insuperable barrier to moral understanding, if affective empathy is intact. For further discussion of the relevance of different types of empathy impairment (seen in autism versus psychopathy) to the capacities required for responsibility, see, e.g., Kennett 2002; Bollard 2013; Nelkin 2016; Dineen 2019. 7  This capacity to feel concern for another is sometimes referred to as “sympathy” (Sebastian 2011). For the purposes of this chapter, unless otherwise indicated, “empathy” will refer to “affective empathy” (as defined in the previous footnote) together with “sympathy.”

392   Elizabeth Shaw others, if one is incapable of experiencing a feeling of caring about other people. On this view, merely understanding, purely intellectually, that there is a rule against seriously harming people, or that others label such actions “wrong,” is not enough for genuine moral understanding (or at least it is not enough for genuine understanding of harm-​ based moral norms). The ability to list such rules while being cut off from their emotional content might be compared to a parrot’s ability to repeat words without grasping their meaning. Theorists who take this view often argue that understanding the moral wrongfulness of harming others requires emotional capacities that are not required for understanding nonmoral norms, for example, social conventions like dress codes (see section 5.1). They argue that one does not really understand what is wrong with, for example, a person stepping on someone’s face for sadistic pleasure, if one has never been capable of feeling any more aversion to the thought of harming others than one would feel about breaching a rule not to step on the grass. It might be objected that people often judge actions that harm a person to be morally wrong without feeling empathy at that time or ever feeling empathy for that person. For example, one might be able to understand the wrongfulness of harms done centuries ago, or the wrongfulness of harming an adversary without feeling empathy for those victims.8 In reply, it might be suggested that moral understanding requires the capacity to empathize with someone, at some point, in order to learn the meaning of moral concepts, even if one does not need to empathize every time one later encounters or applies those moral concepts. By analogy, arguably in order to fully understand what is meant by the word “blue,” a person would have had to see the color blue at some point, but after learning this concept, she can then understand what her friend is talking about when he phones her up and says he has painted his room blue, even though she has not seen her friend’s room. Another objection might be based on an alternative conception of moral understanding based on rationalism, according to which, unlike sentimentalism, moral understanding does not require empathy, but is an intellectual matter analogous to mathematical reasoning (Gill 2007). The rationalist might argue, for example, that if one thinks one’s own pain is bad, one is rationally committed to believing, to be consistent, that it is wrong to inflict pain on others. (A similar objection is discussed in Deigh 1995.) In reply, firstly, it is not clear how the rationalist can, without begging the question, explain the leap from thinking one’s own pain is “bad,” in the sense of disliking it, to thinking that causing others’ pain is “wrong,” in the sense of morally wrong. Secondly, inconsistency is not necessarily always irrational. Putting milk in one’s tea on Monday 8 

The point that emotional response to wrongdoing tends to diminish with spatial and temporal distance was considered by Hume, who maintained that sympathy was the source of moral judgments. He argued that reason tells us that that human suffering which occurs in a distant land is no less bad than suffering which is more immediate. By analogy, he argued, reason tells us that objects do not really diminish in size the further we move away from them. However, this does not show that reason alone could enable us to make moral judgments any more than reason alone without visual sensations could allow us to make judgments about the real and apparent sizes of the objects which we see (Hume 1975). For another broadly Humean response to this kind of problem, see Radcliff (1994).

Legal Responsibility   393 and then, for no particular reason, drinking one’s tea without milk on Tuesday is, in a sense, inconsistent without being irrational (Deigh 1995). Thirdly, even if it were demonstrated that immorality involved an irrational kind of inconsistency, it is not clear that understanding that hurting others is “irrational because inconsistent” is the same as understanding that it is “morally wrong.” Imagine that someone said, “I firmly believe that harming others is morally wrong, and so I never harm others, although I have always felt neutral/​pleased/​amused at other people’s distress in itself. It’s the inconsistency that bothers me.” The sentimentalist would doubt the speaker had genuine moral understanding of what was wrong about harming others. There is not scope in this chapter to engage further in the debate between sentimentalists and rationalists about the nature of moral understanding. I will suggest in the next section that the arguments in favor of sentimentalism may be enough to raise a reasonable doubt about whether individuals who are incapable of empathy can understand moral norms and that this might be enough to justify excusing them. At this point, it is necessary to briefly spell out the connection between moral understanding, moral responsibility, and criminal responsibility. Most philosophers accept that in order to be held morally responsible for failing to act in accordance with moral reasons, one must be capable of guiding one’s actions in accordance with moral reasons, which requires (among other conditions) that one understands such reasons (for an influential account of moral responsibility, see Fischer and Ravizza 1998). Many theorists of criminal law and punishment, in turn, accept that moral responsibility should be a precondition for criminal responsibility. For example, according to retributivism (which may be the dominant view among penal theorists—​Matravers 2016) punishment is only justified if the offender deserves it, in virtue of being morally responsible for a wrongful act. Communication theorists (e.g., Duff 2001) explain the connection between moral understanding and criminal responsibility partly by reference to punishment’s aims of communicating a message of moral disapproval of criminal wrongdoing and attempting to persuade offenders to reform. These aims depend on the offender being able to understand this moral message and to understand the moral reasons for reforming. The literature on the criminal responsibility of psychopaths has been particularly influenced by the communication theory of punishment (see, e.g., Duff 1977; Fine and Kennett 2004; Morse 2008; Shaw 2009). Theorists who maintain that psychopaths should be excused from criminal responsibility typically focus on psychopaths’ eligibility for the insanity defense, which in many common law criminal justice systems is based on the M’Naghten Rules, which state that to qualify for the defense, the accused must have, at the time of the crime, suffered from “a disease of the mind, as not to know the nature and quality of the act he was doing; or if he did know it, that he did not know he was doing what was wrong.”9 However, knowledge of wrongfulness has been interpreted by the courts in some jurisdictions, such as England and Wales, as knowledge of legal rather than moral 9  M’Naghten’s Case (1843) 10 Cl & F 200, per Lord Chief Justice Tindal at 210. On ‘wrongfulness’ see Windle (1952) 2 QB 826.

394   Elizabeth Shaw wrongfulness. However, the Law Commission (2013) has proposed to replace the insanity defense with a broader defense that would allow the jury to consider whether the accused had the capacity to understand the moral wrongfulness of his conduct. This chapter will be based on the assumption, shared by many legal theorists, that the insanity defense (or an equivalent defense) ought to include an incapacity for moral understanding. It should also be noted that arguments for excusing psychopaths from responsibility have been developed that are more consistent with other metaethical theories, such as rationalism (e.g., Nelkin 2016). There is not scope within this chapter to provide an account of these other arguments. However, this chapter exemplifies a methodology that could be adopted by proponents of these other arguments: (1) present the metaethical and normative arguments for a particular set of criteria for responsibility, (2) present the empirical evidence that psychopaths do not meet these criteria, and (3) consider whether these arguments and this evidence has met the required standard of credibility. As will be explained in the next section, the cumulative weight of different kinds of arguments for excusing psychopaths (including arguments not explored in this chapter) could strengthen the overall case for excusing psychopaths, even if there is some uncertainty about the soundness of each argument individually.

3.  PCL-​R Psychopathy Some descriptions of the PCL-​R psychopath seem to suggest that psychopaths lack moral understanding on a sentimentalist view. On these descriptions, psychopaths seem emotionally cut off from other people and do not seem to have ever genuinely cared about anyone (Hare 199910). Most people, including most convicted offenders, care about some people, such as their immediate relatives, friends, or partner. Yet psychopaths apparently show no compunction about hurting, abandoning, or using anyone who might try to form an intimate relationship with them. Even the concept of “honour among thieves” seems alien to the psychopath. While, for instance, criminal gang members typically show some loyalty to other members, psychopaths will betray their associates whenever it suits them. On this account of psychopathy, this lack of an emotional bond with others is not simply the result of psychopaths becoming “hardened” through a criminal lifestyle but stems from a personality disorder that prevented them caring about others even in early childhood. While normally developing two-​ year-​olds typically show some aversion to signs of another’s distress, and will typically stop performing an action when they perceive it is hurting someone and are responsive 10  Although Hare denies that this necessarily means that they should be exempt from legal responsibility.

Legal Responsibility   395 to parents’ disapproval of such behavior, toddlers with psychopathic traits seem unfazed by others’ distress or parents’ disapproval (Kahn 2012). There are descriptions of very young children with psychopathic traits showing chilling detachment toward others’ suffering. One child with psychopathic traits calmly watched a sibling in a pool unable to swim, and failed to get help, saying he was “curious” about what drowning would look like; another gradually cut off a cat’s tail, in a series of small amputations, just to see what would happen (Kahn 2012). However, not all individuals diagnosed as psychopaths using the PCL-​R criteria will fit this description of the “archetypal” psychopath (even though this kind of description is often provided by researchers who assume that the PCL-​R is the best tool for diagnosing psychopathy that we currently have). An incapacity to feel empathy and an incapacity to form enduring relationships are just two factors out of 20. Someone could meet the cutoff score for psychopathy, even if they scored 0 on these factors, meaning that their empathy and ability to form relationships was intact (Maibom 2018). Failure, in certain studies, to exclude certain PCL-​R “psychopaths” with intact empathic abilities may explain some of the inconsistent data discussed in section 5. (For a discussion of ways of improving the exclusion criteria when conducting such studies see Rosenberg Larsen 2018).

4.  Psychopathy and Empirical Uncertainty Jalava and Griffiths (2017) claim that most philosophers who discuss psychopathy take the empirical literature about psychopaths’ supposed deficits at face value, without questioning the reliability and validity of the data. They caution against this approach, arguing that before theorists can say anything useful about the moral and legal responsibility of psychopaths, they must first engage critically and in detail with the science. It is futile, according to Jalava and Griffiths, to embark on philosophical and legal arguments, no matter how sophisticated, before the existence of psychopaths’ deficits has been scientifically established. They draw the following analogy: No matter how legally astute they may be, there would be no point in defense and prosecution lawyers debating the legal implications of knife wounds on a corpse before the coroner had established that the wounds existed in the first place. This raises fundamental questions about how interdisciplinary research should be conducted in this area, given that most philosophers and lawyers are not scientifically trained and must rely heavily on the conclusions of scientific experts about the state of the evidence. Must philosophical and legal theorists wait until a consensus among these scientific experts emerges before they can comment on the legal and philosophical implications? Given that judges and juries cannot just throw up their hands and wait until science progresses, there has to be some kind of default position

396   Elizabeth Shaw in the meantime. What that default position should be is an important ethical and legal question. A second key question concerns the level of certainty that should be required about psychopaths’ deficits before moral and legal consequences should follow. The required level of certainty might well differ, depending on the nature of these consequences, for example, moral versus legal responsibility, full exemption from responsibility versus diminished responsibility, mitigation of sentence versus preventative incapacitation. Thirdly, theorists can also usefully respond to a division of opinion between scientists about how to interpret the evidence by mapping out the legal and philosophical implications of different possible scenarios, that is, if interpretation (a) is correct then certain legal and philosophical implications would follow, and if interpretation (b) is correct then other implications would follow. This third point could generate proposals about a practical way forward, if it were combined with conclusions about the first two questions concerning the required levels of certainty for different legal/​moral consequences and the conclusions about what the “default” position should be. This chapter will focus on the required standard of credibility for holding psychopaths criminally responsible, given that holding people criminally responsible can warrant inflicting serious harm on them and is therefore of considerable practical importance. Arguably, the standard for holding psychopaths morally (but not criminally) responsible might be lower, given the less-​serious practical consequences of doing so.11 Given that moral responsibility is a precondition for criminal responsibility on many influential theories of criminal law and punishment (for overviews see Duff and Hoskins 2017), it is worth spelling out the implications of endorsing both (a) the principle that moral responsibility is required for criminal responsibility and (b) the idea of having a lower standard of credibility for moral versus criminal responsibility. This would imply that, while doubts about the justification for holding psychopaths morally responsible might not preclude holding psychopaths morally responsible (without imposing criminal responsibility), these same doubts about psychopaths’ moral responsibility might preclude holding them criminally responsible. In general, the default principle, in the context of criminal responsibility, is the presumption of innocence. The principle that the prosecution bears the burden of proving the accused’s guilt beyond reasonable doubt is regarded as fundamental in Anglo-​ American legal systems. It has been vividly described as the “golden thread” running throughout the “web” of English criminal law.12 The “beyond reasonable doubt” principle is also, with few exceptions, widely endorsed by legal theorists, although there is still debate about the best justification for the principle and how “reasonable doubt” 11 

Haji (1998) has suggested the opposite—​arguing that it is easier to justify holding psychopaths legally rather than morally responsible. Furthermore, the requirements for some conceptions of moral responsibility might be more demanding than others (e.g., the “basic desert” sense of moral responsibility versus the “moral appraisability” sense, see Pereboom 2006). 12  Woolmington v. DPP [1935] AC 462, 481.

Legal Responsibility   397 should be interpreted (see, e.g., Picinali 2018; Reiman and Van Den Haag 1990). The more general idea that the accused’s guilt must be established to a high degree of certainty is recognized across a wide range of jurisdictions, despite variations in how the principle is formulated (Clermont 2002). This chapter will suggest that one way of dealing with the conflicting and puzzling evidence about psychopaths’ impairments is to apply the beyond reasonable doubt standard, so that if the arguments and evidence are strong enough to raise a reasonable doubt about whether they lack the capacity for moral understanding, then psychopaths should not be held criminally responsible. Given the complexity of the issues, it is beyond the scope of this paper to provide a full justification for applying the beyond reasonable doubt standard in this context. Instead, it will consider two of the main objections that might be raised against this approach and provide some reasons for thinking that these objections can be overcome. The first objection is that “reasonable doubts” concerning the accused’s mental capacities should be treated differently from doubts about other parts of the prosecution’s case. Rather than requiring the prosecution to prove the accused’s sanity beyond reasonable doubt, legal systems typically start with the default position that the accused is presumed to be of sound mind. A burden of proof is placed on the accused to rebut this presumption. This rule is sometimes called “the presumption of sanity.” The rationale behind this presumption is disputed. It seems to be influenced, at least partly, by practical considerations. Ferguson (2018) calls it an “empirically-​based” presumption, resting on the “common-​sense” idea that most people, most of the time have the mental capacities required to be held responsible. If it were very rare for accused persons to suffer from responsibility-​undermining mental disorders, then compelling the prosecution to bring evidence to rule out this remote possibility in every case would waste the court’s time. However, serious mental disorders are much more common among those accused of crimes than the general population, and although mentally disordered offenders rarely qualify for a mental incapacity defense, this is arguably because such defenses are unjustifiably restrictive (Kelly 2018). Other practical considerations include the difficulties the prosecution might face in getting evidence that the accused had the required mental capacities. However, the Law Commission (2013) has persuasively argued that these practical considerations cannot justify placing an onerous burden on the accused to establish that they lacked the capacities required for responsibility. They point to the fact that these practical considerations apply equally to defenses such as automatism, yet it has proved quite workable to require the prosecution to bear the burden of disproving those defenses (once the accused has discharged the very light burden of just presenting enough evidence to raise the defense as a live issue). A more theoretical argument for the “presumption of sanity” is that it shows respect for the accused’s autonomy to prove his “insanity” before making him eligible for hospitalization, which would amount to a serious “indignity” if he were sane (Brudner 1997, 308. Related theoretical arguments are considered and dismissed in Fine and Kennett 2004). However, it is not clear that hospitalizing a sane person is worse than

398   Elizabeth Shaw imprisoning a mentally disordered person. If the accused has raised a mental incapacity defense (which is usually the case—​Law Commission 2013, p. 180), this suggests that the accused has decided imprisonment is the worse option, and it is not clear why the state should try to “protect” them from the consequences of that decision, based on the state’s view of which option is worse. Furthermore, decisions about whether hospitalization and treatment are required can be made separately from the determination of criminal responsibility, based on different criteria (Law Commission 2013). The second objection focuses on my suggestion that the beyond reasonable doubt standard should apply to the philosophical positions (e.g., sentimentalism) that are part of the overall case for excusing the psychopath, not just to the empirical evidence. Various writers have argued that the beyond reasonable doubt standard should apply to theoretical claims, not just empirical ones (see, e.g., Pereboom 2006; Vilhauer 2009; Shaw 2014; Caruso 2018). One plausible defense of this approach is that the beyond reasonable doubt standard rests on a more fundamental principle about which there is widespread agreement among theorists from a range of different perspectives. Vilhauer (2009, 2012) has argued that this fundamental principle is the idea that there is a powerful presumption against inflicting serious harm on others, and that this presumption can only be displaced if there is a high degree of certainty that doing so is justified. It would be arbitrary, on this view, just to require a high degree of certainty about empirical considerations and not theoretical ones. Vilhauer argues that if there is a philosophically valuable debate about an issue, then there must be a reasonable doubt about that issue. This seems plausible. If reasonable people, with relevant expertise, who have thought seriously about a question still disagree about how to answer it, it seems natural to say that there must be a reasonable doubt about the answer. There is a philosophically valuable debate about whether rational faculties alone are sufficient for moral understanding, or whether empathy is also required. Different positions might be taken about how much certainty should be required that psychopaths lack moral understanding before they should be excused. I have suggested that raising a reasonable doubt as to psychopaths’ capacity for moral understanding should be enough, but it might be argued that the threshold should be somewhat higher. Similarly, different positions might be taken as to how to interpret “reasonable doubt.” The fact that reasonable experts in the area who have seriously considered the matter have endorsed a particular argument for excusing psychopaths might be thought enough to raise a reasonable doubt. However, even if it were thought that one line of argument endorsed by one group of experts had not met the required threshold of certainty on its own, if we consider several lines of argument for excusing psychopaths (e.g., arguments based on empathic deficits, arguments based on evaluative deficits, arguments based on control deficits etc.), and each argument raised some doubt whether psychopaths were responsible, the cumulative weight of these doubts could meet the threshold for excusing them.

Legal Responsibility   399

5.  Evidence of Lack of Moral Understanding This section will summarize some of the key empirical findings concerning psychopaths’ moral understanding and emotional capacities. Firstly, it will briefly discuss one type of study that attempts to measure psychopaths’ moral understanding directly, by investigating psychopaths’ ability to appreciate certain features of moral norms supposedly captured by the “moral/​conventional” distinction. It will then discuss a second type of study that may indirectly shed light on psychopaths’ moral understanding by investigating emotional capacities that arguably underlie or enable the acquisition of moral understanding. These studies have generated complex and seemingly conflicting results. However, after describing each type of study, an interpretation of the evidence will be suggested that more or less reconciles these results and that is broadly consistent with the hypothesis that (at least some) psychopaths lack moral understanding. (The results of some studies may not fit with this interpretation due to their use of an over-​ inclusive definition of psychopathy, which may fail to exclude certain individuals with relatively intact empathy, as noted in section 3.)

5.1  The Moral/​Conventional Distinction One type of moral reasoning task relies on the moral/​ conventional distinction. Although, this paradigm has been criticized (e.g., Kelly 2007; Shoemaker 2011), many theorists and cognitive scientists accept that it can shed some light on psychopaths’ moral understanding (e.g., Blair 1995; Fine and Kennett 2004; Levy 2007, 2014). According to this paradigm, violations of moral norms, for example, norms prohibiting assault, tend to have the following characteristics: (1) they tend to be more serious than transgressing conventional norms, (2) they are wrong at least partly because of harm to the victim, (3) they are wrong independent of whether there is any rule against them, and (4) they would be wrong even if an authority figure permitted them. In contrast, breaches of conventional norms (e.g., norms about the clothes to be worn as part of a school uniform) tend to be less serious than moral transgressions, tend not to be harm-​ based, and tend to be rule-​dependent and authority-​dependent. Blair and others, who pioneered research on psychopaths’ understanding of the moral/​conventional distinction, concluded that psychopaths give abnormal responses to questions about it, suggesting that they may not genuinely understand the distinction (Blair et al. 2001, 1995; Blair 1997). However, overall the results of studies in this area have been mixed. Psychopaths draw some distinction between moral and conventional transgressions based on the seriousness characteristic, but to a lesser extent than non-​psychopaths (Blair 1995, 1997). When psychopaths are asked to come up with

400   Elizabeth Shaw a reason on their own why moral transgressions are wrong, they are much less likely to invoke harm-​based considerations (Arsenio and Fleiss 1996; Blair 1997). However, if psychopaths are explicitly asked whether an example of an immoral act causes harm, they tend to answer correctly (Aharoni et al. 2012, 2014). When asked whether the action would still be wrong if there were no rule against it, psychopaths are less likely than controls to distinguish between moral and conventional transgressions on this basis (Blair 1995, 1997). However, when psychopaths are presented with eight transgressions and are explicitly told that half are “moral” and half “conventional,” psychopaths can group the transgressions into each category as accurately as non-​psychopaths (Aharoni et al. 2012, 2014).

5.2  The Moral/​Conventional Distinction: Interpreting the Evidence It is now relatively often suggested that Aharoni et al’s. findings (2012, 2014) discredit Blair’s claim that psychopaths cannot draw the moral/​conventional distinction (see, e.g., Barnes 2018; Jurjako and Malatesti 2018, p. 1013; Jalava and Griffiths 2017, p. 5; and Schlaich Borg and Sinnott-​Armstrong 2013, p. 124). However, there are interpretations of the findings which are consistent with Blair’s claim that psychopaths lack genuine understanding of the features of moral norms that the moral/​convention distinction aims to capture. Levy (2014) points out that (a) Aharoni et al. asked psychopaths how others would categorize the transgressions, which may invite psychopathic participants to “parrot” the expected answers, even if the psychopaths themselves do not endorse or grasp the nature of moral norms; and (b) Aharoni et al. told participants that half of the short list of transgressions were moral and half conventional, which gives them a substantial clue as to what the “right” answer would look like, making it easier to complete the task, even if the psychopathic participants lacked genuine understanding. In contrast, Blair’s design asked participants for their own views, and did not give them this clue. Apart from these empirical debates about how to interpret psychopaths’ responses to the moral/​conventional test, there are theoretical concerns about whether the moral/​conventional distinction accurately captures the nature of moral norms (e.g., Shoemaker 2011; Kelly 2007). Such concerns are sometimes cited as grounds to doubt any argument for excusing psychopaths that relies on psychopaths’ supposed inability to draw this distinction (e.g., Godman and Jefferson 2017). However, while these concerns suggest that the theoretical underpinnings of the moral/​conventional test need to be refined, they do not necessarily undermine the specific examples used by the test to probe psychopaths’ moral understanding. For example, while proponents of the distinction may be mistaken in assuming that moral norms are completely authority-​independent, since the “victim” may have the relevant authority (Shoemaker 2011), this point does not undermine the specific example used in the test that merely assumes, rightly, that the norm prohibiting one child from hitting another child does not depend on whether the

Legal Responsibility   401 teacher gave permission to do this. Similarly, while it may be wrong to assume, in general, that all moral norms are based on welfare, it is plausible that the wrongfulness of the specific moral transgressions that feature in the test depend, at least partly, on considerations concerning the victim’s welfare. There seems to be a plausible interpretation of the moral/​conventional distinction studies that suggests that psychopaths may fail to grasp important features of moral norms. It seems particularly telling that they often fail to come up with welfare-​based justifications when asked why transgressions involving harm to others are wrong. When asked why hitting or kicking a child to make her cry is wrong, some psychopaths gave answers such as “it’s not socially acceptable” (Blair 1995, p. 212)—​the same sort of reason they give for why conventional transgressions are wrong. This is just what would be predicted by the hypothesis that (at least some) psychopaths have an inability to empathize that prevents them from grasping what is morally wrong with hurting others and leads them to see moral norms as a set of arbitrary prohibitions equivalent to mere conventions. This is not undercut by the finding that psychopaths can respond correctly when explicitly asked whether an act harms the victim, as this latter question does not probe whether psychopaths themselves consider the victim’s welfare relevant to moral wrongfulness.

5.3  Emotional Capacities A wide range of studies have been conducted on emotional capacities that arguably underlie or enable the acquisition of moral understanding. These emotional capacities include, firstly, “affective theory of mind”—​the ability to infer from the context what another is feeling; secondly, the ability to recognize another’s emotions, such as distress and fear, from cues such as facial expression and tone of voice; and thirdly, the ability to feel an automatic, aversive emotional response to the recognition that another is afraid or distressed. Recognizing what others feel is necessary in order to understand how one morally ought to treat them. Evidence that psychopaths could not recognize others’ feelings through the means investigated by the studies would not conclusively prove that they have impaired understanding, as they might work out what others feel through different means. However, (as will be explained in more detail in 5.3) evidence of impaired emotion recognition taken together with evidence of impaired emotional responsiveness make more plausible the idea that psychopaths are somehow cut off from others’ emotions and thereby fail to develop a genuine understanding of norms about how others should be treated. Studies on affective theory of mind typically ask participants to work out what various characters in a scene or vignette are feeling, based on information about the context. For example, normal participants may be expected to infer that character A desires object X, and that character B desires the same object, and that A will feel disappointed if B, rather than A, obtains the object. Psychopaths seem to perform the same as controls when the scenarios are relatively straightforward, for example, based on a simple cartoon image

402   Elizabeth Shaw (Sebastian 2012; Shamay-​Tsoory 2010). However, psychopaths seem to perform less well when the scenarios and questions are more complex, for example, when asked to record the different feelings characters experience during the course of a 15-​minute video, and when asked about the relationship between two or more characters’ feelings, rather than just being asked about one character (Sharp and Vanwoerden 2014). Furthermore, although psychopaths’ responses to the simple scenarios are similar to non-​psychopaths, they seem to employ different brain regions when making these decisions. In addition, another type of study suggests that psychopaths do not register other people’s perspectives in the automatic way that non-​psychopaths do. When non-​psychopathic participants were asked to count dots on a screen, they were slower at doing so if they could also see a character on the screen from whose viewpoint a different number of dots were visible (Drayton et al 2018). It seems that non-​psychopaths automatically register the character’s perspective, which interferes with their response about what they can see from their own perspective, causing a delayed reaction. However, this time delay was not observed in psychopaths, suggesting that they did not automatically detect the character’s perspective. Some studies on psychopaths’ ability to recognize others’ emotions based on their facial expression or voice have suggested that psychopaths are less able to identify negative emotions such as sadness and fear, compared to non­psychopaths (Jusyte et al. 2015; Blair et al. 2005). However, when psychopaths’ attention is directed in certain ways, they seem to perform normally. For example, in one study in which psychopaths were less accurate than controls at identifying facial emotions, it was observed that psychopaths were looking at the mouth region of the face, which carries less emotional information than the eye area. When psychopaths were explicitly asked to look at the eye area, their performance was normal (Dadds et al. 2006). However, when the images flashed up so quickly that the participant’s direction of gaze could not make a difference, their ability to recognize facial emotions seemed impaired compared with non-​psychopathic controls (Jusyte et al. 2014). There is also some evidence that psychopaths struggle to distinguish real from feigned distress (Dawel, Wright, Dumbleton, and McKone. 2019). Certain physiological reactions (e.g., increased skin conductance and fear-​ potentiated startle) that are associated with aversive states are normally seen in non-​ psychopaths when viewing images of others in distress, suggesting that they find others’ distress aversive. However, most studies on this topic suggest that these physiological responses to distress are deficient in psychopaths (see references in Maibom 2017a, p 1119). Furthermore, in some studies, regions of the brain, such as the amygdala, associated with processing emotional stimuli in non-​psychopaths, did not activate normally when psychopaths were asked to evaluate facial expressions. This difference in amygdala activation between psychopaths and non-​psychopaths appeared even when the emotional images flashed up so quickly that direction of gaze could not make a difference (Viding et al. 2012). Some studies have also shown that adult psychopaths and children with psychopathic traits have significantly lower amygdala volume than non-​psychopaths (Blair 2005). Blair (2019) summarizes the evidence as follows: “In summary, the existing literature relatively reliably indicates reduced responsiveness

Legal Responsibility   403 to facial expressions, particularly distress cues, in children and adults with conduct problems that may be particularly marked in those with psychopathic traits. The regions implicated across studies are not always consistent . . . but the basic finding of reduced neural responsiveness appears robust.”

5.4  Emotional Capacities: Interpreting the Evidence One plausible interpretation that reconciles the mixed evidence on emotional capacities is that, while most non-​psychopaths can automatically infer from the context what others are feeling, recognize others’ distress and find others’ distress aversive, (at least some) psychopaths’ automatic ability to do these things is significantly deficient (Tillem et al. 2019). In specific circumstances, it seems that (at least some psychopaths) can exercise these capacities (non-​automatically), provided their attention is suitably directed by the experimenter, although it remains possible that individuals with severe psychopathy still have abnormal neurological responses, even in these circumstances. Blair (2019) points out that group-​level results may mask the deficits of individual members of the group. An experiment failing to reveal deficient responses to emotional stimuli, under certain conditions, in the psychopathic group overall (whose members had varied degrees of psychopathy) is still compatible with the hypothesis that a subset of individuals with high psychopathic traits still had reduced responses under these conditions. Blair (2019, p. 295) hypothesises that “. . . if the intensity of this stimulus is sufficiently heightened, via an attentional manipulation that increases the emotional stimulus’ representational strength, group differences are reduced (because the individuals with lower psychopathic traits reach an asymptote level in responding).” Furthermore, it seems plausible that deficits in automatic emotion-​processing leads, firstly, to an impaired ability to infer others’ emotions from the context when the situation is relatively complex and, secondly, to an impaired ability to recognize emotions, such as another’s distress, when there is no opportunity to take the time to focus on certain cues. The hypothesis that psychopaths have reduced emotional responses to others’ distress is supported by the evidence that psychopaths show volume reduction in the amygdala (a brain area associated with these responses). The next question is whether this interpretation of the evidence is consistent with arguments that (at least some) psychopaths have grounds for an excuse based on the idea that their emotional incapacities deprive them of genuine moral understanding. Based on research, such as that discussed earlier showing that psychopaths sometimes have normal behavioral and neurological responses when performing tasks that purport to test morally relevant capacities, Jurjako and Malatesti (2018) claim that “psychopaths seem to have a general capacity to appreciate moral considerations” (p. 1016).13 13  Earlier in the article, Jurjako and Malatesti make the more modest claim that the current evidence does not warrant the conclusions that “psychopaths lack moral understanding” or that they “should be excused from criminal responsibility.” In other words, at this earlier point in their article, they merely

404   Elizabeth Shaw If psychopaths perform normally at least some of the time, this indicates, according to Jurjako and Malatesti, that they have the general capacities required for moral responsibility. Nevertheless, since psychopaths appear to show deficits in certain contexts, Jurjako and Malatesti maintain that it is still an open question whether individual psychopathic wrongdoers had the specific capacities required for responsibility at the time of their wrongful acts. They suggest, however, that occasions when psychopaths lack these specific capacities are likely to be rare, for the following reasons. Firstly, even if psychopaths are unable to automatically register morally relevant information based on certain cues (e.g., facial expressions), they might be able to register this information by consciously attending to these cues. Secondly, even if they cannot register morally relevant information from certain sources (e.g., facial expressions) they might gain this information from other sources. Jurjako and Malatesti suggest, for example, that a psychopath staging an armed bank robbery might not be able to detect the bank clerk’s fear from her facial expression, but could infer it from other signs, for example, screaming, putting her hands in the air, etc. (Jurjako and Malatesti 2018). To show that psychopaths’ (supposed) deficits might not undermine their responsibility, they draw an analogy with a color-​blind person who is responsible for jaywalking, despite his inability to see the red light, because he should have worked out other ways of telling whether the light was red, based on, for example, the order of the lights and other people’s behavior. Only in unusual circumstances, such as the authorities unforeseeably altering the order of the lights, would the color-​blind person be relieved of responsibility. Similarly, they argue, the psychopath would only be non-​responsible in relatively unusual circumstances. However, there are alternative interpretations of the data, which, if correct, would increase the likelihood that an individual with psychopathy was non­responsible for his or her act. If, as suggested before, the evidence indicates that (at least some) psychopaths do not automatically register and respond to others’ morally relevant emotions, this might have undermined their moral development so that they failed to develop into moral agents in the first place. For example, Blair (2017) has argued that psychopathy is a developmental disorder that is often apparent from earliest childhood. As already noted, it has been frequently observed, that, unlike normally developing children, children with psychopathic tendencies (or at least those children with the severe form of the disorder) seem unresponsive to the attempts of parents and teachers to socialize them, seem unable to form loving relationships with others, including primary caregivers, and seem, even as toddlers, capable of inflicting pain without the inhibition about doing this that is seen in non-​psychopathic toddlers (see, e.g., Hare 1999; Kahn 2012). Blair (2017) describes a body of developmental research that supports the proposition that an automatic aversive response to perceiving another’s distress is a crucial part of moral development (see also references in Cushman et al. 2017). When a normally developing child perpetrates or witnesses care-​based transgressions, which involve hurting someone else,

make the negative claim that those who argue for the non-​responsibility of psychopaths have not (yet) proved their case (2018, p. 1013).

Legal Responsibility   405 the child will “learn the badness of the care-​based transgressions because of the pairing of the victim’s distress [which the child observing it finds aversive] with the commission (or observation of someone else committing) the care-​based transgression” (Blair 2017). Caregivers also typically reinforce, in the mind of the child, the idea that such care-​based transgressions are bad by pointing to the victim’s distress (see, e.g., Nucci and Nucci 1982), and it seems plausible that a failure to form a bond of affection with the caregiver (or indeed with anyone) undermines these attempts at reinforcement. Now, it might be objected that the kind of emotionally informed moral understanding that some philosophers have argued is necessary for responsibility is much more complex than this basic association of an automatic aversion to others’ distress with the “badness” of an act. However, in response, it is not being suggested that this basic emotional association is all that moral understanding consists in, but, rather, that it may be the (empirically) necessary foundation upon which sophisticated empathy and genuine moral understanding are built. Similarly, John Deigh (1995, p. 743) argues that the kind of sophisticated empathy necessary for genuine moral understanding emerges gradually “from early experiences of shared feeling . . . [Empathy takes] increasingly mature forms as one’s understanding of what it is to be a human being and to live a human life deepens.” If psychopathy is a disorder of moral development, as described, then adult psychopaths could not be held responsible on the basis that they should (i.e., were under a moral obligation to) have worked out what other people were feeling through non-​ automatic routes, because one needs to be a moral agent in the first place in order to recognize what one’s moral obligations are. This point is brought out in the following passage from Kant. There are certain moral endowments such that anyone lacking them could have no duty to acquire them. They are moral feeling, conscience, love of one’s neighbour and respect for oneself (self esteem). There is no obligation to have these because they lie at the basis of morality . . . All of them are natural predispositions of the mind for being affected by concepts of duty. To have these predispositions cannot be considered a duty; rather, every . . . [moral agent] has them, and it is by virtue of them that he can be put under an obligation . . . For if [a person] had no conscience, he could not even conceive of the duty to have one. . .” (Gregor trans. 1991, p. 400)14

If the previously outlined account of moral development is correct, then the “moral endowments” involved in the normal process of forming a conscience may include automatic emotional responses to others’ distress, experienced at a critical stage of early development. Evidence that (at least some) adult psychopaths have normal neurological

14  I Kant (Gregor trans. 1991). This passage is also cited in Fine and Kennett (2004). This passage in Kant may appear to contradict Kant’s seemingly cognitivist position discussed in the previous section. For a discussion that can shed light on how to interpret these conflicting strands in Kant’s moral theory, see M. Midgley, “The Objection to Systematic Humbug,” Philosophy 53, no. 204 (1978): 147.

406   Elizabeth Shaw responses to others’ distress, which are not engaged automatically, but which emerge under specific experimental conditions, does not show that the required responses were engaged at the relevant developmental stage. However, this evidence might suggest that psychopaths have the ability to acquire moral understanding, if they receive interventions that help to elicit the necessary emotional responses during moral learning. An analogy might be drawn with learning language. Someone might have the neurological architecture necessary to acquire language, under certain conditions, but they do not have the ability to speak, unless their ability to acquire language has been engaged under those conditions. Even if reduced amygdala volume in psychopaths is evidence of the “use it or lose it” principle (Jurjako 2019), if psychopaths did not use the necessary brain regions, because they were not exposed in early childhood to the conditions in which those brain areas would be activated, they cannot fairly be blamed for the underdeveloped state of their brains and the impaired emotional capacities that may result.

6.  An Individualized Approach to Psychopathy There is a tendency among theorists to frame some of their arguments in terms of the legal and moral responsibility of “psychopaths” as a class, or in terms of a subcategory of psychopaths (e.g., severe psychopaths, or unsuccessful psychopaths) as a class (see, e.g., Jalava and Griffiths 2017; Glannon 2017; Glenn et al. 2011; Duff 1977). This approach might seem odd from a medico-​legal perspective, since mental incapacity defenses generally require an individualized assessment of the accused’s capacities at the time of crime. Simply showing that the accused belonged to a class of people with a particular medical diagnosis, by itself, generally cannot settle the question of whether that individual was criminally responsible. Now, if the term “psychopath” were being used as a philosophical (rather than medical) concept, defined in terms of the complete lack of certain capacities that were argued to be prerequisites for moral agency (see Duff 1977), it would make sense to talk of the non-​responsibility of psychopaths in general. However, many theorists use the term “psychopath” to refer to a medical diagnosis, such as “PCL-​R psychopathy,” and it is relatively unusual for it to be the case that all individuals with a particular medical diagnosis are never responsible. This could be the case for conditions such as severe intellectual disabilities, advanced dementia, or advanced Huntingdon’s disease, but it is not true of many other conditions such as depression, bipolar disorder, anxiety disorders, or schizophrenia, which can undermine the responsibility of some sufferers on some occasions. Such considerations suggest that the criminal and moral responsibility of individual PCL-​R psychopaths should be assessed on a case-​by-​case basis (Jalava and Griffiths 2017; Shaw 2016, 2009). Indeed, even if one were employing the word

Legal Responsibility   407 “psychopathy” as a philosophical concept, defined in terms of the absence of certain prerequisites for responsibility, it would still need to be established on a case-​by-​case basis that an individual wrongdoer met this definition of psychopathy. When arguing for an individualized approach, Jurjako and Malatesti (2018) helpfully distinguish between general capacities and specific capacities. A general capacity is an ability that an agent can exercise in a suitably wide range of contexts, but not necessarily in all circumstances; whereas a specific capacity is an ability that an agent can exercise on a particular occasion. For example, a person may have the general capacity to play the piano but might lack the specific capacity to do so in certain circumstances, for example, when she is asleep, or has no access to a piano. They rightly maintain that, when considering whether a psychopath is morally/​criminally responsible for committing a wrongful/​criminal act, it is relevant to consider the psychopath’ s specific capacities to understand and act in accordance with moral/​legal norms at the time of performing the act.

7. Conclusion This chapter drew attention to a question that has been paid insufficient attention in the legal philosophical literature on the criminal responsibility of psychopaths—​who should bear the burden of proving that they are or are not responsible and to what standard? It provided some reasons for thinking that raising a reasonable doubt as to psychopaths’ capacity for moral understanding should be enough to justify exempting them from criminal responsibility and punishment. (Although it may still of course be necessary to hospitalize or treat them—​if effective treatments can be developed—​to prevent them from harming others—​for discussion of this issue, see Shaw 2018). This chapter summarized relevant research on psychopathy and discussed an interpretation of this research that was consistent with psychopaths being entitled to an excuse on sentimentalist grounds (having also argued that any doubt about sentimentalism should be dealt with in favor of the accused.) Finally, this chapter proposed that it should be decided on a case-​by-​case basis whether someone diagnosed with psychopathy should be excused from criminal responsibility.

References Aharoni, Eyal, Walter Sinnott-​Armstrong, and Kent Kiehl. “Can psychopathic offenders discern moral wrongs? A new look at the moral/​conventional distinction.” Journal of Abnormal Psychology 121, no. 2 (2012): 484–​497. Aharoni, Eyal, Walter Sinnott-​Armstrong, and Kent Kiehl. “What’s wrong? Moral understanding in psychopathic offenders.” Journal of Research in Personality 53 (2014): 175–​181. Barnes, Simon. Re-​evaluating the exclusion of psychopathy from the mental disorder defence in Scots law. Juridical Review 1 (2018): 1–​22.

408   Elizabeth Shaw Baron-​ Cohen, Simon et al. Empathy in Autism Spectrum Conditions. https://​ www.autismresearchcentre.com/​project_​1_​empathy . 2019. Accessed November 2019. Blair, James. “A cognitive developmental approach to morality: Investigating the psychopath.” Cognition 57, no. 1 (1995): 1–​29. Blair, James. “Moral reasoning and the child with psychopathic tendencies.” Personality and Individual Differences 22, no. 5 (1997): 731–​739. Blair, R. J. R., Mitchell, D. & Blair, K. (2005). The psychopath. Emotion and the brain. Blackwell. Blair, R. James R. (2008). The cognitive neuroscience of psychopathy and implications for judgments of responsibility. Neuroethics 1 (3):149–​157. Blair, James. “Emotion-​based learning systems and the development of morality.” Cognition 167 (2017): 38–​45. Blair, James. “Dysfunctional neurocognition in individuals with clinically significant psychopathic traits.” Dialogues in Clinical Neuroscience 21, no. 3 (2019): 291–​299. Blair, James, Derek Mitchell, and Karina Blair. The psychopath: Emotion and the brain. Blackwell Publishing, 2005. Blair, James, Jey Monson, and Norah Frederickson. “Moral reasoning and conduct problems in children with emotional and behavioural difficulties.” Personality and Individual Differences 31, no. 5 (2001): 799–​811. Bollard, Mara. “Psychopathy, autism and questions of moral agency.” In Alexandra Perry & C. D. Herrera (eds.), Ethics and Neurodiversity (2013). Newcastle upon Tyne, UK: 238–​259. Brudner, Alan. “Guilt under the charter: The lure of parliamentary supremacy.” Criminal Law Quarterly 40 (1997): 287. Caruso, Gregg D. “Justice without retribution: An epistemic argument against retributive criminal punishment.” Neuroethics 13 (2020): 1–​16. Clermont, Kevin, and Emily Sherwin. “A comparative view of standards of proof.” American Journal of Comparative Law 50 (2002): 243. Cushman, Fiery, Victor Kumar, and Peter Railton. 2017. “Moral learning: Psychological and philosophical perspectives.” Cognition (2017): 1–​10. Dadds, Mark R., Yael Perry, David J. Hawes, Sabine Merz, Alison C. Riddell, Damien J. Haines, Emel Solak, and Amali I. Abeygunawardane. “Attention to the eyes and fear-​recognition deficits in child psychopathy.” The British Journal of Psychiatry 189, no. 3 (2006): 280–​281. Dawel, Amy, Luke Wright, Rachael Dumbleton, and Elinor McKone. “All tears are crocodile tears: Impaired perception of emotion authenticity in psychopathic traits.” Personality Disorders: Theory, Research, and Treatment 10, no. 2 (2019): 185–​197. Deigh, John. “Empathy and universalizability.” Ethics 105 (1995): 743–​763. Dineen, Katy. “Kant, emotion and autism: Towards an inclusive approach to character education.” Ethics and Education 14, no. 1 (2019): 1–​14. Drayton, Lindsey A., Laurie R. Santos, and Arielle Baskin-​Sommers. “Psychopaths fail to automatically take the perspective of others.” Proceedings of the National Academy of Sciences 115, no. 13 (2018): 3302–​3307. Duff, Antony. “Psychopathy and moral understanding.” American Philosophical Quarterly 14, no. 3 (1977): 189–​200. Duff, R. A. (2001). Punishment, Communication, and Community. OUP: USA. Duff, Antony, and Zachary Hoskins. “Legal punishment.” (2017). The Stanford Encyclopedia of Philosophy. https://​plato.stanford.edu/​entries/​legal-​punishment/​. Accessed November 2019.

Legal Responsibility   409 Fine, Cordelia, and Jeanette Kennett. “Mental impairment, moral understanding and criminal responsibility: Psychopathy and the purposes of punishment.” International Journal of Law and Psychiatry 27, no 5 (2004): 425–​443. Fischer, John Martin, & Ravizza, Mark (1998). Responsibility and Control: A Theory of Moral Responsibility. Cambridge University Press. Fischer, J. M., & Ravizza, M. Responsibility and control: A theory of moral responsibility. Cambridge University Press, 2011. Focquaert, Farah, Andrea Glenn, and Adrian Raine. “Psychopathy and free will from a philosophical and cognitive neuroscience perspective.” In W. Glannon (Ed.), Free Will and the Brain (2015): 103–​124. Fox, Adam R., Trevor H. Kvaran, and Reid Griffith Fontaine. “Psychopathy and culpability: How responsible is the psychopath for criminal wrongdoing?” Law & Social Inquiry 38, no. 1 (2013): 1–​26. Glannon, Walter. “Psychopathy and responsibility: Empirical data and normative judgments.” Philosophy, Psychiatry, & Psychology 24, no. 1 (2017): 13–​15. Glenn, Andrea, Adrian Raine, and William Laufer. “Is it wrong to criminalize and punish psychopaths?” Emotion Review 3, no. 3 (2011): 302–​304. Gill, Michael B. “Moral rationalism vs. moral sentimentalism: Is morality more like math or beauty?” Philosophy Compass 2, no. 1 (2007): 16–​30. Godman, Marion & Jefferson, Anneli (2017). On Blaming and Punishing Psychopaths. Criminal Law and Philosophy 11 (1):127–​142. Haji, Ishtiyaque. “On psychopaths and culpability.” Law and Philosophy 17, no. 2 (1998): 117–​140. Hare, Robert. The Hare psychopathy checklist revised, 2nd edn. Multi-​Health Systems, Toronto, 2003. Hare, Robert D. Without conscience: The disturbing world of the psychopaths among us. Guilford Press, 1999. Hume, David. An enquiry concerning the principles of morals. Selby-​Bigge, ed. Clarendon Press, Oxford, 1975. Jalava, Jarkko, and Stephanie Griffiths. “Philosophers on psychopaths: A cautionary tale in interdisciplinarity.” Philosophy, Psychiatry, & Psychology 24, no. 1 (2017): 1–​12. Jurjako, Marko. “Is psychopathy a harmful dysfunction?” Biology & Philosophy 34, no. 1 (2019): 5. Jurjako, Marko, and Luca Malatesti. “Neuropsychology and the criminal responsibility of psychopaths: Reconsidering the evidence.” Erkenntnis 83, no. 5 (2018): 1003–​1025. Jusyte, A., S. V. Mayer, E. Künzel, M. Hautzinger, and M. Schönenberg. “Unemotional traits predict early processing deficit for fearful expressions in young violent offenders: An investigation using continuous flash suppression.” Psychological Medicine 45, no. 2 (2015): 285–​297. Kahn, Jennifer, “Can you call a 9-​year-​old a psychopath” (2012). https://​www.nytimes.com/​ 2012/​05/​13/​magazine/​can-​you-​call-​a-​9-​year-​old-​a-​psychopath.html Kant, Immanuel. “The doctrine of virtue,” in Immanuel Kant, The Metaphysics of Morals, trans. M. Gregor, Cambridge University Press, 1991. Kauppinen, Antti. Moral Sentimentalism. https://​plato.stanford.edu/​entries/​moral-​ sentimentalism/​. 2014. Accessed November 2019. Kelly, Daniel, Stephen Stich, Kevin J. Haley, Serena J. Eng, and Daniel M.T. Fessler. “Harm, affect, and the moral/​conventional distinction.” Mind & Language 22, no. 2 (2007): 117–​131.

410   Elizabeth Shaw Kelly, Erin. The limits of blame: Rethinking punishment and responsibility. Cambridge, MA: Harvard University Press, 2018. Kennett, Jeanette. “Autism, empathy and moral agency.” The Philosophical Quarterly 52, no. 208 (2002): 340–​357. Law Commission for England and Wales, Criminal Liability: Insanity and Automatism, A Discussion Paper. Law Com DP, 2013 Levy, Neil. “The responsibility of the psychopath revisited.” Philosophy, Psychiatry, & Psychology 14, no. 2 (2007): 129–​138. Levy, Neil. “Psychopaths and blame: The argument from content.” Philosophical Psychology 27, no. 3 (2014): 351–​367. Litton, Paul. “Psychopathy and responsibility theory.” Philosophy Compass 5, no. 8 (2010): 676–​688. Maibom, Heidi L. “The mad, the bad, and the psychopath.” Neuroethics 1, no. 3 (2008): 167–​184. Maibom, Heidi. “Psychopathy: Morally incapacitated persons.” Handbook of the Philosophy of Medicine (2017a): 1109–​1129. Maibom, Heidi. “Introduction to philosophy of empathy.” The Routledge Handbook of Philosophy of Empathy (2017b): 1–​9. Maibom, Heidi. “What can philosophers learn from psychopathy?” European Journal of Analytic Philosophy 14, no. 1 (2018): 63–​78. Malatesti, Luca, and John McMillan, eds. Responsibility and psychopathy: Interfacing law, psychiatry and philosophy. Oxford University Press, 2010. Matravers, Matt. De-​moralising retributivism: agency, blame and humanity in criminal law theory and practice. In J. Jacobs, & J. Jackson (Eds.), Routledge Handbook of Criminal Justice Ethics (pp. 144–​162). Routledge. Midgley, Mary. “The objection to systematic humbug.” Philosophy 53, no. 204 (1978): 147–​169. Montgomery, Charlotte, Carrie Allison, Meng-​Chuan Lai, Sarah Cassidy, Peter Langdon, and Simon Baron-​Cohen. “Do adults with high functioning autism or Asperger Syndrome differ in empathy and emotion recognition?” Journal of Autism and Developmental Disorders 46, no. 6 (2016): 1931–​1940. Morse, Stephen J. “Psychopathy and criminal responsibility.” Neuroethics 1, no. 3 (2008): 205–​212. Nelkin, Dana. Fine Cuts of Moral Agency: Dissociable Deficits in Psychopathy and Autism. In S. Matthew Liao & Collin O'Neil (eds.), Current Controversies in Bioethics. New York (2017): 47–​66. Pereboom, Derk. Living without free will. Cambridge University Press, 2006. Picinali, Federico. “Can the reasonable doubt standard be justified? A reconstructed dialogue.” Canadian Journal of Law & Jurisprudence 31, no. 2 (2018): 365–​402. Radcliffe, Elizabeth S. “Hume on motivating sentiments, the general point of view, and the inculcation of.” Hume Studies 20, no. 1 (1994): 37–​58. Ramadan, Hisham M. “Reconstructing reasonableness in criminal law: Moderate jury instructions proposal.” Journal of Legislation 29 (2002): 233. Reiman, Jeffrey, and Ernest Van Den Haag. “On the common saying that it is better that ten guilty persons escape than that one innocent suffer: Pro and con.” Social Philosophy and Policy 7, no. 2 (1990): 226–​248. Rosenberg Larsen, Rasmus. “False-​positives in psychopathy assessment: Proposing theory-​ driven exclusion criteria in research sampling.” European Journal of Analytic Philosophy 14, no. 1 (2018): 33–​52.

Legal Responsibility   411 Sebastian, Catherine, Eamon McCrory, Charlotte Cecil, Patricia Lockwood, Stéphane De Brito, Nathalie Fontaine, and Essi Viding. “Neural responses to affective and cognitive theory of mind in children with conduct problems and varying levels of callous-​unemotional traits.” Archives of General Psychiatry 69, no. 8 (2012): 814–​822. Sebastian, Catherine L., Nathalie M.G. Fontaine, Geoffrey Bird, Sarah-​Jayne Blakemore, Stephane A. De Brito, Eamon .JP. McCrory, and Essi Viding. “Neural processing associated with cognitive and affective Theory of Mind in adolescents and adults.” Social Cognitive and Affective Neuroscience 7, no. 1 (2011): 53–​63. Shamay-​Tsoory, Simone G. Hagai Harari, Judith Aharon-​Peretz, and Yechiel Levkovitz. “The role of the orbitofrontal cortex in affective theory of mind deficits in criminal offenders with psychopathic tendencies.” Cortex 46, no. 5 (2010): 668–​677. Sharp, Carla, and Salome Vanwoerden. “Social cognition: Empirical contribution: The developmental building blocks of psychopathic traits: Revisiting the role of theory of mind.” Journal of Personality Disorders 28, no. 1 (2014): 78–​95. Shaw, Elizabeth. “Psychopaths and criminal responsibility.” (2009): 13(3), 497–​502. Shaw, Elizabeth. “Psychopathy, moral understanding and criminal responsibility.” European Journal of Current Legal Issues 22, no. 2 (2016). Shaw, Elizabeth. “The treatment of psychopathy: Conceptual and ethical issues.” In Neurolaw and Responsibility for Action: Concepts, Crimes and Courts, pp. 244–​263. Cambridge University Press, 2018. Shaw, Elizabeth. “Free will, punishment and criminal responsibility.” (2014). Available at: https://​www.era.lib.ed.ac.uk/​handle/​1842/​9590. Accessed November 2019. Shoemaker, David W. (2011). Psychopathy, Responsibility, and the Moral/​ Conventional Distinction. Southern Journal of Philosophy 49 (s1):99–​124. Tillem, Scott, Shou-​An Chang, and Arielle Baskin-​Sommers. (2019 in press). Comparison of socio-​affective processing across subtypes of antisocial psychopathology. In Farah Focquaert, Elizabeth Shaw, & Bruce Waller (eds.), The Routledge Handbook of the Philosophy and Science of Punishment: Routledge. Vilhauer, Benjamin. “Free will and reasonable doubt.” American Philosophical Quarterly 46, no. 2 (2009): 131–​140. Vilhauer, Benjamin. “Taking free will skepticism seriously.” The Philosophical Quarterly 62, no. 249 (2012): 833–​852.

CHAPTER 20

Resp onsibi l i t y a nd Di stribu tiv e J u st i c e Richard Arneson

The thought that fair shares for individuals depend on whether they behave responsibly is popular and plausible. But the response that this way of thinking is narrow-​minded and misses something important is also strongly appealing. This chapter outlines responsibility-​oriented conceptions of distributive justice and some criticisms they have attracted. One critic is Karl Marx. In the letter that became known as Critique of the Gotha Program, Karl Marx wrote that when social progress enables us to move altogether beyond the “narrow horizon of bourgeois right,” society can inscribe on its banner, “From each according to his ability, to each according to his needs!” (Marx 1977: 615). In Marx’s mindset, moving beyond the bourgeois right appears to involve reaching a stage of society in which no rules specifying individual rights and duties need to be promulgated and enforced. Well brought-​up adults will seamlessly coordinate to cooperate with each other as needed. Their patterns of cooperation will bring it about that each individual gives to society according to her ability and takes from society according to her needs. Hence society in this higher phase of communism will fulfill the norm that Marx here announces and can proclaim it on social banners without false advertising. One might imagine that economic production in this utopia takes the form of a giant pot, into which people contribute as they see fit, and from which they withdraw goods as they see fit, without the need for social control beyond individual sensibility. This fascinating passage raises many questions. A flat-​footed literal-​minded response is that so far from being the statement of an uncontroversial social ideal, “From each according to her ability and to each according to her needs” would be unfair and unjust. One simple criticism is that different individuals might find it differentially onerous to contribute according to their abilities. Contributing to social production to the best of your ability might be incredibly painful and difficult for you, and contributing to the

Responsibility and Distributive Justice    413 best of my ability might be incredibly enjoyable and easy for me. So intuitively, fairness would require that an adjustment be made, and that you contribute less and I contribute more. In the paragraph from which the Marx quotation just rehearsed was stated, Marx also remarks that at the stage of social development he is envisaging, labor will have become “life’s prime want.” But that stipulation is compatible with the troubling scenario just described. Labor is the prime want for both of us, as we both seek fulfillment and service to others, nonetheless laboring might still be difficult and painful for you and easy and enjoyable for me. Another concern is that the developed abilities that an individual has at any given time depend in part on ability-​developing and ability-​preserving actions she has taken, or declined to take, in the past. Moreover, intuitively, individuals might be responsible for these actions, in such a way that what distributive justice requires for them and from them now depends in some part on their past exercises of responsibility. Just detecting each person’s present abilities does not suffice to determine what the person fairly owes to society. You and I might have very different developed abilities at present. But the history of how this situation came about might preclude its being the case that “from each according to her abilities, to each according to her needs” identifies what is fair and just for the two of us now. For example, you might have worked assiduously in high school and post-​secondary education while I goofed off. You worked hard and conscientiously to develop your abilities and I did not. Also, you might have taken reasonable prudent precautions in your choice of leisure time activities over the past few years. As a young adult, you climbed mountains while wearing a protective helmet and taking reasonable safety precautions with ropes and chocks. At that same life stage, I skateboarded downhill at speeds that were thrilling and fun but excessively risky for self and others. Having crashed several times, my mental and physical abilities are now less than yours and, more important, less than they would have been had I behaved reasonably over the preceding past few years. There are further questions that arise about the individual actions that individuals take that predictably bring it about that they will have one or another set of abilities and other traits that bear on their capacity and willingness to be socially productive, to serve the common good, to an appropriate extent, and by their lifetime activities. You might see that society faces a looming crisis in the form of a shortage of healthcare workers. Responding, you develop abilities and get the education that makes you ready to take on a career in medicine. I insist on serving the common good by becoming a poet, and work to develop my modest talents at poetry, for which there is little social need. There is already a surfeit of poets. Again, all else equal, “from each according to his ability” is not a fair standard for job allocation between you and me, given how we have positioned ourselves for entry into the labor market. Same goes, mutatis mutandis, for “to each according to his needs.” The needs that an individual has at a time are sometimes the result of unforeseeable events that fall on him in ways that are beyond his power to control, as when a sudden lurch in economic conditions in a distant part of the globe reverberates widely and results in the loss of his

414   Richard Arneson job. At the other extreme, my need for narcotics today may be the result of long-​term drug use on my part. My need for expensive medical care to treat my heart disease may be directly traceable to my diet and exercise habits sustained long-​term. In between the extremes, what present themselves to the individual as implacable needs may stem from variously reasonable or unreasonable choices made in the face of good or bad fortune that itself is traceable to conditions for which the individual might bear various degrees of responsibility. This is not to say that society—​the members of a political community regarded together—​ought to turn a cold shoulder to individuals who are in bad predicaments of their own making. But when resources are scarce, not all can always be helped. “To each according to his needs” suggests that when people’s needs cannot all be fully satisfied, society’s response should be to bring it about that each person’s important needs are satisfied in the same proportionate degree. But equally meeting, say, half of the needs for medical care of the malnourished child who needs food and medical attention, and of the alcoholic who needs a new liver, and of the tourist who broke his leg wandering around in a remote area and needs an expensive rescue operation, is not obviously what should be done, as it would be if “to each according to her needs” were the applicable norm. On some views, these implications of “to each according to her needs” show that conforming to this norm is quite obviously what ought not to be done (see Anderson 1999 versus Arneson 2000). To be fair, it should be noted that Marx suggested his slogan as a distributive norm that the desirable communist society of the future would satisfy only provided that several conditions are met. We have already mentioned that in the higher phase of communism, according to Marx, labor will have become life’s prime want. Marx also supposes that the subordination of the individual to the division of labor will have ceased and that “all the springs of cooperative wealth flow more abundantly.” The ending of the division of labor brings it about that individuals are free to shift from one to another productive role just as they please. The remaining assumption, to the effect that wealth becomes more abundant, leaves it open to what extent abundance prevails. Marx may have a maximal conception of abundance, as inaugurating an effective end of scarcity, so that available resources exceed the ensemble of demands that people might conceivably make on them (see Cohen 1995). If that is the future state in which society can truthfully write Marx’s slogan on its banner, we are in the realm of utopia, and not Rawlsian realistic utopia, but something more akin to heaven on Earth (Rawls 1999). But in that imagined blissful setting, why care whether individuals contribute according to their abilities, since their contributing less would impose no costs on others, and why limit social distribution to satisfaction of needs, however these are defined, since available resources would suffice to satisfy everyone’s needs and in addition everybody’s heartfelt and trivial desires? To specify conditions in which the implementation of Marx’s slogan could make sense and be feasible is also to specify conditions in which implementing it would be contrary to justice and morality.

Responsibility and Distributive Justice    415 Or so some would hold. Others would vigorously disagree. This is contested terrain. The discussion to this point, criticizing Marx’s view, takes it for granted that what we owe to one another by way of distributive justice rights and duties varies significantly depending on how individuals comport themselves and are disposed to comport themselves.

1.  Key Concepts Distributive justice as understood in this chapter names a broad topic (Cohen 2008). Distributive justice principles specify what is a fair distribution of benefits and burdens across individual persons (and other animals as well, for simplicity this important consideration is set to the side and ignored). Narrower conceptions of distributive justice are identifiable. Some identify the topic of distributive justice as the specification of what would be fair division of the benefits of social cooperation (Rawls 1999). Suppose individuals lived on separate islands and did not interact. Here there is no social cooperation, but questions of distributive justice in the wide sense might still arise. Maybe a person who happens to live on an island with lush resources is required by distributive justice to share resources with those less favorably situated. Responsibility is a workhorse term that plays several roles. To speak of a person’s responsibilities might be to refer to her duties or obligations. A person who is responsible for some children’s safety has a duty to keep them safe. Sometimes to describe a person as responsible, for example, for her own health or safety is not to ascribe a duty regarding her health or safety to her but rather to deny that any other person is under a duty to make good any shortfall that comes about in her health or safety. To say that someone behaves responsibly in some context is to say that she does a good job executing some responsibilities she has in that context. To hold someone responsible for their actions or omissions or for the outcomes of their actions and omissions is to treat them as responsible for these things or to regard them as it would be appropriate to regard someone who was responsible for these things. So understood, holding responsible depends for its meaning on someone’s being responsible. What is that? Thomas Scanlon observes that if one is responsible for doing an action, it is then attributable to one. He writes, “To say that a person is responsible, in this sense, for a given action is only to say that it is appropriate to take it as a basis for moral appraisal of that person” (Scanlon 1998: 248). Moral appraisal of an action done by a responsible agent is moral appraisal of that agent. What establishes this connection between agent and deed? One suggestion is that one has rational agential capacities—​cognitive, volitional, and affective, at a sufficient level, and nothing interferes with one’s opportunity to exercise the capacities on this occasion of choosing and doing.

416   Richard Arneson

2.  Sufficientarianism, Utilitarianism, and Outcome Egalitarianism Not all doctrines of social justice hold that at the fundamental level, the determination of what counts as fair distribution of benefits and burdens across persons varies with individual responsibility factors. On a plain sufficientarian view, justice requires that all be provided sufficiently good opportunities, resources, or capabilities, so that all have good-​enough life prospects or a good enough level of basic rights protection or the like (Benbaji 2005). The requirement might be that one is sustained at the good enough level throughout one’s life. On such a view, individuals do not lose entitlement to sufficiency by behaving in ways that violate responsibility standards. If one behaves in a super-​responsible, highly virtuous way, one does not thereby become entitled to more sacrifice on the part of others to maintain one at sufficiency, and if one behaves in subpar, vicious ways, one does not thereby become entitled to less. We might say, for the sufficientarian, the entitlement to sufficiency is not conditional on good behavior (Anderson 1999). Other justice doctrines are likewise unencumbered by commitments to any claim to the effect that, at the fundamental level of principle, what one is entitled to get from others by way of treatment or provision is diminished or amplified by one’s choices of conduct. Utilitarianism holds that one ought always to do whatever would maximize the aggregate of good or welfare or the like, summed across persons (Eggleston 2014). How others behave has only instrumental, not intrinsic reasons-​giving force affecting the utilitarian calculation of what one should do. One’s conduct might make one a more or less effective transformer of resources into welfare, for self or others, and if so, that would affect the contribution to aggregate good that provision of resources to one could be expected to bring about. But one’s conduct does not in itself constitute a reason to reduce or increase one’s fair share of resources in utilitarian accounting. An outcome egalitarianism would be another example of a justice doctrine that rejects the idea that one’s own conduct, for which one is rightly deemed responsible, has a bearing on what we are entitled to by way of treatment at the hands of others. Outcome egalitarian justice holds that social arrangements should be set so that each person over the course of her life gets the same or receives the same from others, with what counts as the same being measured by whatever is the right standard for this purpose. Here utilitarianism, by virtue of being a single-​principle morality, more definitively excludes responsibility considerations from playing a role in the determination of what is fair than, for example, sufficientarian or egalitarian doctrines that take their place as one principle among several in an overall doctrine. For example, consider the sufficientarian doctrine espoused by Elizabeth Anderson. Her label for the position is democratic equality. In a nutshell, the position is that all members of society should be continuously enabled to function as full members of democratic society. Any shortfall one suffers in capacity to be a full participating member of democratic society is to be

Responsibility and Distributive Justice    417 boosted to the full satisfactory level, no questions asked. But being enabled to be a full participating member of democratic society, though perhaps a demanding standard, does not require that one have ample wealth or particularly favorable life prospects or that one’s life is proceeding successfully by one’s own lights. In Anderson’s view, there is a division of labor between social and individual responsibility. Society is obligated to maintain each of us at the good-​enough democratic participation level. Beyond that level, individual responsibility holds sway. Whatever ambitions any of us might have, above or in addition to being a full participant, each of us will have to fashion a set of goals and a plan of life to seek to fulfill them. The rest of us are not on the hook for making good any shortfall in the life outcomes anyone reaches. Beyond democratic sufficiency, each of us is responsible for how her own life goes—​meaning that no one else has any moral duty to bring it about that one’s life attains gains above sufficiency resources or achieves one’s prime life aims. The responsibility each of us has to fashion her own life and accept the consequences for herself, good or bad, of pursuing it, attaches to us given a background fair framework for interaction, maintenance of which is also a moral requirement that falls on all of us as community members. What counts as a fair framework, and also sets the democratic participation capacity sufficiency line, is fixed by a more rock-​bottom moral requirement that our relationships with others be relations of equality. According to Anderson and others (Scheffler 2003, 2005; Kolodny 2014), thinking of justice in distributive terms commits us to misguided priorities. What is fundamental in this domain is something entirely different—​namely, arranging social life to avoid certain inequalities in status, power, and consideration. Unjust distribution is distribution that impedes or violates relating as equals. As the thumbnail description of Anderson’s democratic equality indicates, the relational egalitarian does not reject personal responsibility as a justice consideration; far from it. She denies that anything in the ball park of distributive justice is the slot into which responsibility considerations significantly fit. Many questions arise here. One might hold that maintenance of social equality and fair distribution are separate and independent components of justice, rather than holding with Anderson and others that the latter is subordinate (Lippert-​Rasmussen 2016). Partly agreeing with Marx’s skepticism that fair distribution is what ultimately matters, G. A. Cohen defends the socialist ideal, viewed as balancing relational and distributive egalitarian concerns (Cohen 2009). One might hold that distributive justice broadly construed is morally fundamental and that the kinds of social relationships we ought to sustain are whatever ones in our circumstances would best promote distributive justice. One might hold that relating as equals, respecting one another as rational agents, itself requires conditioning the claims we make on others on the quality of our exercise of responsible agency. In the remainder of this chapter these interesting questions are set aside. (For discussion, see Lippert-​Rasmussen 2018.) This chapter explores the idea that the benefits we are entitled to expect from others, and the burdens they are entitled to impose on us, depend in part on how we and they exercise our agency. Two interpretations of this idea are discussed, under the headings of choice and fault/​desert. The idea is subject to

418   Richard Arneson challenge on the ground that we are not the sort of creatures that can exercise responsible agency; we lack free will. Even helping ourselves to assumptions that are friendly to the idea that we are do have responsibility capacity, we find that the free will challenge gives us reason to revise, maybe substantially, our common-​sense notions of what we are responsible for.

3.  Luck Egalitarianism Our starting point is a doctrine that is known as “luck egalitarianism.” It holds that luck, good and bad fortune, falls on us, but this luck comes in different flavors, which make a difference to what turn out to be our distributive fair shares. Luck egalitarianism so understood has two separate and distinct components, which we might identify as luckism and egalitarianism. The egalitarian component states a moral presumption that fair shares are equal shares, as measured by whatever is the correct standard for assessing people’s condition. The luckism component says that the presumption in favor of equal shares can be weakened or perhaps at the limit even extinguished when people arrive at unequal shares by processes for the outcome of which they are rightly held personally responsible. One suggestion: Unchosen and uncourted luck does not dampen the moral presumption for equality. Chosen and courted luck does dampen this presumption. Luckism is not necessarily tied to egalitarianism. One could affirm luck-​oriented sufficientarianism, luckist maximin, luckist prioritarianism, and so on. On the face of it, the luck-​oriented idea that what should count as fair distributive shares is sensitive to personal responsibility considerations is compelling. Consider the sufficiency doctrine that Anderson espouses. Imagine two persons who started adult life with resources, skills, and opportunities that enable each one to be a full participant in democratic society over a full span of life. Maria carefully husbands her enabling means, but is laid low by a disease spread by an unforeseeable epidemic. Without an infusion of extra medical care, her condition will unavoidably be below sufficiency. Dick squanders his enabling means, failing to maintain his work skills, indulging in prolonged alcohol abuse, sneering at exercise, and eating unhealthy foods. He succumbs to predictable health ailments, expensive to treat. Without an infusion of extra medical care, his condition will unavoidably be below sufficiency. If it is not possible to help both Dick and Maria, the fact that the misfortune he suffers lay within his power to control, and was avoidable by reasonable actions he might have undertaken, might be thought to render it the case that Maria’s claim to aid in these circumstances is stronger and takes priority over his. However, as already mentioned, norms of personal responsibility as they might figure as components of social justice doctrines are many and various. Luckism is one path among others. Following this path, we see that it forks. Larry Temkin formulates the luck egalitarian idea in these words: “It is morally bad—​ unjust and unfair—​if some are worse off than others through no fault or choice of their

Responsibility and Distributive Justice    419 own” (Temkin 1993:13, see also Lippert-​Rasmussen 2016 and Segall 2013). Let’s add: It is morally bad—​unjust and unfair—​if some are just as well off as others through no merit or choice of their own. Let’s also interpret this as a pro tanto norm, not one that absolutely ought to hold sway come what may. Putting these claims together still leaves room for interpretation. Let us go with this: It is in itself always morally better to some degree if people’s condition is closer to equality rather than further from it, and also in itself morally better to some degree if people’s condition is closer to being in proportion to their fault, merit, or choice. The two values need to be balanced against each other, and perhaps against other considerations, to obtain an all things considered verdict as to what are fair shares. The alternation of fault or choice indicates that different norms of individual responsibility (or luckism) might be in play here. Ronald Dworkin has championed choice (Dworkin 2000, 2011; for discussion see Arneson 2018). To see its appeal, consider some examples. High-​stakes gambler.  Alf and Bert voluntarily enter a casino and engage in high-​ stakes gambling from which one emerges poor, one rich. Nose-​to-​the-​grindstone worker.  Alicia works ferociously hard at her job and puts in overtime hours steadily over the years. Bea works less hard and takes time off work for leisure time pursuits. Over time Alicia has greater wealth, Bea less. Adventurous climber.  Kay enjoys the thrill of big climbs on treacherous mountains. Mel enjoys the safety of climbing in a gym. Over time Kay suffers injuries and needs rescues; Mel does not.

Notice that as described, nobody’s choices are necessarily unreasonable or imprudent. The risks that Alf and Bert and Kay take may be good bets; the play-​it-​safe choices of Mel may be equally good decisions. Alicia and Bea choose different packages of work and leisure, but for each of them, the package chosen may be rational. The notion of personal responsibility as choice does not make fair shares vary depending on whether an individual is making good or bad choices, or better or worse choices than others are making. The relevant distinction is between brute and option luck, characterized by Dworkin as follows: “Option luck is how deliberate and calculated gambles turn out—​whether someone gains or loses through accepting an isolated risk he or she should have anticipated and might have declined. Brute luck is a matter of how risks fall out that are not in that sense deliberate gambles” (Dworkin 2000, 73). In the three examples described, the responsibility as choice norm dictates that if the individuals start adult life with fair (for the egalitarian, equal) shares, no further compensation is appropriate to equalize the conditions of Alf and Bert, or those of Alicia and Bea, or those of Kay and Mel. (If Alicia and Bea are facing choices with certain outcomes, this is a degenerate or limit case of option luck. Each takes a “gamble” with the outcome of choice known in advance for sure.) For option luck to be fully in play, the individual who takes a gamble must have some reasonable alternative to taking that gamble. Chased by an angry bear, I might choose to

420   Richard Arneson jump off a cliff, risking injury if I land on rocks below and drowning if I land in a river that exceeds my swimming ability. But if the risk of a bear attack was unforeseeable and no reasonable alternative to escape the bear is available, the outcome that befalls me is brute not option luck. A complication here is that a shadow of option luck remains: Suppose I choose to face the bear, even though the cliff jump gives a better prospect of escaping death or serious injury. The availability of a reasonable alternative course of action renders this a somewhat option luck choice.

4.  Deservingness There is an alternative, desert-​oriented standpoint from which personal responsibility might be assessed (Eyal 2007). We can imagine further detail that might obtain in the three gambling choices described. High-​stakes gambler. Alf ’s spouse needs an expensive medical treatment not covered by insurance. Alf can gain the money he needs to provide his spouse the medical treatment and a good chance of survival only if he gambles big and wins, and the casino offers the best available odds of success. Bert, in contrast, borrows against his meager retirement savings in order to indulge his yen for high-​stakes gambling. His aim is to gain a big wad of money for a splurge vacation in Phoenix. Nose-​to-​the-​grindstone worker. Alicia neglects her family responsibilities in order to pursue her career to the max. Bea has no lucrative work skills but happens to have great abilities to convert spare time into fulfilling life activities. She trades paltry extra wages, which would not do anyone much good, for free time that expectably is the key that unlocks the door to a richly flourishing life. Adventurous climber. Kay correctly observes that her community is greatly in need of competent volunteers for wilderness rescue teams. She hones her mountain skills with a view to volunteering for valuable rescue service. Mel is simply trying to maximize her expected fun over the long run of her life span. If we ask, who is more deserving then in common-​sense terms, Alf looks to be more deserving than Bert, so more eligible for compensation if he loses at gambling than Bert would be with equal losses, and more entitled to keep his winnings, if he is lucky, than Bert would be with equal winnings. In a similar way, Bea looks to be more deserving than Alicia, and Kay more deserving than Mel. Alicia makes the conventionally prudent choice, saves for the future, and Mel does the best he can to avoid unnecessary risk-​taking. To focus on the question at issue here, do not consider in the first instance how institutions should be set up to handle a possible claim for compensatory help by one of these individuals, who ends up suffering large losses. How institutions and social practices should be arranged may reflect an array of considerations. We would certainly not want first-​responders arriving at an emergency scene to inquire into the process by which the emergency victims came to be in peril at this moment. We want them to ask no questions

Responsibility and Distributive Justice    421 and just do what they can to do as much good as they can, saving as many as can be saved. Perhaps tax law enforcers would be hopelessly overburdened if we inserted fine-​grained issues of deservingness or risk responsiveness into the determination of what tax rates apply to individuals and what tax bills are owed. Perhaps attempts to make treatment of individuals by state policies and by officials administering the policies sensitive to deservingness and character would inevitably be self-​defeating, counterproductive. Ultimately the moral principles we embrace should make a difference to the design and operation of social practices. But there are reasons to abstract away from feasibility and implementability issues when pondering what from a moral standpoint matters, what moral principles we should embrace. Most obviously, what morally matters and what is feasible are just different questions. Even if slavery is entrenched so deeply in society now that no antislavery policy is implementable, the question, is slavery right and just makes sense, and the answer can be a resounding No even if somehow this has no immediate policy implications. Secondly, what is implementable shifts. If we don’t separate the issues, we may fail to notice possible paths we could now take, that would bring it about that in the future what is now unimplementable becomes implementable. Third, to ask what is feasible is usually not to ask what is literally physically possible but rather something like this: What means that we could take to achieve candidate goal to one or another degree would be worthwhile to undertake given the moral cost of pursuing them? To determine these trade-​offs of values we need to know what are the magnitudes of the values and disvalues at stake in such exercises. Eyal’s point is simple and, once stated, hard to brush aside. Suppose someone is in a predicament and needs help. Or suppose someone is in possession of resources that could be deployed to help those in predicaments, dire conditions. In some cases, something in the situation makes us hesitate, doubtful that the person in trouble really has a strong claim for help, or doubtful that the manner in which the person came to be in possession of resources that could be used for altruistic ends really leaves us morally free in one way or another to expropriate her resources for good causes. So to speak, good or bad luck can sometimes stick to a person, in the sense that there arise moral reasons not to remove the bad luck and not to take away the person’s good luck. But the factor that inhibits us, Eyal insists, is never the sheer fact that the person has undertaken risk to his prospects (or at the limit a certain loss). In his words: “That someone incurs a disadvantage without having culpably to risked incurring it is, in a central respect, unjust. If, however, that disadvantage results from that person’s own culpable choice to take that risk, then (barring prioritarian considerations) that disadvantage can remain perfectly just” (Eyal 2007: 4). To clarify, we should allow that deservingness can take a positive or negative form. Being culpable (morally blameworthy), one incurs moral discredit, and likewise, being morally praiseworthy, one incurs moral credit. The claim being considered is that moral discredit and credit is a fundamental distributive justice building block. (Eyal pairs deservingness with a prioritarian norm of distributive justice, but as mentioned already, we could pair deservingness with any of several such norms.)

422   Richard Arneson We might balk immediately at the identification of justice-​relevant deservingness with moral deservingness. Doesn’t the fastest runner in the race deserve to finish first, and the brilliantly successful scientist the Nobel Prize? In response: Any form of activity in which people seek to encourage excellent achievement will give rise to standards of merit and deservingness. If you want to run a bridge club that promotes excellent card play, you had better reward the best players. For this purpose they are the most deserving. They could be gangsters or moral predators—​that does not matter for the purpose at hand. If by some odd causal chain rewarding the worst players best promoted good play, there would be no objection to heaping awards on the players who can never manage a finesse or count the cards that have been played so far in the hand. The suggestion then is that deservingness is often an instrumental concern. People are rewarded for having and exercising traits, rewarding which serves the purposes of the enterprise. In a well-​run pirate operation, it would make sense to regard the more ferocious and loyal pirates as meriting reward. There should be no suggestion here that at some fundamental moral level, justice requires that people who are more deserving in these common-​sense ways are entitled to better lives or greater shares of resources. Our question is, is there some idea of rewarding the deserving, such that doing that is intrinsically not merely instrumentally valuable? Further response: Even if we thought it intrinsically desirable that, for example, a prize for the best player in some domain should go to the player who plays best, this thought can float entirely free from the idea that there is some sense of being deserving, such that it is in itself morally better, all else equal, that good fortune, over the course of someone’s life, goes to the deserving. We are following the suggestion that there is such a sense of being deserving, and that it consists of moral deservingness. The suggestion Eyal and I are following is actually a bit narrower; it identifies the relevant moral deservingness with moral praiseworthiness and moral blameworthiness. We’ll need eventually to say more about what this might come to. Another point to flag here is that someone might abide the idea of people’s being morally deserving or undeserving in particular contexts and settings but reject the idea that there is any moral imperative or norm to the effect that it is morally fitting that people’s good or bad fortune be proportionate to their overall lifetime deservingness score. This chapter focuses on whether the rejected idea just mentioned can be justified. One reason for this focus is that if we confine the idea that fair shares should track desert to particular contexts it will be hard to be sure we have clearly distinguished our instrumental convictions –​to the effect that, for example, if we define “good teacher” appropriately, rewarding good teachers will bring it about that children learn more and better –​from our convictions, if any, regarding what is in itself morally fitting. However, this is not to deny that distributive justice considerations can arise in particular contexts, where significant burdens and benefits will fall on people, depending on what we do, or what rules we enforce regarding this context, and the question arises, what distribution would be in itself fair. One such context is war (McMahan 2011).

Responsibility and Distributive Justice    423

5.  Desert and Free Will For many of us, once we start thinking of global deservingness as rendering it desirable to bring about overall good or bad fortune for a person, we face free will skepticism. Maybe when we get down to it, no one is really truly deserving in the way that could underwrite this connection (Pereboom 2001, 2014; also McKenna and Pereboom 2016). Not everyone is tempted down this path. Shelly Kagan, author of the definitive treatment on desert for our time, writes, “I find myself far more confident of the claim that some people are more morally deserving than others than I am of any particular philosophical theses concerning the supposed metaphysical preconditions of desert.” Kagan adds that whatever these metaphysical conditions of desert might be, “I simply assume—​if only for the sake of discussion—​that whatever they are, they are met” (Kagan 2012: 130). This stance understates the rational threat posed by the skeptical challenge (Nagel 1979). For consider a familiar argument: 1. If human choices are events brought about by prior causes, they are beyond our power to control. 2. We are morally responsible, at most, for what is within our power to control. 3. Human choices are events brought about by prior causes. 4. So, we are morally responsible for none of our choices. Every phrase in this argument can be variously interpreted. Interpretations of this and similar arguments are legion in recent discussions of free will and moral responsibility. For purposes of this chapter, we simply assume that some suitably refined version of 1–​4 constitutes a valid argument. The compatibilist denies 1, asserting that the fact that human choices are caused events does not rule out their being within the power of the choosing agent to control, in some sense that preserves the agent’s moral responsibility for choice. According to the free will libertarian, premise 3 is false. The intuitive basis for rejecting 3 is that our ordinary experience of choosing between alternatives does not involve any experience of events prior to choice that predetermine the choice, so that it is not really up to us to choose any one of the alternatives presented. This may not be any strong reason to reject 3, because our ordinary experience of choosing might simply be shaped by lack of conscious awareness of the panoply of causal factors that shape the choice we are now making (see Nagel 1987). Nonetheless, the “free will” hypothesis is best viewed as the empirical claim that the full scientific account of human choice making will somehow vindicate the ordinary chooser’s experience. How might that be? Our current science offers no clue. If the free will hypothesis is correct, future science will have to differ sharply from current brain science in ways we cannot anticipate. Trying to give a coherent account of the free will idea would seem to

424   Richard Arneson be a mug’s game, akin to trying to characterize non-​Euclidean geometry before its core concepts have been invented and formulated. The abstruse free will issue has immediate ramifications for any distributive justice doctrine according to which fair shares for individuals depend in part on what those individuals deserve. We can be differentially deserving only in ways that lie within our power to control. If no significant aspect of our agency lies within our power to control, we are morally responsible for none of it, and so we cannot be differentially deserving. Everyone’s deservingness score will be the same. This flattening of individual deservingness would not show that a distributive principle according to which fair shares for individuals should vary with their deservingness must be false (but see Scheffler 2005 for a contrary view). It might be true, but it would become an idle wheel for the determination of fair shares. (The principle might be construed as making a hypothetical claim: If people vary in deservingness, their fair shares will vary with their varying deservingness scores.) If compatibilism is true—​that is, if premise 1 in the argument above is false—​then we might be responsible for some of our choices and might become differentially deserving even in a world in which human choices are all caused events. If compatibilism prompts doubts (for a clear simple expression of doubt, see Nagel 2014; for discussion of the issue McKenna and Pereboom 2016 and the references cited therein; and for a recent attempt to dissolve the doubt, List 2019), we have extra reason to investigate whether background conditions might yet obtain, such that it can make sense to embrace distributive justice principles that incorporate deservingness as intrinsically helping to fix the content of fair shares. So suppose we begin by denying premise 3 (“Human choices are events brought about by prior causes”). Where does that leave us?

6.  Quality of Will Remarks by John Rawls pose further problems for the idea that distribution according to desert can plausibly become a component in distributive justice principles. Common-​sense precepts of desert on cursory examination look unsuited for this role. He adds: The precept which seems intuitively to come closest to rewarding moral desert if that of distribution according to effort, or perhaps better, conscientious effort. Once again, however, it seems clear that the effort a person is willing to make is influenced by his natural abilities and skills and the alternatives open to him. The better endowed are more likely, other things equal, to strive conscientiously, and there seems to be no way to discount for their greater good fortune. The idea of rewarding desert is impracticable. (Rawls 1999: 274).

Responsibility and Distributive Justice    425 Rawls does not assert that individuals never make choices for which they might be, to some degree, morally responsible. Suppose they do. This would not plausibly imply that choices float entirely independently above the world of caused events. Whatever our free choice might be, it is conditioned by causal factors and entangled with them. Rawls does not deny there could be such a thing as desert, but determining that one person is more or less deserving, given causal entanglement, is beyond our power. This would be true for an acting individual as well as for others observing her. What choices one makes is partly a result of the decision problems one faces, and these may often be beyond her power to control. One person is tempted and fails to resist the temptation, and other is not tempted; one person has the opportunity to act nobly; another lacks the opportunity. In another terminology, we might say that a person’s deservingness cannot depend on luck that is beyond her power to control. This suggestion might push us to hold that what one is truly responsible for, at most, is the orientation of one’s will—​one’s disposition to do what is right. How circumstances conspire to trigger the dispositions one has is beyond one’s power to control, so premise 2 above limits responsibility to the quality of one’s will, and rules out assessing true deservingness by the quality of choices displaying one’s underlying will. This cannot be right either. Disposing one’s will to do what is right is ambiguous. This can refer either to disposing oneself to do whatever is in fact right, so far as one can determine what that is, or it can refer to disposing oneself to do this and that particular thing or type of thing, which one takes to be in fact right. But even if anyone can make a good faith effort to try to discern what is morally right, people have different levels of ability and face variously favorable opportunities for succeeding in such efforts. Moral truths may be difficult to discern. Figuring out what is right may require solving a complex equation, which you can understand and solve, and I cannot. Whether becoming morally deserving depends on trying to do what one subjectively takes to be moral, or what really is moral, is controversial. What can be called a pure quality of will view holds that what matters is degree of concern that is directed to what is really morally right from the recognition that this is so. Here is Nomy Arpaly stating the view: “An action is blameworthy just in case the action resulted from the agent’s caring inadequately about what is morally significant—​where this is not a matter of de dicto caring about morality but de re caring about what is morally significant” (Arpaly 2003: 84). Suppose that in a certain setting what is really morally significant is the fact that one’s act would help the needy. Suppose that in this setting what one sincerely but falsely believes to be morally significant is to refrain from coddling people’s weakness by helping them avoid facing difficult problems and tasks. Suppose one turns away from a drowning person struggling in deep water, with a view to avoiding coddling. One acts from concern for what one de dicto takes to be morally required but does not act from any concern at all for what is de re really morally required. For Arpaly, one’s act is morally blameworthy (for a view on this position opposed to Arpaly’s but nonetheless still rejecting the control principle, see Rosen 2003 and 2004; and for critical discussion from a pro-​Arpaly stance, Harman 2011).

426   Richard Arneson This is not an implausible position, but it flatly goes against the control principle (premise 3). If it does not lie within one’s power of control to discern what is really morally significant, one cannot be morally blameworthy for failing to discern this and act from concern for it. So holds the control principle. One might reject the control principle on the ground that it generates counterintuitive judgments we ordinarily confidently make, such as that Nazis are morally blameworthy for evildoing even if that evildoing is done from sincere belief that one is doing the right thing, a belief formed after conscientious and careful moral reflection. But the control principle is itself rooted in common-​sense conviction. “I couldn’t help doing it” when truthfully and correctly uttered blocks adverse moral assessment to the effect that the speaker is morally blameworthy for doing it. My phobia against touching snakes may render it the case that I couldn’t help running away from a snake encounter when the morally right act in the circumstances would have required prolonged snake handling. Lacking knowledge that the dark shadow in my car’s path is a paralyzed person and lacking any indication that further checking would make any sense, I couldn’t help running over the person even though, given the actual facts, that was the morally wrong thing to do. Arpaly’s quality of will account distinguishes the Nazi example involving acting on false moral belief and the other examples involving compulsive desire and false empirical belief. The conscientious Nazi does evince inadequate moral concern de re but the snake avoider and the unfortunate driver do not. However, this result is secured by moral gerrymandering: The principle is formulated, one might say rigged, to get this result. Suppose that if I had an IQ of 110 I could arrive at a correct factual belief on a matter germane to choice of right action, but lacking the extra intelligence, I could not arrive at a correct factual belief. Here I am not morally blameworthy if I act wrongly from false belief. But in another scenario, if I could arrive at the correct moral belief if I had slightly higher IQ, but could not with my actual 108 level, and I act from false moral belief, I am morally blameworthy according to Arpaly. I lack concern for what is truly morally significant. But you could just as well say, in the factual belief case, that I act from inadequate concern for what is in fact morally right in my circumstances. I’m not concerned at all to do what is in fact right, because I have no way of knowing that particular act is the one that in my circumstances is uniquely morally right.

7.  Conclusion The position so far advanced is that what renders one morally deserving, in the way that boosts one’s fair share, is caring de dicto for what is morally significant—​the right and the good. To care is to be disposed to choose and strive in accordance with morality’s dictates. One is morally responsible for caring in this way, or not, to the extent this lies within one’s power to control. A companion principle complements this doctrine: The

Responsibility and Distributive Justice    427 more difficult and painful it would be to do what I take to be morally right, and the more cost I see that I will incur if I do it, other things being equal, and adjusting for the moral stakes involved, the less blameworthy I am if I fail to do it, and the more moral credit I earn if I do it. (Making a great conscientious effort to do what one takes to be right, at great cost and sacrifice, may render one morally praiseworthy and deserving independently of whether one actually succeeds in choosing and executing what one takes to be right.) If we accept the control principle, a sensible continuity norm will recommend accepting the companion principle as well. If it lies within my power to control whether I do the right thing, but just barely, and only at great sacrifice, the assessment of my failing to do right if that is what happens should not differ greatly from the assessment that would have attached to me had this failure been just barely beyond my power to control. The position adumbrated so far does not yet fully respond to Rawls’s causal entanglement worries. Even on the assumption we have free will, it still must be entangled in a host of causes that continuously partially act as forces on choice and action, and sometimes, perhaps often, perhaps almost always render this free will reduced, impotent, or nearly so. I assume these causes are excusing in negative cases and in positive scenarios, deservingness-​diminishing. So how can one tell whether anyone, oneself or another, is truly deserving, and to what degree? By sheer luck, for example, one person will be strongly motivated to put forth conscientious effort toward figuring out what is right and conforming one’s will to that, and another person will be beset with strong contrary desires (accepting moral luck, one would find acting from bad motives that just befall one, and also find being led by self-​interest not to seek out inconvenient moral truths or evidence, as blameworthy, as in Harman 2011). One person by luck will be faced with decision problems that happen to fit her dispositions so that she acts well; another not. And two persons identically disposed may face different decision problems. Even if each is somehow identically responsible for the orientation of his or her will, one will get the opportunity to act well, or badly, the other not. As Rawls states, the idea of rewarding desert is impracticable. The notional reply to entanglement worries is that a person’s raw or common-​sense deservingness score must always be adjusted, up or down, in order fully to offset causal entanglement. One subtracts the influence on choice and effort of all causal forces operating on the individual to push choice or effort one way or another. We can’t do this, of course, in any remotely fine-​grained way. Can we can do so in a coarse-​grained way, by institutional procedures or introspective judgment or inference from behavioral observation, in some settings, sufficiently adequately so that the project of rewarding desert is practicable and not counterproductive? This is a large question, or maybe a nest of questions. (From a first-​person perspective, I am myself in no doubt that I have been seriously morally blameworthy sometimes in ways that render me less deserving and entitled to less by way of fair shares than I would otherwise have been. But such confident judgment is not self-​certifying.) These questions this essay leaves open and unaddressed. Of course, if the questions are open, that means Marx and Rawls, who

428   Richard Arneson in different ways want to downgrade desert as a factor that should influence people’s shares, might yet be shown to be right.

References Anderson, E. (1999). “What Is the Point of Equality?” Ethics 109: 287–​337. Arneson, R. (2000). “Luck Egalitarianism and Prioritarianism.” Ethics 111: 339–​349. Arneson R. (2018). “Dworkin and Luck Egalitarianism: A Comparison,” in S. Olsaretti, ed., The Oxford Handbook of Distributive Justice. Oxford: Oxford University Press, 41–​64. Arpaly, N. (2003). Unprincipled Virtue. Oxford: Oxford University Press. Benbaji, Y. (2005). “The Doctrine of Sufficiency: A Defence.” Utilitas 17: 310–​332. Cohen, G. A. (1995). “Self-​Ownership, Communism, and Equality: Against the Marxist Technological Fix,” in his Self-​Ownership, Freedom and Equality. Cambridge: Cambridge University Press. Cohen, G. A. (2008). Rescuing Justice and Equality. Cambridge, MA: Harvard University Press. Cohen, G. A. (2009). Why Not Socialism? Princeton, NJ: Princeton University Press. Dworkin, R. (2000). Sovereign Virtue: The Theory and Practice of Equality. Cambridge, MA: Harvard University Press. Dworkin, R. (2011). Justice for Hedgehogs. Cambridge, MA: Harvard University Press. Eggleston, B. (2014). “Act Utilitarianism,” in B. Eggleston and D. Miller, eds., The Cambridge Companion to Utilitarianism. Cambridge: Cambridge University Press, 125–​145. Eyal, N. (2007). “Egalitarian Justice and Innocent Choice.” Journal of Ethics and Social Philosophy 2: 1–​18. Harman, E. (2011). “Does Moral Ignorance Exculpate?” Ratio 24: 443–​468. Kagan, S. (2012). The Geometry of Desert. Oxford and New York: Oxford University Press. Kolodny, N. (2014). “Rule Over None II: Social Equality and the Justification of Democracy.” Philosophy and Public Affairs 42: 297–​336. Lippert-​Rasmussen, K. (2016). Luck Egalitarianism. London and New York: Bloomsbury Academic. Lippert-​ Rasmussen, K. (2018). Relational Egalitarianism: Living as Equals. Cambridge: Cambridge University Press. List, C. (2019). Why Free Will Is Real. Cambridge, MA: Harvard University Press. Marx, Karl (1977). “Critique of the Gotha Program,” in D. McClellan, Karl Marx: Selected Writings. Oxford: Oxford University Press, 1977, 210–​217. McKenna, M., and Pereboom, D. (2016). Free Will: A Contemporary Introduction. New York and London: Routledge. McMahan, Jeff. (2010). “The Just Distribution of Harm Between Combatants and Non-​ Combatants,” Philosophy and Public Affairs 38 (4):342–​379. Nagel, T. (1979). “Moral Luck,” in his collection Mortal Questions. Cambridge: Cambridge University Press, 24–​38. Nagel, T. (1987). “Free Will,” in Nagel, What Does It All Mean? A Very Short Introduction to Philosophy. New York and Oxford: Oxford University Press, 47–​58. Pereboom, D. (2001). Living Without Free Will. Cambridge: Cambridge University Press. Pereboom, D. (2014). Free Will, Agency, and Meaning in Life. New York and Oxford: Oxford University Press. Rawls, J. (1999). A Theory of Justice, rev. ed. Cambridge, MA: Harvard University Press.

Responsibility and Distributive Justice    429 Rosen, G. (2003). “Culpability and Ignorance.” Proceedings of the Aristotelian Society 103(1): 61–​84. Rosen, G. (2004). “Skepticism About Moral Responsibility.” Philosophical Perspectives 18: 295–​313. Scanlon, T. (1998). What We Owe to Each Other. Cambridge, MA: Harvard University Press. Scheffler, S. (2003). “What Is Egalitarianism?”. Philosophy and Public Affairs 31: 5–​39. Scheffler, S. (2005). “Choice, Circumstance, and the Value of Equality.” Politics, Philosophy and Economics 4: 5–​28. Segall, S. (2013). Equality and Opportunity, Oxford: Oxford University Press. Temkin, L. (1993). Inequality. Oxford and New York: Oxford University Press.

PA RT I X

R E SP ON SI B I L I T Y, N E U RO S C I E N C E , A N D P SYC HOL O G Y

CHAPTER 21

MORAL RESP ONSI BI L I T Y AND NEUROS C I E NC E Alfred R. Mele

Skepticism about the existence of moral responsibility has a variety of sources. This chapter’s topic is the most prominent neuroscientific source of this skepticism, namely, studies that purport to show that various decisions are made unconsciously (that is, are made at times at which we are not conscious of making them). I have argued in detail elsewhere that, in fact, the data do not warrant the judgment that participants in these studies made decisions unconsciously.1 Here I explain why, even if they did unconsciously make decisions, that poses no significant challenge to the proposition that moral responsibility exists and is widespread.

1.  Some Prominent Studies Assume that participants in the studies I will be discussing actually made the decisions the experimenters attributed to them. What did they decide to do? In highly influential experiments by Benjamin Libet, they decided when to flex a wrist (1985). In a much-​ discussed fMRI study (Soon et al. 2008), they decided which of two buttons to press. In a more recent study (Fried et al. 2011), they decided when to press a key.2 Participants had no reason to prefer the decided-​upon option to similar alternative options—​and vice versa. They had no reason to prefer a particular moment for beginning to flex a wrist or press a key over nearby moments and (in the study by Soon et al.) no reason to prefer either button over the other. 1  On Benjamin Libet’s studies, see Mele 2009; on work reported in Soon et al. 2008, see Mele 2014, pp. 200–​205; and on work reported in Fried et al. 2011, see Mele 2018. 2  Fried et al. mention another study of theirs in which participants select which hand to use for the key press (2011, p. 553).

434   Alfred R. Mele Participants are asked to report on the time at which they had certain conscious experiences—​variously described as experiences of an urge, intention, or decision to do what they did. They make their reports after they act. When participants in Libet’s studies are regularly reminded not to plan their wrist flexes, and when they do not afterward say that they did some such planning, an average ramping up of EEG activity (beginning 550 ms before muscle motion begins) precedes the average reported time of the onset of the conscious experience (200 ms before muscle motion begins) by about a third of a second (1985). Libet claims that decisions about when to flex were made at the earlier time, 550 ms before muscle motion (1985, p. 536). Although I disputed that claim in Mele 2009, here, for the sake of argument, I am supposing that it is true. Chun Siong Soon, Marcel Brass, Hans-​Jochen Heinze, and John-​Dylan Haynes, commenting on Libet’s studies, write, “Because brain activity in the SMA consistently preceded the conscious decision, it has been argued that the brain had already unconsciously made a decision to move even before the subject became aware of it” (2008, p. 543).3 To gather additional evidence about this, they use functional magnetic resonance imaging (fMRI) in a study of participants instructed to do the following “when they felt the urge to do so”: “decide between one of two buttons, operated by the left and right index fingers, and press it immediately” (p. 543). Soon et al. found that, using readings from two brain regions (one in the frontopolar cortex and the other in the parietal cortex), they were able to predict with about 60% accuracy (see Soon et al. 2008, supplementary ­figure 6; Haynes 2011, p. 93) which button participants would press several seconds in advance of the button press (p. 544). In another study, Soon et al. instruct participants to “decide between left and right responses at an externally determined point in time” (2008, p. 544). The participants are to make their decision when shown a cue and then execute their decision later, when presented with a “respond” cue (see their supplementary material on “Control fMRI experiment”). They report that one interpretation of their findings in this study is that “frontopolar cortex was the first cortical stage at which the actual decision was made, whereas precuneus was involved in storage of the decision until it reached awareness” (p. 545). Itzhak Fried, Roy Mukamel, and Gabriel Kreiman use depth electrodes to record directly from the brain (2011). They report that “a population of SMA [supplementary motor area] neurons is sufficient to predict in single trials the impending decision to move with accuracy greater than 80% already 700 ms prior to subjects’ awareness” (p. 548) of their “urge” (p. 558) to press a key. By “subjects’ awareness” here, Fried et al. mean the “awareness” time that participants later report: The authors recognize that the reports might not be accurate (pp. 552–​553, 560). Unlike Libet, they seem sometimes 3  Readers should not take this quotation to attribute two decisions to the participants, one made unconsciously and the other made consciously. Rather, the idea is that the participants make their decisions unconsciously and become conscious of those decisions later: “the conscious decision” is meant to refer to a decision that the participant has become conscious of.

MORAL RESPONSIBILITY AND NEUROSCIENCE     435 to treat decisions as conscious by definition (p. 548). Possibly, in their thinking about their findings, they identify the participants’ decisions with conscious urges. If that is how they use “decision,” their claim is that on the basis of activity in the SMA they can predict with greater than 80% accuracy what time a participant will report to be the time at which he was first aware of an urge to press 700 ms prior to the reported time. But someone who uses the word “decision” differently may describe the same result as a greater than 80% accuracy rate in detecting decisions 700 ms before the person becomes aware of a decision he already made.

2.  Some Terminological Background Some terminological background will prove useful. I start with decisions. The decisions at issue in this chapter are decisions to do things. They should be distinguished from decisions about what is the case, as in “After waiting for an hour, Alice decided that Bob probably was not going to show up.” In this chapter, when the word “decisions” is used without qualification, it is shorthand for “decisions to do things.” To decide to do something, as I understand it, is to perform a momentary action of forming an intention to do it (Mele 2003, ch. 9). In principle, momentary actions of this kind may or may not be based on conscious reasoning about what to, and decisions about what to do may be made in the absence of conscious reasoning about what to do. If and when a decision is based on conscious reasoning, the reasoning is not part of the decision; it is rather a cause of the decision. Alice, in a Libet experiment, decided at noon to flex her wrist right then and Bob decided at noon, after a great deal of reflection, to quit his job tomorrow. As I see things, what this amounts to is that, at noon, Alice performed a momentary action of forming an intention to flex her wrist right then and Bob performed a momentary action of forming an intention to quit his job tomorrow. I leave it open that such momentary actions can be unconscious. My conception of deciding has at least two virtues for present purposes. It is straightforward, and it does not beg any questions against those who view the data I have described as posing a serious threat to the existence of moral responsibility.4 Proximal decisions are, in the basic case, decisions to do a thing right then (Mele 1992, pp. 158–​160). Alice’s decision, made at noon, to flex her wrist right then is a proximal decision. (Bob’s decision is a distal decision; see Mele 1992, pp. 158–​160.) Having the expression “proximal decision” ready to hand will save some ink.

4  Jeff Miller and Wolf Schwarz report that although I reserve “the term ‘decision’ for the final resolution at the end of ” a process, they “prefer to use ‘decision’ as short-​hand for the entire process rather than reserving it for the final termination” (2014, p. 18). (They refer to this process as “the decision-​ making process.”) I realize that what Miller and Schwarz represent as shorthand may be regarded by some readers as ordinary usage. Also, I certainly do not deny (as Miller and Schwarz observe, p. 18) that decisions are associated with processes that issue in them.

436   Alfred R. Mele What does it mean to say that a person made a decision unconsciously? The measure of consciousness in the studies at issue is a report the participant makes. For example, after she flexes and the Libet clock comes to a halt, Alice may move a cursor to point p on the clock to indicate that p is when she was first conscious of a proximal decision to flex (or to indicate that p is when she made this proximal decision). Experimenters who believe that Alice actually made the decision at p—​350 ms and take Alice’s report seriously will conclude that Alice made her decision unconsciously and became conscious of it about a third of a second after she made it. (For grounds for caution about the accuracy of participants’ “consciousness” reports, see Mele 2009, ch. 6 and 2013b, pp. 77–​82.) For a decision to be made unconsciously is for it to be made at a time at which the person is not conscious of making it. The preceding sentence is an important one to keep in mind, and I will be returning to it. My final terminological observation is about moral responsibility. I take seriously both words in the expression “morally responsible.” As I understand moral responsibility, it is a moral matter. So, in my view, people are not morally responsible for actions of theirs that fall outside the sphere of morality—​or, more precisely, actions of theirs that morality is not in the business of prohibiting, requiring, or encouraging. (I owe this way of putting things to Josh Gert.)

3.  An Unsuccessful Argument for Skepticism How might one build a bridge from the assertion that in the studies described in section 1 decisions are made unconsciously to the conclusion that (probably) no one is morally responsible for anything? Here is one way to do it. Argument 1 1. In the experiments at issue, all the decisions studied are made unconsciously. 2. So probably all decisions are made unconsciously. 3. Any person who makes a decision unconsciously is not morally responsible for that decision. 4. So probably people are never morally responsible for their decisions. (From 2 and 3.) 5. People who are never morally responsible for their decisions are not morally responsible for anything.5 6. So probably no one is morally responsible for anything. (From 4 and 5.) 5  A proponent of this premise may believe that all possible instances of moral responsibility either are or derive from moral responsibility for decisions. For my purposes in this chapter, there is no need to take a stand on this premise.

MORAL RESPONSIBILITY AND NEUROSCIENCE     437 Although I have argued elsewhere that premise 1 is unwarranted (see note 1), I assume here that it is true for the sake of argument. My focus in the present section is on the inference expressed in premise 2. Is that inference warranted? That depends on how representative the decisions (allegedly) made in these studies are of decisions in general. (Reminder: Only decisions to do things are at issue.) Regarding the issue of representativeness, one point to notice is that the decisions made in these studies are not moral decisions. That is, these decisions are not located in the moral sphere of life. In this way, they differ from decisions that do have a home in that sphere—​decisions about what to do in the sphere of conduct that morality is in the business of prohibiting, requiring, or encouraging. A related point in linked to my observation that participants in the studies at issue had no reason to prefer the decided-​upon options to other relevant options and vice versa. This normally is not the case when people make decisions in the moral sphere. The decisions allegedly made in these experiments are morally neutral decisions made under conditions of indifference. How similar are they to morally significant decisions made in situations in which agents are weighing competing reasons? And how similar are they to nonmoral decisions made in situations in which people are not indifferent about their leading options and weigh competing reasons? In the experiments at issue, participants are instructed to be spontaneous. They are instructed to avoid thinking about when to flex or click or about which button to press. If they decide without consciously reasoning about what to do, they are being good, obedient experimental subjects. But many decisions are preceded by prolonged conscious reasoning. This is another difference between the decisions under investigation in the studies at issue and many other decisions. Here I pause to mention the possibility that when agents are consciously weighing reasons and apparently deciding (partly) on the basis of their conscious assessment of reasons, the decisions they make may be less likely to be unconscious ones. The conscious processing may increase the likelihood of conscious deciding.6 Philosophers who believe that we sometimes act freely or, more formally, exercise free will, disagree about whether we can do this in situations that feature indifference about the leading options. As some philosophers—​so-​called restrictivists—​conceive of free will (Campbell 1957, pp. 167–​174; Kane 1989, p. 252; van Inwagen 1989), people can exercise it only when they make significant moral or practical decisions in the face of temptation or competing motivation; and some other philosophers are much less restrictive about free will (Berofsky 2012, pp. 65–​66 Fischer and Ravizza 1992; O’Connor 2000, pp. 101–​107). If the restrictivists are right, alleged findings of the sort described in section 1 tell us little about free will: The decisions and overt actions being investigated are outside the sphere of free will in addition to being outside the sphere of morality. But even if there can be free decisions in cases of indifference, the disagreement between restrictivists and others may point to an interesting difference among kinds of allegedly

6 

For a model of conscious deciding, see Mele 2009, pp. 40-​44.

438   Alfred R. Mele free decisions. Perhaps people who consciously struggle with temptation or competing motivation during a process that leads to a decision are more likely to decide consciously than are people who unreflectively select among options with respect to which they are indifferent. So how representative are the decisions allegedly made in studies of the sort described in section 1 of decisions in general? The decisions made in the studies are morally neutral ones that are made under conditions of indifference and (apparently) are not preceded by conscious reasoning about what to do. But many decisions are located within the moral sphere and are not made under conditions of indifference, and are preceded by conscious reasoning about what to do. These points constitute a serious obstacle to attempts to generalize from the alleged finding that the former decisions are made unconsciously to the conclusion that all decisions are made unconsciously. As I have observed elsewhere (Mele 2009, p. 83), automatic tie-​breaking mechanisms are at work in many ordinary cases in which we are indifferent between or among the available options, and it is rash to assume that what happens in situations featuring indifference is also what happens in situations in which unsettledness about what to do leads to careful, protracted, conscious reasoning about what to do. Even if some action-​ties are broken for us unconsciously, well before we are aware of what we “decided” to do, it does not follow from this that we never consciously make decisions. As I paused to reflect on what point to make next, I felt tempted to launch into a certain anecdote. I reflected—​consciously, I report—​on whether the anecdote would be appropriate and, within a minute or so, I decided to tell it. So here it is. I was once a subject in a Libet-​style experiment. My plan was to do my best to act as a naive subject would. After I got wired up and settled in, I waited for a conscious urge to flex to pop up in me so I would have something to report when it was time to make my first consciousness report. I waited until I was pretty sure no suitable urge was going to pop up on its own. I wondered what to do, gave the matter some thought, and then hit on a strategy. It was to say “now” to myself silently (many times over the course of the experiment), to flex my wrist in response to the silently uttered cue, and then, a little later, try to report where the hand was on the Libet clock when I said “now.” That was the strategy that, after a bit of conscious reflection, I decided to employ. If I decided to say “now” each time I said it (or decided to flex each time I flexed), those decisions were arbitrary pickings of a moment at which to say now (or arbitrary pickings of a moment to flex). I had no reason to prefer that moment to nearby moments and no reason to prefer any of the latter to the former. But my decision to adopt the strategy I described was not an arbitrary selection from an array of options that I was indifferent about. I selected the first option that occurred to me, after deeming it promising. My strategic decision was based on reflection—​pretty quick thinking, I am pleased to report—​on how to get out of my bind and play my role as experimental

MORAL RESPONSIBILITY AND NEUROSCIENCE     439 subject. The proximal “now”-​saying decisions (if that is what they were) differ in this way from my decision about a strategy to employ. And this difference should raise a large red flag for anyone who is tempted to generalize from the claim that the former decisions are made unconsciously to the much grander claim that all of my decisions are made unconsciously. Argument 1 does not mention free will, but a related argument does. Replacing premise 3 with the following premise yields such an argument: Any person who makes a decision unconsciously does not make it freely and therefore is not morally responsible for that decision. Robert Deutschländer, Michael Pauen, and John-​Dylan Haynes conducted a series of studies designed to test the contention that “the simple experiments used in neuroscience do not match typical concepts of free will held by laypeople” (2017, pp. 232–​233). They find evidence that, among other things, lay folk are happy to count various spontaneous, arbitrary decisions as freely made. I am not surprised. And I should emphasize that my argument here in no way depends on decisions of this sort not being free; nor have I ever endorsed restrictivism. I have argued in this section that decisions of this sort differ from decisions of some other kinds in such a way that we cannot properly generalize from the alleged finding that decisions of the former kind are made unconsciously in various experiments to the claim that all decisions are made unconsciously. Elsewhere, I have made the same point when free will, rather than moral responsibility, was the primary topic of discussion (Mele 2013a, pp. 781–​782; also see Mele 2009, pp. 85–​87). Before I bring this section to an end, two more items of business need attention. The first is a disclaimer. I mentioned the possibility that we are more likely to decide consciously when we are consciously reasoning about what to do than when we are performing our assigned tasks as obedient participants in studies of the kind described in section 1. But I am not suggesting that engagement in conscious reasoning about what to do might make us better judges about when, exactly, we were first conscious of deciding to do what we did. In cases in which such reasoning seems to issue in a decision to A, our attention is primarily focused on the issue we are reasoning about, on weighing pros and cons, and the like—​not on the question when we were first conscious of a decision to A. In Libet-​style studies, participants are focusing on something else entirely. Because they are instructed to be prepared to report on the moment of first awareness of their decision (or urge or intention) after they act, their attention is focused on detecting (or generating) a conscious proximal decision (or urge or intention) to flex a wrist or press a button, and on the clock, so they can report on the time of their experience later.7 (Their attention is also focused to some extent on the task of flexing or pressing.) Of course, experimenters can conduct studies in which participants who watch a Libet clock are instructed to deliberate about whether to A or B, make a decision, execute it, and then report on when they first became aware of their decision. The attention they devote to 7  On generating (as opposed to merely detecting) a conscious intention to flex in these studies, see Mele 2009, pp. 34-​36.

440   Alfred R. Mele the deliberative process will decrease the attention they can devote to the timing task. There is strong evidence that participants in Libet-​style studies, after they act, are still in the process of estimating where the hand was on the clock when they first became conscious of their urge (or intention or decision) to flex or click (see Mele 2013b, pp. 77–​ 782, for discussion and references); and the further division of attention in the imagined deliberative-​decision study may make the estimating task significantly more difficult. However, a person might make a decision consciously—​that is, be conscious of making it when he makes it—​even if, when asked, his estimate of the time at which he made his decision is mistaken. Finally, a potential source of serious confusion should be eliminated. The sentence “Alice made her decision unconsciously” can be read in at least two different ways: (1) Alice made her decision at a time at which she was not conscious of making it; (2) Alice made her decision independently of any conscious thinking about what to do. As I have indicated (in a sentence in section 2 that I said I would return to), a reading of type 1 is the operative one here. Libet, for example, claimed that the participants in his experiments made their proximal decisions to flex unconsciously on grounds that he took to support the claim that they were not conscious of making the decisions at the time at which they actually made them. If a reading of type 2 were the operative one, things would be even worse for proponents of Argument 1. Building a sturdy bridge from the finding that, when people are instructed not to think about what to do and apparently comply with that instruction, they make trivial decisions in the sphere of indifference independently of any conscious thinking about what to do, to the conclusion that people make all decisions independently of any conscious thinking about what to do to would be an astounding feat. Even so, some people slide from readings of type 1 to readings of type 2. The following point may help thwart any temptation a reader may have to make this slide. Suppose that Alice thinks long and hard about what to do and then makes a decision. If she becomes conscious of the decision only after she makes it, she counts as unconsciously making the decision on reading 1. And this is so even if her conscious reasoning played an important role in generating her decision—​that is, even if her decision does not count as unconsciously made on reading 2.

4.  Another Unsuccessful Argument for Skepticism In the present section I grant, for the sake of argument, that all decisions are made unconsciously (on a reading of type 1), and I explain why, even then, an argument in the spirit of Argument 1 is unsuccessful. Consider the following argument.

MORAL RESPONSIBILITY AND NEUROSCIENCE     441 Argument 2 a. All decisions are made unconsciously. b. Any person who makes a decision unconsciously is not morally responsible for that decision. c. So people are never morally responsible for their decisions. (From a and b.) d. People who are never morally responsible for their decisions are not morally responsible for anything. e. So no one is morally responsible for anything. (From c and d.) I accept premise a for the sake of argument, as I indicated. I take no position in this article on premise d (see note 5). Premise b is the primary topic of this section. Sometimes people consciously reason about what to do before they decide what to do. Suppose it is discovered that, in all such cases, the decisions they make are made about 50 ms before they become aware of them. Suppose that their conscious reasoning issues in decisions that it takes just a bit of time to register consciously. The combination of these suppositions and premise b yields the conclusion that in none of these cases is the agent morally responsible for the decision made. What does that say about premise b? Given the suppositions in play, conscious reasoning about what to do can play whatever role believers in moral responsibility believe it plays in much morally responsible decision-​ making—​unless they believe that this role is played only if the onset of consciousness of a decision does not lag behind the making of the decision. But what recommends the latter belief? Attention to a remark by David Rosenthal will prove instructive in this connection. Rosenthal writes, “there is experimental evidence that we come to be conscious of our decisions only after those decisions have been formed . . .; so consciousness cannot play a role in determining what we decide even when our decisions are conscious” (2009, p. 246). The inference here is invalid: Consciousness of a decision we make does not play “a role in determining” our making it, even if we are conscious of the decision right when we make it. We should look for causes of decisions in things that precede them; and among the candidates to be considered for causal contributors in some cases are such things as conscious reasoning, conscious beliefs, and conscious desires (or the neural realizers of these things). These things continue to be candidates even on the supposition that we do not become conscious of the decisions produced until 50 ms after we make them. Recall the two readings mentioned earlier of “Alice made her decision unconsciously”: (1) Alice made her decision at a time at which she was not conscious of making it; (2) Alice made her decision independently of any conscious thinking about what to do. In effect, Rosenthal has claimed that making a decision unconsciously on reading 1 is sufficient for making that decision unconsciously on reading 2. Someone may contend that, despite his invalid inference, Rosenthal has put his finger on the real worry, namely, that our conscious reasoning about what to do has no effect on our decisions. Some people do harbor this worry. But the experiments described in section 1 do not justify it. In these experiments, by design, there is nothing to reason

442   Alfred R. Mele about.8 What would support the worry on the table now is evidence that whenever we do consciously reason about what to do, that reasoning has no significant effect on what we decide, and the experiments at issue provide no such evidence. Daniel Wegner writes, “it has to be one way of the other. Either the automatisms are oddities against the general backdrop of conscious behavior causation in everyday life, or we must turn everything around quite radically and begin to think that behavior that occurs with a sense of will is somehow the odd case, an add-​on to a more basic underlying system” (2002, p. 144). Similarly, someone might contend that if conscious reasoning does not play a role in producing some decisions, it does not play a role in producing any decisions: “it has to be one way or the other”; either all decisions are produced at least partly by conscious reasoning or no decisions are. I can think of nothing that recommends acceptance of this contention. Consider an experiment in which participants are instructed to move from point A to point B in a lab without using their feet. Point A is marked by a sign that reads “point A.” It hangs from a bar near one end of a set of monkey bars. That bar is within arm’s reach of a participant. A “point B” sign hangs from a bar about five feet away. The clever participants reach up and grab the bar at point A with both hands, raise their feet, and swing—​bar by bar—​from point A to point B. A casual observer reasons as follows: It has to be one way or the other; either all human movements from one place to another involve the use of feet or none do; the movements just observed did not involve the use of feet; so no human movements from one place to another involve the use of feet. The reasoning about conscious reasoning in the preceding paragraph does not stand on much firmer ground. Return to premise a. On a reading of type 2, what the premise asserts is that all decisions about what to do are made independently of conscious reasoning about what to do. That assertion does sound like bad news for moral responsibility. But, as I have explained, it is not even remotely supported by the studies described in section 1. On a reading of type 1, premise a asserts the following: All decisions about what to do are made at times at which the people who make them are not conscious of making them. This assertion is not well supported by the experimental findings, as I explained in section 3. And what I have argued in the present section is that even if the assertion is true, its truth is compatible with widespread moral responsibility. I selected a lag time of 50 ms for dramatic effect. A lag time of 1 ms would have been more dramatic. I leave it to readers to speculate about at what point an increasing lag between a person’s making a decision on the basis of conscious reasoning and the onset of his consciousness of that decision would pose a serious problem for moral responsibility (for discussion, see Mele 2013a). Readers who come to have a view about this may then ask whether there is any good evidence that the lag is never shorter than the threatening lag they have identified. 8  In light of the anecdote I reported in section 3, this is a bit of an exaggeration for at least one participant.

MORAL RESPONSIBILITY AND NEUROSCIENCE     443

5.  Souls to the Rescue? Things are looking bleak for Arguments 1 and 2. (And things are even worse if, as I have argued elsewhere [see note 1], the belief that decisions are made unconsciously in the studies at issue is not warranted by the data.) When I get to this point on a related topic—​ skeptical arguments about free will that are fueled by experiments of the kind described in section 1—​I sometimes am told that I and other like-​minded people are missing the point. The point, it is claimed, is that the studies at issue show that decision-​making is a brain process and anyone who knows what “free will” means knows that free decisions cannot issue from brain processes! Elsewhere, I have quoted several lively expressions of this view of free will (Mele 2014, p. 196). Here I reproduce just one of those quotations—​my personal favorite. It is from neuroscientist P. Read Montague: Free will is the idea that we make choices and have thoughts independent of anything remotely resembling a physical process. Free will is the close cousin to the idea of the soul—​the concept that “you,” your thoughts and feelings, derive from an entity that is separate and distinct from the physical mechanisms that make up your body. From this perspective, your choices are not caused by physical events, but instead emerge wholly formed from somewhere indescribable and outside the purview of physical descriptions. This implies that free will cannot have evolved by natural selection, as that would place it directly in a stream of causally connected events. (2008, p. 584)

Anyone who sides with Montague on this and also believes that moral responsibility depends on free will in such a way that any possible world without free will is a world without moral responsibility will endorse the proposition that moral responsibility is spooky in the same way. Because I discuss the claim that free will depends on substance dualism at some length in Mele 2012, I will be relatively brief here. In response to claims like Montague’s, I conducted some simple survey studies (first reported in Mele 2012). In the study in which physicalism was made most salient, I polled a group of 90 Florida State University undergraduates taking a basic philosophy course that did not deal with free will. The students read the following text: “We’re interested in how you understand free will. Please read the following sentences and answer the questions by circling your answer.” About half then read Story 1 below before they read Story 2; the others read the stories in the opposite order. The students were instructed not to change their answer about the story they read first after reading the other story. The questions were these: “Did John have free will when he made his decision?” and “Is this your first philosophy class after high school?” Participants’ options for answers were “Yes” and “No.” Story 1: In 2019, scientists finally prove that everything in the universe is physical and that what we refer to as “minds” are actually brains at work. They also show exactly

444   Alfred R. Mele where decisions and intentions are found in the brain and how they are caused. Our decisions are brain processes, and our intentions are brain states. Also, our decisions and intentions are caused by other brain processes. In 2009, John Jones saw a 20 dollar bill fall from the pocket of the person walking in front of him. He considered returning it to the person, who did not notice the bill fall; but he decided to keep it. Of course, given what scientists later discovered, John’s decision was a brain process and it was caused by other brain processes. Story 2: In 2019, scientists who work for a secret military organization finally develop a fool-​proof compliance drug. The drug is used to make people decide to do various things. Whenever they give a person the drug and then suggest a course of action, that person is irresistibly caused to decide to take that course of action. They make their suggestions through a tiny computer chip that they implant in a person’s brain. These chemists gave the compliance drug to John Jones, a very honest man. When John saw a 20 dollar bill fall from the pocket of the person walking in front of him, they suggested keeping it. John considered returning it to the person, who did not notice the bill fall; but, of course, he decided to keep it. After all, the combination of the compliance drug and the suggestion forced John to decide to keep it. The results are interesting. Almost three-​quarters (73%) of the respondents said that John had free will when he made his decision in Story 1, and only about one-​fifth (21%) said this about John in Story 2.9 The strong negative response to Story 2 indicates that the great majority of respondents do not take the free-​will-​no-​matter-​what perspective.10 And Story 1, in which physicalism is very salient, yields a strong “free will” response. That story leaves no place in the universe for nonphysical entities to be at work. These findings clash with the claim that ordinary usage of the expression “free will” treats free will as a supernatural power that depends on the existence of immaterial souls. Although I certainly do not see my polls as foolproof, they definitely provide better evidence about ordinary usage of “free will” than does, for example, a randomly selected neuroscientist’s opinion about what that expression means. A recent follow-​up study was part of a much larger study (Vonasch, Baumeister, and Mele 2018). Participants read one of four randomly assigned probes. The two probes that are most pertinent for present purposes are reproduced below. (The study was conducted using Mechanical Turk; 386 participants both completed the questionnaire and passed the comprehension check.) 9  Of the participants who saw Story 1 first (N = 43), 79.07% answered yes to the question about that story, and 25.58% answered yes to the question about Story 2. The figures for those who saw the stories in the reverse order were 68.09% vs. 17.02%. The figures for students who were taking their first philosophy course (N = 53) and those who were not (N = 37) were very similar: grand averages for yes answers were 71.70% vs. 22.64% for the first-​time students and 75.68% vs. 18.92% for the others. 10  On this perspective, see Feltz et al. 2009, pp. 16-​19.

MORAL RESPONSIBILITY AND NEUROSCIENCE     445 Story A: Imagine that scientists finally discover that there are no souls. So, for example, they discover that you don’t have a soul, that your friends and neighbors don’t have souls, and so on. Story B: Imagine that scientists finally discover that there are no souls. So, for example, they discover that you don’t have a soul, that your friends and neighbors don’t have souls, and so on. Like everyone else, John doesn’t have a soul. One day, he sees a twenty dollar bill fall out of the pocket of the person in front of him. He picks it up and keeps it for himself.

Participants responded to three statements about each story on scale from 1 to 7, where 1 is “strongly disagree” and 7 is “strongly agree.” The following statements accompanied Story A: People have free will in this situation; People can make conscious decisions in this situation; People are morally responsible for their actions in this situation. The statements accompanying Story B were specifically about John: John has free will in this situation; John can make conscious decisions in this situation; John is morally responsible for his actions in this situation. Counting answers of 1, 2, and 3 as expressing disagreement and answers of 5, 6, and 7 as expressing agreement, we calculated the percentage of people who agreed with the statements and the percentage who disagreed. The results (rounded to the nearest whole number) are reported in the table below. Story A Free Will Conscious Decisions Morally Responsible

Story B

Agree

Disagree

Agree

Disagree

73% 75% 70%

18% 19% 17%

90% 91% 86%

4% 3% 9%

Here again we have evidence that a substantial majority of nonspecialists do not believe that free will depends on souls, and the same goes for moral responsibility.11 Where does this leave us? Right where we were before substance dualism and souls entered the discussion. Things look very bleak for Arguments 1 and 2.

6.  Parting Remarks In this chapter, I explored the most prominent neuroscientific grounds offered for skepticism about the existence of moral responsibility. Studies that purport to show 11  As discerning readers will have surmised, one of the issues investigated in Vonasch et al. 2018 is effects on lay judgments about free will, moral responsibility, and conscious decision making of the concreteness and abstractness of probes, and another was relationships among judgments on these three topics.

446   Alfred R. Mele that various decisions are made unconsciously are recruited to do the heavy lifting. Elsewhere, I argued that the claim that participants in these studies made decisions unconsciously is unwarranted (see note 1). In this chapter, I argued for two main points. First, even if these participants did unconsciously make decisions, we cannot properly generalize from that to the claim that all decisions are made unconsciously. Second, even if all decisions are made unconsciously because any consciousness we have of decisions we make routinely lags behind the making of them, this leaves the existence of widespread moral responsibility wide open. Anyone who once believed that the studies at issue show that conscious reasoning never plays a role in the production of decisions, and who worried about the existence of moral responsibility for that reason, should now understand how wide of the mark—​and how confused—​that belief is. And anyone who contends that I am missing the point because the existence of free will and moral responsibility depends on the existence of immaterial souls seems to be grasping at straws.12

References Berofsky, B. 2012. Nature’s Challenge to Free Will. New York: Oxford University Press. Campbell, C. 1957. On Selfhood and Godhood. London: Allen and Unwin. Deutschländer, R., M. Pauen, and J. D. Haynes. 2017. “Probing Folk-​Psychology: Do Libet-​Style Experiments Reflect Folk Intuitions About Free Action?” Consciousness and Cognition 48: 232–​245. Feltz, A., E. Cokely, and T. Nadelhoffer. 2009. “Natural Compatibilism versus Natural Incompatibilism: Back to the Drawing Board.” Mind and Language 24: 1–​23. Fischer, J., and M. Ravizza. 1992. “When the Will Is Free.” Philosophical Perspectives 6: 423–​451. Fried, I., R. Mukamel, and G. Kreiman. 2011. “Internally Generated Preactivation of Single Neurons in Human Medial Frontal Cortex Predicts Volition.” Neuron 69: 548–​562. Haynes, J. D. 2011. “Beyond Libet: Long-​Term Prediction of Free Choices from Neuroimaging Signals.” In W. Sinnott-​Armstrong and L. Nadel, eds. Conscious Will and Responsibility (Oxford: Oxford University Press), 85–​96. Kane, R. 1989. “Two Kinds of Incompatibilism.” Philosophy and Phenomenological Research 50: 219–​254. Libet, B. 1985. “Unconscious Cerebral Initiative and the Role of Conscious Will in Voluntary Action.” Behavioral and Brain Sciences 8: 529–​566. Mele, A. 1992. Springs of Action. New York: Oxford University Press. Mele, A. 2003. Motivation and Agency. New York: Oxford University Press. Mele, A. 2009. Effective Intentions. New York: Oxford University Press. Mele, A. 2012. “Another Scientific Threat to Free Will?” Monist 95: 422–​440. Mele, A. 2013a. “Unconscious Decisions and Free Will.” Philosophical Psychology 26: 777–​789. Mele, A. 2013b. “Vetoing and Consciousness.” In A. Clark, J. Kiverstein, and T. Vierkant, eds. Decomposing the Will (Oxford: Oxford University Press), 73–​86.

12  This article was made possible through the support of a grant from the John Templeton Foundation. The opinions expressed here are my own and do not necessarily reflect the views of the John Templeton Foundation.

MORAL RESPONSIBILITY AND NEUROSCIENCE     447 Mele, A. 2014. “Free Will and Substance Dualism: The Real Scientific Threat to Free Will?” In W. Sinnott-​Armstrong, ed. Moral Psychology, Volume 4: Free Will and Moral Responsibility (Cambridge, MA: MIT Press), 195–​207. Mele, A. 2018. “Free Will and Consciousness.” In D. Jacquette, ed. Bloomsbury Companion to the Philosophy of Consciousness (London: Bloomsbury Publishing), 371–​388. Miller, J., and W. Schwarz. 2014. “Brain Signals Do Not Demonstrate Unconscious Decision Making: An Interpretation Based on Graded Conscious Awareness.” Consciousness and Cognition 24:12–​21. Montague, P. R. 2008. “Free Will.” Current Biology 18: R584–​R585. O’Connor, T. 2000. Persons and Causes: The Metaphysics of Free Will. New York: Oxford University Press. Soon, C. S., M. Brass, H. J. Heinze, and J. D. Haynes. 2008. “Unconscious Determinants of Free Decisions in the Human Brain.” Nature Neuroscience 11: 543–​545. van Inwagen, P. 1989. “When Is the Will Free?” Philosophical Perspectives 3: 399–​422. Vonasch, A., R. Baumeister, and A. Mele. 2018. “Ordinary People Think Free Will Is a Lack of Constraint, Not the Presence of a Soul.” Consciousness and Cognition 60: 133–​151. Wegner, D. 2002. The Illusion of Conscious Will. Cambridge, MA: MIT Press.

CHAPTER 22

Resp onsibi l i t y a nd C onsciou sne s s Matt King and Peter Carruthers

1.  Introduction Intuitively, consciousness matters for responsibility. A lack of awareness generally provides the basis for an excuse or at least for blameworthiness to be mitigated. If you are aware that what you are doing will unjustifiably harm someone, it seems you are more blameworthy for doing so than if you harm them without awareness. There is thus a strong presumption that consciousness is important for responsibility. The position we stake out in this chapter, however, is that consciousness, while relevant to moral responsibility, is not necessary. The background for our discussion is an emerging consensus in the cognitive sciences that a significant portion, perhaps even a substantial majority, of our mental lives takes place unconsciously. For example, routine and habitual actions are generally guided by the so-​called “dorsal stream” of the visual system, whose outputs are inaccessible to consciousness (Milner and Goodale 1995; Goodale 2014). And there has been extensive investigation of the processes that accompany conscious as opposed to unconscious forms of perception (Dehaene 2014). While there is room for disagreement at the margins, there is little doubt that our actions are much more influenced by unconscious factors than might intuitively seem to be the case. At a minimum, therefore, theories of responsibility that ignore the role of unconscious factors supported by the empirical data proceed at their own peril (King and Carruthers 2012). The crucial area of inquiry for those interested in the relationship between consciousness and responsibility concerns the relative strength of that relationship and the extent to which it should be impacted by findings in the empirical sciences. While the nature of consciousness continues to be much debated, most of those debates concern the nature and extent of so-​called “phenomenal” consciousness (which is how experiences feel, or what they are like for their subjects). This is arguably

Responsibility and Consciousness    449 irrelevant to our topic, for the intuition that consciousness is tightly connected to responsibility has little to do with the “feely” character of conscious experience, we submit, and much more to do with awareness.1 It is lack of awareness of the circumstances or consequences of one’s actions that is thought to excuse, not the presence or absence of felt qualities of one’s experiences while acting. And there is much less dispute about the nature of awareness, which seems to coincide with the notion of so-​called “access consciousness” (Block 1995). Moreover, the best—​and widely accepted—​theory of access consciousness is the “global workspace” or “global broadcasting” model (Baars 1988, 2002; Dehaene 2014). On this account, mental representations become conscious when they pass a threshold of activation (generally resulting from attentional signals being directed toward them), which results in them being made widely available to multiple subsystems in the brain—​specifically, to systems for planning, decision-​making, and verbal report; for forming long-​term memories; and for issuing in full-​blown emotional reactions. At any rate, this is the framework for thinking about consciousness that we adopt in the present chapter.

2.  The Relevance of Consciousness Even if consciousness is intuitively relevant to responsibility, it remains an open question precisely how and to what extent it is relevant. One natural suggestion is that consciousness is in some sense required for responsibility. In order to be responsible for an action, its main psychological causes (not only its goals, but the beliefs and perceptions that guide it) must be conscious. Call this the consciousness condition. While we agree that consciousness is relevant to responsibility, we deny the consciousness condition. The consciousness condition has some intuitive support. We often attempt to excuse our behavior by pointing out that we were unaware of relevant details. It is customary to excuse bumping into someone by saying, “I didn’t see you there.” When we lack important information, our ignorance can serve to mitigate blame. More significantly, we are more apt to excuse behavior the more unconscious it is. In some cases, such as that of a sleepwalker, we may even exculpate the agent altogether. Notably, however, it is just as apparent that we don’t always excuse people for lack of awareness. If I forget a friend’s birthday, I’m apt to feel guilty about doing so and thus apologize. Moreover, such responses appear rational. We hold others to account not just for what they notice and do, but what they fail to notice, and so fail to do (Smith 2005). It is common practice to blame people for their negligent actions, where they are unaware of an unjustified risk of harm their conduct poses (usually through a failure to take reasonable care).2

1 

2 

Though see Shepherd 2015. See also n. 8. This is not to say such practices are without controversy. See King 2009.

450    Matt King and Peter Carruthers There is disagreement about whether such cases really show that awareness isn’t required. Defenders of the consciousness condition might insist that responsibility is only possible in such cases through some prior exercise of consciousness-​involving responsibility. So, while the agent at the time is unaware that today is her friend’s birthday, we can nonetheless “trace” her responsibility back to some prior responsible action (or omission).3 Such claims are controversial, however. First, any such tracing often leaves us with an identical problem to resolve. For example, if one traces culpability for forgetting a friend’s birthday back to culpability for failing to take sufficient steps to remember (such as putting a reminder in one’s calendar), then we would still have to explain responsibility for that prior forgetting.4 So we would need some final terminus for responsibility wherein the consciousness condition was satisfied. That seems implausible for a wide variety of cases. Second, one’s culpability for some prior act (or omission) is often significantly less than one’s culpability for the present one, so the strategy still leaves us without a fully adequate explanation.5 While we allow that some cases of unconscious action might be explained via tracing back to some previous conscious episode, we deny that all cases are best explained this way. A full defense of this claim, however, would steer us too far aside from the main discussion. Instead, we will simply observe that we often fail to excuse others for lack of awareness, and in doing so we do not always look to trace their culpability back to some prior episode. More dramatically, not only does the fact that some action was unconsciously caused not always mitigate responsibility, it occasionally seems to exacerbate it. It is a familiar thought that what we do spontaneously and without conscious reflection can reveal a deeper truth about ourselves than more careful, deliberate action. Rather than settling matters, then, reflection on our ordinary practices reveals that we have diverging commitments when it comes to the consciousness condition. On the one hand, we often excuse behavior precisely because of its unconscious nature. On the other hand, we don’t universally excuse such behavior. Any discussion of whether, and in what sense, consciousness is necessary for responsibility will need to engage with these observations. Beyond our ordinary experience, scientific findings raise a variety of concerns. Some results have been taken to suggest that consciousness plays no significant role in the production of action or responsible choice (Libet 1993; Wegner 2002). This would suggest, if consciousness is required for responsibility, that we are much less responsible than we think. But such research has been widely undermined both philosophically and

3  E.g., see Smith 1983; Wieland and Robichaud 2017. For alternative strategies, see King 2014, 2017; Sher 2009; Smith 2011. 4  See King 2009, 2014; and Sher 2009 for critiques of this form. 5  To use an example from Sher 2009, forgetting one’s dog in the car on a hot day is worse than deciding to bring the dog with you while running errands on a hot day. And, of course, one would still need to show that one was conscious of the risk to the dog for the tracing strategy to work.

Responsibility and Consciousness    451 empirically.6 This shouldn’t be surprising, since it is quite implausible to suppose that consciousness has no role to play in action production and responsibility, even if it isn’t strictly necessary. More plausible worries concern the manner in which our actions and decisions might be affected or directed in ways of which we are unaware. For instance, it has been repeatedly found that people can have affective biases of which they lack awareness (and which influence their everyday behavior), as well as having their expectations of other people influenced by racial, gender, and other stereotypes that they wouldn’t endorse consciously (Banaji and Greenwald 2013).7 There is plenty of disagreement about the significance of such findings. Controversy over the research is perhaps to be expected given our general ambivalence about the precise relevance of consciousness to responsibility. As already indicated, unconscious factors seem to excuse in some kinds of cases but not in others, while in still others they may actually heighten responsibility. Given the lack of consensus, both from our ordinary practices and scientific research, greater attention is required to refine the relationship between consciousness and responsibility. In the limited space available to us, it is impossible to canvass every possible variant of the consciousness condition, nor every area in which consciousness (or unconsciousness) might be implicated. In light of this, we plan to examine the consciousness condition through three of the types of case so far mentioned: automaticism (like sleepwalking), forgetting, and implicit bias. While by no means exhaustive, they represent a suitable spectrum along which to pursue the issue.

3.  Some Generalities Regarding Responsibility Before proceeding to cases, however, it is important to consider in what way consciousness might be relevant to responsibility generally. After all, while consciousness could be brutely necessary, in such a way that the consciousness condition would be independent of any other necessary conditions on responsibility, such a position would be

6 

For a thorough review of the original research and subsequent critiques, see Levy 2014, pp. 14–​ 26. Additionally, it’s worth noting that most appeals to Libet-​and Wegner-​style experiments assume something like the consciousness condition, rather than arguing to it from the data. 7  In addition to implicit bias, some claim that seemingly irrelevant features of our situation can affect our judgments or reasons-​responsiveness in ways of which we aren’t aware (i.e., the so-​called “situationism” literature). To cite just one example, having previously held a hot cup of coffee can positively influence one’s assessment of others (Williams and Bargh 2008). But there is simply no consensus either on the exact nature of these effects, their scope, or the degree to which they threaten responsibility. Some relevant recent discussions include Caruso 2015; Herdova 2016; Holroyd 2012; Mavda 2018; and Schlosser 2013.

452    Matt King and Peter Carruthers explanatorily disappointing.8 We would be left without an answer for why consciousness matters. Instead, it is more promising to suppose that if consciousness is necessary, it is because it makes possible or facilitates the satisfaction of some other condition on responsibility. Indeed, even though we sometimes excuse people for actions done in ignorance, we don’t generally think this is simply because they lacked awareness. Rather, a more plausible idea is that consciousness tends to correlate with, or is indicative of, something else that is itself essential for responsibility. At least, this is how we propose to proceed. We will examine the prospects for the consciousness condition within a basic framework of what else matters for responsibility. There is a plethora of theories of moral responsibility. Indeed, it is fashionable these days to distinguish different conceptions of responsibility, which differ from one another in their dimensions or requirements.9 For our purposes, we propose to work with a unified, monistic conception of responsibility, in which to be responsible is to deserve (or to merit) blame for bad things and praise for good things. This is, in our estimation, a perfectly common conception and is the one most often invoked in traditional philosophical debates about free will. Moreover, it arguably comes with the most stringent conditions.10 So if consciousness isn’t required for responsibility thus conceptualized, we should expect the result to generalize to other accounts as well. Nevertheless, we don’t rule out that there might be further complexities that arise in evaluating the prospects for the consciousness condition across other plausible conceptions of responsibility. Instead, we aim for a useful generality in our discussion, so that those who prefer alternative characterizations of responsibility can be invited to modify our discussion accordingly. We believe the basic claims will remain unchanged. Indeed, no matter the account one offers of responsibility, some general features of agency tend to be prominent: specifically, choice, control, and coherence.11 These features are not exclusive of one another, and there is plenty of disagreement over which are necessary for or most central to responsibility. Nevertheless, they represent a plausible core of responsibility-​relevant capacities.

8  Shepherd 2015 reports that folk judgments of responsibility do seem to be influenced by judgments about phenomenal consciousness. But it isn’t clear from the findings what the precise relation between those two judgments is or how to apply them to informational awareness. In any case, even if the folk judged consciousness to be necessary, that wouldn’t show it was directly so. 9  David Shoemaker (2015) has perhaps the strongest such defense, contending that there are multiple independent conceptions of responsibility that differ in terms of their conditions and associated warranted responses. Others contend, for example, that there are conceptions that do not involve any claims about desert (Pereboom 2016). While growing in popularity, this approach is not without its opponents (cf. Smith 2012). 10  Among those that are remotely plausible, that is. We grant that ours isn’t the most demanding conception out there. Strawson 1986, for instance, characterizes moral responsibility to be a matter of one’s deserving eternal punishment or reward. That strikes us (and many others) as far too demanding a notion, and one that doesn’t correspond to whatever conceptions might be at play within our ordinary practices of holding others responsible. 11  The following characterizations are modeled on King and May 2018.

Responsibility and Consciousness    453 Consider choice first. Some accounts state that responsibility requires agents to select from among genuinely available alternatives, free of external constraints (Kane 1996). What matters here is the agent being free to choose. (While not required, most such views favor incompatiblism between free will and determinism.) Other theories stress the capacity to exercise effective control (Mele 1995). Since actions that an agent didn’t control, like accidents, look to be ones for which she isn’t responsible, a popular requirement on responsibility is that the agent guided her conduct appropriately. Many theories develop this idea in terms of being “reasons responsive” (Fischer and Ravizza 1998). Agents exercise the relevant control in cases where they are capable of being sensitive to the salient moral features of their circumstances, and of guiding their behavior appropriately in light of those reasons. Still other theories emphasize the coherence between the action (and its motivations) and the agent’s overall psychology (Frankfurt 1971; Sripada 2016). The thought is that an agent is more responsible for an action the more it reflects her moral self, or the better integrated its motivations are with the full set of her motivations and values. Consider the difference between a committed racist who utters a racist remark and someone who simply makes an insensitive comment (one that is out of character). There is an intuitive difference here that is captured by the idea that responsible action is one that manifests the agent’s true commitments and values. We can calibrate our understanding of the consciousness condition to these core capacities. On the first idea, if what matters is choice, we might wonder whether consciousness is required for choices to be genuinely available or properly selected by an agent. On the second, we might wonder whether we can respond in the requisite way to reasons represented only unconsciously. On the third, we can wonder whether unconscious mental processes properly reflect our values or not. Let us first consider, and quickly reject, the idea that choice among alternatives grounds the consciousness condition. While few claim that consciousness is required for free choice, the idea is not without its defenders. The thought is that consciousness provides the space within which we might find indeterministic mental processes, sufficient to underwrite the sorts of distinctive capacities that free will and responsibility require (Hodgson 2012). If neural processes are governed by deterministic physical laws, perhaps consciousness might supply the right sort of non-​deterministic arena in which free choices can occur. Such views are not popular, and for good reason. First, they are overly speculative, often optimistically relying on future science to vindicate their claims. Second, they confront a dilemma, depending on whether or not physicalism about consciousness is endorsed. If physicalism is maintained, then it would have to be claimed that only certain sorts of neural processes are indeterministic (namely, those implicated in consciousness), because otherwise it would be possible for the requisite indeterminism to be present in unconscious processes. Alternatively, if physicalism is denied, then it would need to be claimed that nonphysical processes (e.g., conscious deliberation) can have physical effects (action). The first horn of the dilemma leaves us with two fundamentally

454    Matt King and Peter Carruthers different types of neural processes, whereas the second conflicts with the causal closure of physics. An alternative way of motivating the consciousness condition on choice might be to claim that the ultimate foundation for responsibility requires a kind of deliberative competition between options. For example, Kane (1996) argues that we are responsible for our choices because of previous “self-​forming actions,” which are (roughly) decisions in which an agent is torn in two directions (e.g., whether to take a new job in a new city or stay where one is), and it is undetermined which option they will choose. Whatever the agent decides will, in some sense, help shape the person they are to be. One might then insist that such indeterministic competition only occurs in conscious deliberation. However, we still don’t have good grounds for the consciousness condition. We have already seen that it is unlikely that consciousness can supply the requisite indeterminism. And in addition, Kane must claim that the kind of choice that leads to self-​forming actions will always involve conscious deliberation. This means limiting responsible actions to those that are “self-​forming” in the required sense (or that flow from aspects of one’s character fixed by such actions). While we cannot definitely reject such a view about responsibility here, it is a significant, and in our view implausible, commitment to take on.12 Although we leave aside the relevance of consciousness to choice, it might still be relevant for control or coherence. Indeed, the most recent and sustained defense of the consciousness condition argues that consciousness is required on either of these two grounds. This is where we go next.

4.  Defending the Consciousness Condition In the most thorough defense to date, Neil Levy (2014) argues that what matters for responsibility is informational awareness. It is our lack of awareness of morally relevant features of our conduct that excuses us, when we are excused. Drawing on the global broadcasting model of consciousness, he holds that the basic function of consciousness is to integrate our attitudes with one another and coordinate our activities over time. As such, consciousness is necessary for a kind of rational agency. Behavior driven by conscious states will be better integrated—​more coherent and consistent—​than behavior driven by unconscious states. Consciousness is thus required to control our behavior properly and to display the requisite flexibility in our responsiveness to reasons. Additionally, Levy argues that consciousness is necessary if a coherence condition is to be fully met. Since conscious contents are those that are globally broadcast and 12 

Notice, also, that such a view will be required to take on some sort of “tracing” mechanism, to account for the very plausible cases of responsible choice that do not involve torn decisions.

Responsibility and Consciousness    455 made available to other consuming systems in the mind, the contents of beliefs, values, and other attitudes, when consciously entertained, will be better integrated with our other beliefs, values, and attitudes. Only conscious contents can be compared with and evaluated against the full suite of our other attitudes, beliefs, and values. Thus, when we are conscious of the springs of our actions those actions more fully express our true selves, since more of our evaluative stance has been brought to bear on them. The conscious decisions we come to, and the actions we then perform, result from competition among a larger (perhaps maximal) set of our personal-​level attitudes. Levy emphasizes personal-​ level attitudes because, he says, only mental representations that are broadcast widely in the mind will be appropriately predicated of the person herself, rather than some lower-​level component of her mind. Since we hold persons morally responsible, we are interested only in facts about the person rather than about her subsystems. I am no more responsible for how certain components of my visual system operate than I am for my digestion, so the thought goes. And in order for me to be informationally aware of something, it must be personally available—​that is, available as a personal-​level representation. On Levy’s account, personally available information is information that is (1) easily and effortlessly retrievable and (2) online. Information that is easily retrievable is apt to be triggered by many ordinary cues (rather than leading questions, say). To be online is to actually play a role in guiding behavior. As an empirical fact, according to Levy, the only information that will meet both conditions is conscious information (that is, information that is globally broadcast). Some have denied that consciousness, in this sense, is necessary for responsibility. For instance, Arpaly (2002) notes a number of cases in which agents act and are seemingly responsible for doing so despite their motivations remaining introspectively opaque. They are not aware, it seems, of why they do the things they do, and yet they strike many of us as responsible, even praiseworthy, for acting. Huck Finn is a popular case study. In a famous scene of Mark Twain’s novel, Huck has an opportunity to turn in the escaped slave Jim to a group of men hunting for him. Huck believes he ought to reveal Jim’s location (for he shares the mistaken moral views of his time), but he can’t bring himself to do it, whether out of unconscious sympathy or an unrecognized recognition of their shared humanity. Interpretations of the case vary, but at least some theorists are led to deny a consciousness condition because of such cases, wherein agents act for reasons of which they are unaware and yet reveal their praiseworthy or blameworthy qualities. Likewise, we at least occasionally experience guilt when we realize that our past motivations for some action were less than noble. And we are all familiar with cases where people have “blindspots.” In such cases, we recognize the true motivations of another (and hold her responsible accordingly), although the person herself is unaware of those motives. Levy is committed to denying responsibility in such cases. Though agents may sometimes have their actions guided by online representations, this is insufficient for responsibility if those guiding representations fail to be incorporated into their personal-​level representations. He gives the example of a woman who forgets her anniversary and so plans a date with friends, but in getting dressed she chooses a particular necklace to wear

456    Matt King and Peter Carruthers because it is the necklace her spouse gave her on their first anniversary. The belief that today is her anniversary is online, since it is guiding her behavior, but while she might be able to retrieve that information if asked a leading question, it isn’t easily retrievable under a wide variety of circumstances. This seems to follow as a consequence of her forgetting the anniversary in the first place. For she is likely to have encountered cues recently that should have led to retrieval of the belief. For example, she saw her husband multiple times over the past week, she looked at the date when planning the dinner with friends, and she regularly passes photos from their wedding in the hallway. If she forgot their anniversary despite regular cues related to that belief, it would seem that “the mere fact that some of her behavior is guided by the knowledge that it is her wedding anniversary does not seem to establish that [she is responsible for forgetting]” (33). Information that is only easily retrievable is likewise insufficient to underpin responsibility, for Levy. This is because one can, through happenstance, fail to encounter any of the cues that would lead to easy retrieval of the information if they were present. Suppose, for example, that had the woman been home she would have remembered her anniversary as she always does, but she is away on a trip in the days leading up to it, traveling in a remote part of the country with no contact with her husband. With nothing to remind her, she may remain unaware of the relevant information, and we would be more apt to excuse in these circumstances. As Levy says, the information needs to be both easily retrievable and online for responsibility. For Levy, then, consciousness is generally relevant because it affords us the ability to integrate divergent beliefs and values, and to coordinate our activities over time. More specifically, only when we are conscious of the morally significant features of our actions is such information personally available and hence reflective of the person’s moral agency. Only then can we be responsible for our actions.

5.  Unconscious Evaluation Levy’s defense of the consciousness condition is unconvincing, and its shortcomings suggest we should be skeptical of any such necessary condition on responsibility. In this section, we argue that the relationship between consciousness and responsibility is more complicated than the consciousness condition allows. We agree that consciousness is relevant to responsibility, but we are responsible for much that is unconscious. To bring out these points, we return to three common (and divisive) kinds of case in the literature: automaticism (e.g., sleepwalking), forgetting, and implicit bias.

5.1  Automaticism It can be tempting to think there must be some necessary consciousness condition. After all, at the extreme, there are cases of apparent automaticism, where even relatively

Responsibility and Consciousness    457 complicated actions are performed by individuals who nonetheless have little or no awareness of what they are doing. Kenneth Parks provides a vivid (and disturbing) example. In May of 1987, after falling asleep, Parks got up, got dressed, and drove 23 kilometers to his in-​laws’ house, whereupon he assaulted his father-​in-​law and repeatedly stabbed his mother-​in-​law, before driving to the police station. Once there, he told police that he thought he had killed people, only then (apparently) noticing the severe wounds to his own hands (Broughton et al. 1994). Parks’s memory of the events was patchy at best, and his defense of somnambulism was credible given a history of disordered sleep, plausible triggering conditions (he was under extreme stress at the time), and lack of any motive (he got along very well with his in-​laws). The jury agreed, and Parks’s acquittal under a plea of “non-​insane automatism” was upheld by the Canadian Supreme Court (R. v. Parks 1992). The evidence suggests that most of what Parks did was driven by low-​level, unconscious representations. And it might be on those grounds alone that he was acquitted. But we think it would be a mistake to conclude from this that consciousness is necessary for responsibility. This is because the underlying causes of sleepwalking are not equivalent to the full range of unconscious processes that operate in normal non-​pathological behavior. Two findings are especially relevant here. The first is that sleepwalking occurs during the deepest phases of non-​rapid-​eye-​movement sleep (Plazzi et al. 2005). The second is that sleepwalking is associated with activation of sensory-​motor pathways between the thalamus and motor regions of the cingulate cortex, combined with sustained deactivation of most other regions of the cortex, including prefrontal executive and decision-​making areas (Bassetti et al. 2000; Terzaghi et al. 2012; Januszko et al. 2016). Moreover, although the amygdala can remain responsive to low-​level threat-​like stimuli during sleepwalking, the remaining components of the sleepwalker’s cortical and subcortical evaluative networks are likewise deactivated (Terzaghi et al. 2012). A plausible explanation for why Parks wasn’t responsible for his actions, then, is not merely that his actions were unconsciously caused, but rather that they were caused in ways that bypassed both his value systems and his decision-​ making capacities, all of which were shut down by sleep. His actions thus reflected nothing of his beliefs and values.

5.2  Forgetting The fact that a process is unconscious does not by any means entail that it bypasses relevant evaluative systems of the agent. This is significant, for if unconsciously caused action can still reflect an agent’s assessment of her reasons or values, then one could still be responsible for unconscious action, either through control or coherence. Indeed, there are grounds for thinking that evaluation is already implicated in determining whether or not a content becomes conscious in the first place. In particular, the evidence suggests that whether or not some information gets globally broadcast often reflects a

458    Matt King and Peter Carruthers competition for attention, resulting in a decision to attend (albeit an unconscious one). This claim will require some background and elucidation.13 It is widely agreed among cognitive scientists that working memory provides the workspace within which representations can be consciously activated, sustained, and manipulated (Miyake and Shah 1999; Baddeley 2003; Alloway and Alloway 2013). It is also widely agreed that entry into working memory (and consciousness) depends heavily on the direction of top-​down attentional signals, which are under intentional control (Gazzaley 2011; Tamber-​Rosenau et al. 2011; Unsworth et al. 2015). These are thought to boost the activity of some representations over the threshold for global broadcasting while at the same time suppressing others. Hence there are such phenomena as inattentional blindness, where people can remain unconscious of even quite salient stimuli because their attention is directed elsewhere (Dehaene 2014).14 In addition to the top-​down attentional network controlled by current goals and intentions, there is also a competing network—​generally described as a saliency network—​ whose function is to compete for the resources of top-​down attention, allowing a new set of contents to become conscious (Corbetta and Shulman 2002; Corbetta et al. 2008; Menon and Uddin 2010). The term “saliency” isn’t really appropriate to describe this network, however, since it is sensitive to much more than just low-​level features such as stimulus intensity (a loud bang) or sudden change (a shape starting to wriggle in the grass as one approaches). On the contrary, the network also evaluates unattended stimuli (as well as unattended memories that may have become active in the context) against standing values and standing goals (Awh et al. 2012). So it is better described as the relevance network. The interactions between these two systems are responsible for the well-​known cocktail-​party effect. One might be fully engaged in conversation with someone at a party, attending to (and thus conscious of) what they are saying. But at the same time the relevance network monitors the surrounding conversations. If someone happens to use one’s name, this is apt to win the competition for top-​down attentional resources, and the sound of the name pops suddenly into one’s stream of consciousness. One has, in effect, decided that the mention of one’s name is more relevant to one’s concerns than whatever one’s friend happens to be saying, and so one directs one’s attention accordingly. Indeed, scientists who model such phenomena think that one is engaging in the functional equivalent of a cost-​benefit analysis, weighing the costs and benefits of one’s current focus of attention against the alternatives (Kurzban et al. 2013).15 13 

The ideas outlined in the next few paragraphs are fully developed, and supported with an extensive survey of the empirical literature, in Carruthers 2015. 14  In one famous example, participants watching a basketball game are directed to count the number of times a player in a white t-​shirt touches the ball. In the course of the game a man in a gorilla suit walks onto the center of the court, beats his chest, and walks off. Participants tasked with attending to the players in white t-​shirts often fail to notice the gorilla altogether, despite his being fully visible to participants not so tasked. Examples are readily available on YouTube. 15  See Carruthers 2015 for an extended argument that unconscious decisions to redirect attention genuinely deserve to be described as such. They are events of the same type, realized in the same brain networks, as the decisions that issue from conscious forms of reasoning and deliberation.

Responsibility and Consciousness    459 The interactions between these networks are also thought to be responsible for the phenomenon of mind wandering (Fox et al. 2015). One might be attending to a task, such as reading an article for work. But at the same time the relevance network will be monitoring surrounding stimuli as well as memories that have been sparked into activity by the circumstances. If any of these surrounding stimuli or memories are deemed relevant enough (and especially if one’s primary task is experienced as aversive), then one will be apt to find oneself suddenly imagining a tropical beach or planning a recipe for dinner. With this background in place, return to the phenomenon of forgetting. Consider Levy’s example of the wife who forgets that today is her anniversary. We surmised earlier that any number of cues could have activated the belief, from passing some wedding photos to looking at the date on the calendar when planning her meeting with friends. Levy concludes that the belief isn’t personally available because it wasn’t in fact activated and globally broadcast. An alternative possibility, however, is that it wasn’t broadcast precisely because it wasn’t judged important enough given her current goals and values. The sight of the date might indeed have evoked the knowledge that it is her anniversary, but this wasn’t deemed relevant enough to attract attention, given her focus on arranging her meeting. As a result, it never became conscious. Recall that Levy denies that any (unconscious) component of the mind can constitute a state or attitude that is attributable to the person herself, rather than just a component. Only consciously entertained representations are properly of the person (personal-​level), as they better reflect the agent’s evaluative stance, having been available for competition and coordination with the agent’s other values and goals. Only such representations can thus ground responsibility for subsequent action. However, rather than failing to be a personal-​level judgment and hence being an unsuitable candidate for attribution to the agent herself, unconscious competition for attention looks prima facie relevant both to an agent’s evaluative stance and her responsiveness to reasons. Think of how natural it would be to say of the wife that she doesn’t care enough about her anniversary, and for her spouse to blame her for forgetting. Indeed, if she cared enough or thought it important enough, it would be more likely that the cues she encounters would be judged significant enough to attract attention and enter consciousness—​which is just to say that if she cared enough, she wouldn’t have forgotten. Of course, accepting a role for unconscious judgments of the sort we have outlined isn’t sufficient to settle questions of responsibility. That she doesn’t attend to the activated belief that today is her anniversary doesn’t by itself show that she fails to care enough about her marriage or her spouse. All it shows is that, under those various circumstances in which the belief was activated, it failed to direct her attention when competing against her current goals. Nonetheless, the nature of such competition should plausibly inform our judgments about (and degree of) her blameworthiness for forgetting. Indeed, it will be consistent with our ordinary assessments of such cases. If she has been harried at work and preoccupied with important matters, then we might expect her attention to be primarily directed at those representations, rather than being easily shifted to her anniversary,

460    Matt King and Peter Carruthers especially if the latter is particularly non-​salient this year (say, it is their 12th anniversary). It is precisely such considerations that could serve to mitigate her blameworthiness for forgetting, or would at least be taken to support understanding on the part of an aggrieved spouse. In contrast, we wouldn’t as easily expect her attention to continue to be held by more trivial matters, like her fantasy football league rosters. And, likewise, such considerations wouldn’t plausibly excuse her at all. Of course, forgetting an anniversary, while significant, is not especially morally serious (although this might vary with particular relationships). Tragically, people do forget about more important things, such as leaving a child unattended in a hot car. It is implausible to suppose such parents fail to care about their children. After all, they are sincerely grief-​stricken when realizing what they have done. The unsettling implication of the role of unconscious attitudes in the competition for attention, however, is that it may well be the case that such parents did fail to have proper concern for their children under the circumstances. If the belief that the child was still in the car was activated (for example, when they used their keys to open their office door) but attention was nevertheless firmly focused on the meeting they were running late for, then the fact that there was risk to the child was not assessed as significant enough to redirect their attentional focus.16

5.3  Implicit Bias The last 20 years have seen an explosion of interest in forms of implicit (unconsciously operating) bias toward members of social groups (Banaji and Greenwald 2013). There are two main forms of implicit bias. One is affective in nature, such as negative feelings (e.g., disliking) toward Black people. The other is cognitive, and includes a variety of stereotypes, such as “men are leaders,” “women are caring,” “Asians are good at math,” and so on. Measures of implicit attitudes are known to dissociate from people’s overt attitudes, and apparently predict some real-​world behavior independently of the latter.17 The question for us is whether implicit attitudes provide a locus of responsibility. The answer may depend, in part, on the best account of the nature of those attitudes. On 16 

Much will depend on the particular facts in cases of this sort. It is consistent with forgetting one’s child in the car that no belief about the risk of harm to the child is activated (even if the presence of the child in the car is). If not activated, then we agree that failing to direct attention to the belief that the child is still in the car doesn’t reflect any agential judgment about relative importance. Of course, in most actual cases we won’t have sufficient evidence about which representations were activated to draw firm conclusions. 17  For a meta-​analysis suggesting that the real-​world effects of implicit attitudes are only minor ones, see Oswald et al. 2013. For a careful and measured reply, see Greenwald et al. 2015, who point out that even minor effects can multiply over time, as well as across large groups of individuals. They also point out that many of the null results included in Oswald et al.’s meta-​analysis concern effects that should not have been theoretically predicted, such as testing for approach or avoidance behavior (which are among the predicted effects of affective attitudes) following measures of implicit stereotypes (which are cognitive in nature).

Responsibility and Consciousness    461 one view, they are low-​level associations among properties, such as between Black and bad or between woman and caring (Gawronski and Bodenhausen 2006). As such, they wouldn’t be characterized by Levy (2014) as personal-​level, and hence he would say that one cannot be held responsible for their effects. On another view, in contrast, implicit attitudes are just attitudes like any other (Carruthers 2018). The difference is merely that other influences mostly swamp the effects of implicit biases on one’s verbal behavior (e.g., when asked how one feels about Black people, or whether one would endorse the claim that women are caring). Other influences would include one’s egalitarian beliefs, one’s desire to avoid social criticism, a desire to protect one’s positive self-​image, and so on. One consideration in support of this view is that the gap between implicit and explicit measures of one’s attitudes is significantly reduced when one is required to answer swiftly, or when one has to answer while under cognitive load of some sort (Hofmann et al. 2005). Under such conditions, one’s implicit attitudes are more likely to break through and find expression in one’s speech. On the account of implicit stereotypes developed by Carruthers (2018), for example, stereotypes are just generic beliefs, like “birds fly” or “mosquitoes bite.” Generics seem to be the mind’s default mode of generalizing. They are acquired extremely swiftly (especially for negative properties), and even “some,” “most,” and “all” statements are apt to be stored and recalled as generics (Leslie and Gelman 2012; Leslie 2015). While most people will assent to “mosquitoes bite,” even though they know that strictly speaking only female mosquitoes do, they are much less likely to assent to “Black men are dangerous” or “women aren’t leaders.” But this isn’t because of any intrinsic difference in the beliefs in question. It is rather that one’s answers, in the latter cases, are moderated by other attitudes. No one thinks we need to be fair to mosquitoes or to give male mosquitoes their due. But many of us think that individual people should be treated as such, and that one shouldn’t form expectations about them in advance. Nevertheless, if one has acquired the relevant stereotypes, then one will form such expectations when one is responding unreflectively. If implicit attitudes are just attitudes that one doesn’t consciously acknowledge, then it is obvious that actions that manifest such attitudes will reveal something about one’s beliefs and values; and as such, those actions may be ones for which one can be held responsible. But what if implicit attitudes are merely associations among ideas? Here, matters may be more complicated. If exposure to a Black face causes a negative affective response in a subject, this implies that she negatively evaluates that Black face. To the extent that affect is a product of one’s evaluative systems, therefore, even low-​level associations might reflect an agent’s (perhaps unmediated) evaluations. Such affects may be relevantly indistinguishable, however, from similar affective responses to, say, foods the subject dislikes. Of course, there is arguably an important moral difference between disliking Black people, on the one hand, and disliking broccoli, on the other. Such feelings may nonetheless still tell us something about the agent’s values. Regardless of whether we can be responsible for our implicit attitudes (understood as associations) directly, however, holding individuals responsible for the actions that result seems much less problematic. After all, complicated actions, like hiring decisions,

462    Matt King and Peter Carruthers are not plausibly driven solely by some low-​level association between ideas. Indeed, an extensive range of other representations and contents will be activated, many of which will be globally broadcast, throughout any responsible deliberative process. A hiring agent who rejects an otherwise qualified candidate out of an associated dislike or for a perceived lack of leadership qualities (when these associations fail to fit the facts) misevaluates the candidate. Such errors look to be reasonable candidates for responsible (indeed, blameworthy) action. Knowing about these biases, for example, would provide one with good grounds for critiquing the agent’s evaluations of the candidates, and perhaps for sanctioning her accordingly.18

6.  What Is Left for Consciousness? While we have critiqued the consciousness condition on responsibility, we concede that consciousness is often important for responsibility. Decisions that are made following conscious reflection will implicate a wider range of one’s beliefs and values than decisions that are spontaneous and that manifest the workings of unconscious processes of one sort or another. This is because spontaneous, intuitive, decisions will generally result from a narrow range of values and information. For example, a swiftly made decision not to hire a Black candidate because “he doesn’t feel right for the job” may result simply from an underlying affective bias against Blacks. Had the person reflected further, however, a wider range of attitudes would have come into play, including egalitarian beliefs, strongly felt affective reactions against injustice, and so on. Indeed, there is surely a world of difference between a harried and rushed interviewer who makes an unreflective spur-​of-​the-​moment (but biased) decision, and someone who reflectively decides not to hire the Black candidate because he doesn’t like Blacks and doesn’t want any to work in his company. That there is a difference, however, is consistent with claiming that both are blameworthy. It seems to us obvious that reflective deliberation often increases one’s blameworthiness for an action of the very same type as one performed spontaneously and unreflectively. If I carefully and consciously calculate the precise words with which to injure you most, I am more blameworthy than if I unthinkingly snap at you after a hard day’s work. And this can be true even if what is said is exactly the same. Though we have argued that consciousness is not necessary for responsibility, this does not entail that it is irrelevant. On the contrary, consciousness plays an important and morally significant role in action production, though its role is nonetheless indirect. It is by no means obvious that consciousness always increases one’s degree of responsibility for an action, however. One sort of exception might be cases where one’s careful

18 

What count as appropriate sanctions here will depend on further moral (and nonmoral) considerations apart from just her blameworthiness.

Responsibility and Consciousness    463 and calculated conscious decision serves to hide one’s true feelings, which might be elicited in spontaneous, non-​deliberative action. Another might include cases of willful ignorance, where one purposely blinds oneself to potentially relevant considerations. In the former, one might be more blameworthy for the unconscious action, and in the latter there may be no difference. (Whether or not the latter involves “tracing,” which we explicitly left to the side, is contentious; see King 2017; Sarch 2016.)

7.  Moving Forward If unconscious evaluation and decision-​making are prevalent and relevant in the ways we have indicated here, it suggests possible implications for other topics. Here we pursue just one: the traditional understanding of mens rea in criminal liability. To be criminally liable requires performing prohibited conduct (actus reus) with a culpable mental state (mens rea). It isn’t enough to kill someone, say. One must do so purposely, knowingly, recklessly, or negligently.19 Traditionally, recklessness is understood as the minimum standard of culpability for serious offenses, and conscious awareness marks the dividing line between it and negligence.20 In cases of both recklessness and negligence, the defendant’s conduct poses a substantial and unjustifiable risk. What distinguishes recklessness from negligence is that the reckless agent is aware of that risk, but disregards it, whereas the negligent agent is not aware of the risk (but should have been). The emphasis on conscious awareness is meant to support and justify the grades of mens rea. The involvement of consciousness is required for serious offenses because it tracks the degree of culpability. A defendant who consciously tries to kill the victim is more culpable than one who is only conscious of the fact that something he is doing for other reasons will kill the victim, who is more culpable than one who is only conscious of ignoring a substantial risk that he will kill the victim. And all of these defendants are more culpable than one who isn’t aware that he risks killing the victim, though he should have been so aware. Thus, conscious awareness is thought to track culpability, which in turn helps underwrite the use of graded mens rea categories for distinguishing between more and less serious offenses. If unconscious agential evaluation is possible, however, then the use of consciousness to mark this divide appears misguided. Consider a very basic legal requirement to respect the legitimate interests of others. I am obligated to exercise a degree of care in my actions, so as not to risk serious harm to those around me. Recklessness, then, occurs when I consider how my actions might affect the interests of others in ways that 19  Here we follow the Model Penal Code’s gradations of culpability (MPC, sec. 2.02), which sought to improve on the common law’s mental state categories. See American Legal Institute 1985. 20  The traditional understanding in common law also indicates that conscious awareness marks a divide in terms of “malice.”

464    Matt King and Peter Carruthers ought to constrain my action, but I proceed with the action anyway. We submit that it is an open question whether such disregarding must be conscious. For it may be that the knowledge that one’s action will risk harm to others is activated, but then deemed not relevant enough to one’s current concerns to attract attention and become conscious. This case strikes us as falling somewhere between the traditional categories of recklessness and negligence. It may not be as culpable as a case where someone is consciously aware of the risk to others, but dismisses it. It is surely more culpable, however, than a case where knowledge of the risk to others is never activated at all. Unsurprisingly, then, consciousness may be important though not necessary for serious criminal liability (for which recklessness generally serves as the minimum standard), just as it is for moral responsibility.

References Alloway, T., and Alloway, R. (Eds.). (2013). Working Memory: The Connected Intelligence. New York: Psychology Press. American Legal Institute. (1985). Model Penal Code: Official Draft and Revised Comments, Philadelphia, PA. Arpaly, Nomy (2002). Unprincipled Virtue: An Inquiry Into Moral Agency. Oxford University Press. Awh, E., Belopolsky, A., and Theeuwes, J. (2012). Top-​down versus bottom-​up attentional control: A failed theoretical dichotomy. Trends in Cognitive Sciences 16, 437–​443. Baars, Bernard J. (1988). A Cognitive Theory of Consciousness. Cambridge University Press. Baars, Bernard J. (2002). The conscious access hypothesis: Origins and recent evidence. Trends in Cognitive Sciences 6 (1):47–​52. Baddeley, A. (2003). Working memory: Looking back and looking forward. Nature Reviews Neuroscience 4, 829–​839. Banaji, M., and Greenwald, A. (2013). Blindspot: Hidden Biases of Good People. New York: Delcorte Press. Bassetti, C., Vella, S., Donati, F., Wielepp, P., and Weder, B. (2000). SPECT during sleepwalking. Lancet 356, 484–​485. Block, N. (1995). A confusion about the function of consciousness. Behavioral and Brain Sciences 18, 227–​247. Broughton, R., Billings, R., Cartwright, R., Doucette, D., Edmeads, J., Edwardh, M., Ervin, F., Orchard, B., Hill, R., and Turrell, G. (1994). Homicidal somnambulism: A case report. Sleep 17, 253–​264. Carruthers, P. (2015). The Centered Mind: What the Science of Working Memory Shows us About the Nature of Human Thought. Oxford: Oxford University Press. Carruthers, P. (2018). Implicit versus explicit attitudes: Differing manifestations of the same representational structures? Review of Philosophy and Psychology 9, 51–​72. Caruso, G. (2015). If consciousness is necessary for moral responsibility, then people are less responsible than we think. Journal of Consciousness Studies 22, 49–​60. Corbetta, M., and Shulman, G. (2002). Control of goal-​directed and stimulus-​driven attention in the brain. Nature Reviews Neuroscience 3, 201–​215.

Responsibility and Consciousness    465 Corbetta, M., Patel, G., and Shulman, G. (2008). The reorienting system of the human brain: From environment to theory of mind. Neuron 58, 306–​324. Dehaene, S. (2014). Consciousness and the Brain: Deciphering How the Brain Codes Our Thoughts. New York: Viking Press. Fischer, J.M., and Ravizza, M. (1998). Responsibility and Control: A Theory of Moral Responsibility. Cambridge: Cambridge University Press. Frankfurt, H. (1971). Freedom of the will and the concept of a person. Journal of Philosophy 68, 5–​20. Fox, K., Spring, R.N., Ellamil, M., Andrews-​Hanna, J., and Christoff, K. (2015). The wandering brain: Meta-​analysis of functional neuroimaging studies of mind-​wandering and related spontaneous thought processes. NeuroImage 111, 611–​621. Gawronski, B., and Bodenhausen, G. (2006). Associative and propositional processes in evaluation: An integrative review of implicit and explicit attitude change. Psychological Bulletin 132, 692–​731. Gazzaley, A. (2011). Influence of early attentional modulation on working memory. Neuropsychologia 49, 1410–​1424. Goodale, M. (2014). How (and why) the visual control of action differs from visual perception. Proceedings of the Royal Society B 281, 20140337. Greenwald, A. G., Banaji, M. R., & Nosek, B. A. (2015). Statistically small effects of the Implicit Association Test can have societally large effects. Journal of personality and social psychology, 108(4), 553–​561. Herdova, M. (2016). What you don’t know can hurt you: Situationism, conscious awareness, and control. Journal of Cognition and Neuroethics 4, 45–​7 1. Hodgson, D. (2012). Rationality + Consciousness = Free Will. Oxford: Oxford University Press. Hofmann, W., Gawronski, B., Gschwendner, T., Le, H., and Schmitt, M. (2005). A meta-​analysis on the correlation between the Implicit Association Test and explicit self-​report measures. Personality and Social Psychology Bulletin 31, 1369–​1385. Holroyd, J. (2012). Responsibility for implicit bias. Journal of Social Philosophy 43, 274–​306. Januszko, P., Niemcewicz, S., Gajda, T., Wolynczyk-​Gmaj, D., Justyna, P., Gmaj, B., Piotrowski, T., and Szelenberger, W. (2016). Sleepwalking episodes are preceded by arousal-​related activation in the cingulate motor area: EEG current density imaging. Clinical Neurophysiology 127, 530–​536. Kane, R. (1996). The Significance of Free Will. New York: Oxford University Press. King, M. (2009). The problem with negligence. Social Theory and Practice 35, 577–​595. King, M. (2014). Traction without tracing: A (partial) solution for control-​based accounts of moral responsibility. European Journal of Philosophy 22, 463–​482. King, M. (2017). Tracing the epistemic condition. In P. Robichaud and J.W. Wieland. (Eds.), Responsibility: The Epistemic Condition. Oxford: Oxford University Press, 266–​280. King, M., and Carruthers, P. (2012). Moral responsibility and consciousness. Journal of Moral Philosophy 9, 200–​228. King, M., and May, J. (2018). Moral responsibility and mental illness: A call for nuance. Neuroethics 11, 11–​22. Kurzban, R., Duckworth, A., Kable, J., and Myers, J. (2013). An opportunity cost model of subjective effort and task performance. Behavioral and Brain Sciences 36, 661–​679. Leslie, S-​J. (2015). The original sin of cognition: Fear, prejudice, and generalization. The Journal of Philosophy 114, 393–​421.

466    Matt King and Peter Carruthers Leslie, S.J., and Gelman, S. (2012). Quantified statements are recalled as generics. Cognitive Psychology 64, 186–​214. Libet, Benjamin W. (1993). The neural time factor in conscious and unconscious events. In G. R. Bock & James L. Marsh (eds.), Experimental and Theoretical Studies of Consciousness. (Ciba Foundation Symposium 174). pp. 174–​123. Libet, B. (1999). Do we have free will? Journal of Consciousness Studies 6, 47–​57. Levy, N. (2014). Consciousness and Moral Responsibility. Oxford: Oxford University Press. Mavda, A. (2018). Implicit bias, moods, and moral responsibility. Pacific Philosophical Quarterly 99, 53–​78. Mele, A. (1995). Autonomous Agents. New York: Oxford University Press. Menon, V., and Uddin, L. (2010). Saliency, switching, attention and control: A network model of insula function. Brain Structure and Function 214, 655–​667. Milner, A., and Goodale, M. (1995). The Visual Brain in Action. Oxford: Oxford University Press. Miyake, A., and Shah, P. (Eds.) (1999). Models of Working Memory. Cambridge: Cambridge University Press. Oswald, F. L., Mitchell, G., Blanton, H., Jaccard, J., & Tetlock, P. E. (2013). Predicting ethnic and racial discrimination: A meta-​analysis of IAT criterion studies. Journal of Personality and Social Psychology, 105(2), 171–​192. Pereboom, D. (2016). Omissions and different senses of responsibility. In A. Buckareff, C. Moya, and S. Rosell (Eds.), Agency and Moral Responsibility. Palgrave-​Macmillan, 179–​191. Plazzi, G., Vetrugno, R., Provini, F., and Montagna, P. (2005). Sleepwalking and other ambulatory behaviors during sleep. Neurological Science 26, 193–​198. R v Parks. (1992). 2 S.C.R. 871. Sarch, A. (2016). Equal culpability and the scope of the willful ignorance doctrine. Legal Theory 22, 276–​311. Schlosser, M. (2013). Conscious will, reason-​responsiveness, and moral responsibility. Journal of Ethics 17, 205–​232. Shepherd, J. (2015). Consciousness, free will, and moral responsibility: Taking the folk seriously. Philosophical Psychology 28, 929–​946. Sher, G. (2009). Who Knew? Responsibility Without Awareness. Oxford: Oxford University Press. Shoemaker, D. (2015). Responsibility From the Margins. Oxford: Oxford University Press. Smith, A. (2005). Responsibility for attitudes: Activity and passivity in mental life. Ethics 115, 236–​271. Smith, A. (2012). Attributability, answerability, and accountability: In defense of a unified account. Ethics 122, 575–​589. Smith, H. (1983). Culpable ignorance. Philosophical Review 92, 543–​571. Smith, H. (2011). Non-​tracing cases of culpable ignorance. Criminal Law and Philosophy 5, 115–​146. Sripada, C. (2016). Self-​Expression: A deep self theory of moral responsibility. Philosophical Studies 173, 1203–​1232. Strawson, G. (1986). Freedom and Belief. Oxford: Oxford University Press. Tamber-​Rosenau, B., Esterman, M., Chiu, Y-​C., and Yantis, S. (2011). Cortical mechanisms of cognitive control for shifting attention in vision and working memory. Journal of Cognitive Neuroscience 23, 2905–​2919. Terzaghi, M., Sartori, I., Tassi, L., Rustoni, V., Proserpio, P., Lorusso, G., Manni, R., and Nobili, L. (2012). Dissociated local arousal states underlying essential clinical features of non-​rapid

Responsibility and Consciousness    467 eye movement arousal parasomnia: An intracerebral stero-​electroencephalographic study. Journal of Sleep Research 21, 502–​506. Unsworth, N., Fukuda, K., Awh, E., and Vogel, E. (2015). Working memory delay activity predicts individual differences in cognitive abilities. Journal of Cognitive Neuroscience 27, 853–​865. Wegner, D. (2002). The Illusion of Conscious Will. Cambridge, MA: MIT Press. Wieland, J.W., and Robichaud, P. (2017). Blame transfer. In P. Robichaud and J.W. Wieland (Eds.), Responsibility: The Epistemic Condition. Oxford: Oxford University Press, 281–​298. Williams, L.E. and Bargh, J. (2008). Experiencing physical warmth promotes interpersonal warmth. Science 24, 606–​607.

CHAPTER 23

Resp onsibi l i t y a nd Situati oni sm Brandon Warmke

1.  You and Your Situations: Three Pictures Here is a truism: Human action requires being in some situation or other. Cracking open a can of beans requires being around a can of beans. Here is another truism: Being morally responsible for what you do requires being in some situation or other. You cannot be morally praiseworthy for pulling a drowning child from a pond unless you are near a pond. You cannot be morally blameworthy for stealing your neighbor’s rake unless you are close enough to a rake. Situations play an ineliminable role in our being responsible for what we do. What kind of role? How should we picture the relationship between ourselves and the situations in which we act? Here is one picture: Situations are a stage. They provide the setting for our agency to unfold. Being temporally and spatially located near a can of beans enables me to open it. Furthermore, our situations provide reasons for action. A child drowning nearby gives me reason to act. This first picture is not true to life. Our situations are not mere stages because they aren’t always so cooperative. Features of our environment sometimes undermine our ability to act freely and responsibly. For one thing, our situations often include other agents who deceive us, coerce us, and manipulate us. Situations also conspire against us “naturally,” without the intervention of other agents. You act on an optical illusion. A hallucinogenic herb falls from a tree into your coffee while camping. While piloting a plane, you suffer from hypoxia, a condition in which your brain becomes deprived of sufficient oxygen, resulting in a deterioration in mental clarity. To correct course, you need the ability to recognize that something has gone wrong, an ability itself compromised by

Responsibility and Situationism    469 your hypoxia.1 Our situations are therefore not as benign as the first picture suggests. According to a second picture, our situations are occasional traps. Generally, we are free and responsible, but sometimes our situations trip us up. While this second picture may not be the happiest of arrangements, it is some consolation that these impediments to free and responsible agency are few and far between. Or perhaps not. In recent decades, the so-​called situationist research program in psychology and philosophy has revealed an uncomfortable fact about us: Situations affect our behavior in surprising and powerful ways. Initially, these “situationists” used empirical studies in psychology to argue that human behavior is highly sensitive to apparently insignificant situational features. Therefore, some concluded, we do not have robust moral character traits.2 The basic idea was that whether you do the honest or compassionate thing has little if anything to do with whether you are an honest or compassionate person, and much to do with what kinds of situation you happen to encounter.3 We lack stable moral character traits that we “carry around” with us across situations. Instead, these situationists claimed, our behavior is highly dependent on how we are affected by situational features. To see why they thought this, let’s look at a few classic studies from the situationist literature. In the most famous of Milgram’s “shock” experiments, researchers told subjects they were testing a teaching strategy: that learning is most effective when incorrect responses are punished (1969). Subjects were asked to administer shocks when the “learner” (a confederate) gave the wrong answer to a question. (No real shocks were administered, though the subjects didn’t know this.) Starting with 15V, each successive incorrect answer was to be greeted with an increased shock. Prerecorded reactions to the shocks ranged from a surprised “Ouch!” to screaming and banging on the wall. At 330V, the confederate ceased to respond. Sixty-​five percent (26 of 40) of subjects administered shocks up through 450V, long after the confederate stopped responding. Many normal people will apparently be willing to harm others significantly when put in the right situation. In one part of Darley and Batson’s 1973 “Good Samaritan” study, seminary students were asked to walk across campus to deliver a talk on Jesus’s parable of the Good Samaritan.4 In their path, they encountered a man slumped in an alleyway (a 1 

Ballantyne (2015).

2 The locus classicus is Ross and Nisbett’s The Person and the Situation (1991), although much of the

“situationist” empirical research had been going on for decades. Flanagan (1991), Doris (1998, 2002), Harman (1999), and Merritt (2000) are credited with bringing this empirical work to the attention of philosophers. 3  I say “the basic idea,” but there are many different claims going by the name “situationism” as regards skepticism about character traits. I will not rehearse them here, but see Miller (2014: ch. 4) and McKenna and Warmke (2017) for a review. Hence it is helpful to distinguish, as Nelkin (2005) does, “situationism” as some specific empirical or philosophical thesis, from the “situationist literature,” the empirical studies themselves. 4  In this parable, a Jewish traveler is stripped of his clothing, beaten, and left to die on the side of the road. Both a priest and a Levite (likely an assistant to a priest) pass by before a Samaritan man

470   Brandon Warmke confederate). Overall, 40% offered to help. However, whether someone helped appeared to depend largely on the extent to which they were told to hurry to give the talk. In “low” hurry situations, 63% helped; in “medium” hurry situations, 45% helped; and in “high” hurry situations, only 10% stopped to help. Being asked to hurry apparently makes a big difference as to whether you stop to help on the way to give a talk about stopping to help. In Latané and Rodin’s 1969 “Lady in Distress” experiment, participants in a room were tested to see if they would help someone next door upon hearing a loud crash and screaming in the next room. (“Oh, my god, my foot . . . I . . . I . . . can’t move it. Oh my ankle, I can’t get this thing off me!” the confederate moaned and cried for a minute, gradually getting more subdued.) Seventy percent of those who had been in the room alone helped. Only 7% helped if they had been in the room with a confederate who was also not helping (cf. Latané and Darley 1968). Apparently, the mere presence of another person who doesn’t help leads us to not help, too. In Isen and Levin’s 1972 study of mood effects, those who found a dime in the return slot of a phone booth were much more likely to help a confederate pick up dropped papers (88%) than those who had not found a dime (4%).5 In another study of such mood effects, Isen and Reeve (2005) found that subjects who were given a $2 box of chocolates worked more quickly and accurately on a task than those who didn’t. They also reported enjoying the task more. Apparently, little mood boosts lead many to do acts of kindness they wouldn’t have otherwise done. Matthews and Canon (1975) concluded that ambient noise can affect helping behavior. While a subject sat in a waiting room, a confederate would drop a stack of books and papers. With 48db of normal background noise, 72% helped. When “white noise” was introduced, helping behavior decreased. At 65db of white noise, 67% helped. When the white noise was increased to 85db, 37% helped. They ran a similar study outdoors with a confederate who, while wearing a cast, dropped a stack of books getting out of a car. Eighty percent of subjects stopped to help with normal background noise. Sixteen percent stopped to help with a loud lawnmower running in the background. Apparently, the sound of a lawnmower makes us less helpful. Baron (1997) found that you could affect people’s behavior by changing the smells in the air. When asked to make change for a dollar bill, 22% of males and 17% of females did so when passing a clothing store. When passing a Cinnabon (cinnamon rolls) or Mrs. (Samaritans were an ethnoreligious group widely despised by Jews) stops to help the traveler, bandage his wounds, and pays for him to convalesce at an inn (Luke 10:30–​37). 5 

As is commonly pointed out, there were replication problems with this study, not to mention its small sample size (24 females, 17 males). But as Miller (2013: ch. 3) explains, other studies of mood effects vindicate their power. I should also say that for present purposes, I will set aside worries about the so-​ called replication crisis in psychological research. I do this for two reasons. First, whatever the status of individual studies, the broad “situationist” research program, buttressed by many kinds of studies, is robust. Psychologist Matthew Lieberman writes, “If a social psychologist was going to be marooned on a deserted island and could only take one principle of social psychology with him it would undoubtedly be ‘the power of the situation’ ” (2005: 746). Second, we can abstract from the specific empirical evidence and still ask a meaningful philosophical question: What kinds of situational influences on our agential capacities would undermine or degrade our capacities for free and morally responsible action?

Responsibility and Situationism    471 Fields (cookies), however, rates of helping behavior shot up: 45% of males and 61% of females made change.6 The smell of sweet treats apparently makes us more helpful. The results are surprising.7 Why would something as morally insignificant as finding a dime, or being asked to hurry, or being in a room with someone else make such a large impact on whether we do right the right thing? As John Doris and Dominic Murphy put it, our situations apparently “affect how circumstances are observed, interpreted, and evaluated, thereby powerfully affecting our moral behavior” (2007: 36). As I mentioned before, the conclusions initially drawn from this empirical literature were, to varying degrees, forms of skepticism about the existence or efficacy of moral character traits. Surely, a truly compassionate or honest person would do the compassionate or honest thing across situations, not just when the smells and sounds were right. Because our behavior is so inconsistent across situations, it was alleged that “global” or “cross-​situationally stable” character traits are rare, if they exist at all. This “first-​wave” situationism relied on the empirical research to raise doubts about the existence of moral character traits.8 More recently (though these “waves” have some chronological overlap), philosophers have again drawn upon these studies to challenge another common way of thinking of ourselves. Whereas first-​wave situationism raised doubts about the existence of moral character traits, second-​wave situationism draws upon the situationist literature to motivate skepticism about free and morally responsible agency.9 The second-​wave challenge, which I develop in more detail later, alleges that if our behavior can be so easily “pushed around” by morally irrelevant situational features, then we often do not act freely and therefore are often not morally responsible for what we do.10 So we have: First-​wave situationism: challenges the manifest image of morally mature adults as having robust moral character traits. 6 

For further review, see Miller (2014: ch. 4). Hence Miller’s (2017) use of the term “surprising dispositions” to describe the experimental results: We are disposed to respond to certain situational features with certain beliefs, desires, and emotions in ways that have an unexpected and significant impact on our behavior. As Miller notes, the observation that these results are surprising is a common one: Ross and Nisbett (1991: 46); Flanagan (1991: 292); Doris (2002: 63, fn. 5); Nahmias (2007: 4); Russell (2009: 253, 277). 8  The literature is vast, but for starters, see discussions in Flanagan (1991), Kupperman (2001), Sreenivasan (2002), Annas (2003), Miller (2003; 2014), Montmarquet (2003), Kamtekar (2004), Tucker (2004), Sabini and Silver (2005), Vranas (2005), Webber (2006), Wielenberg (2006), Russell (2009), Sarkissian (2010), Slingerland (2011), Alfano (2013), and Rodgers and Warmke (2015). 9  See discussions in Doris (2002), Nelkin (2005), Doris and Murphy (2007), Nahmias (2007), Talbert (2009), Brink (2013), Ciurria (2013), Mele and Shepherd (2013), Vargas (2013b), Herdova and Kearns (2015, 2017), McKenna and Warmke (2017), Miller (2017), Sartorio (2018), and Talbert and Wolfendale (2019). 10  For the record, let’s define some terms. By “free will” I mean the ability of a person to control her conduct in the strongest sense necessary for moral responsibility (Mele 2006). This freedom condition is one necessary condition on morally responsible agency (the other usually called the “epistemic condition”), an exercise of which is the gateway to moral responsibility for conduct (McKenna 2012). The kind of moral responsibility for conduct I have in mind involves basic desert: “For an agent to be morally responsible for an action in this sense is for it to be hers in a way that she would deserve to be blamed if she understood that it was morally wrong” (Pereboom 2014: 2). Desert is basic in that “the agent would deserve to be blamed . . . just because she has performed the action, given an understanding of its moral 7 

472   Brandon Warmke Second-​wave situationism: challenges the manifest image of morally mature adults as typically acting freely and responsibly.

To get an initial grip on the situationist threat to moral responsibility, consider Leigh, who stops to make change for someone at the mall. You might have thought that she saw she could help someone and consequently freely and responsibly did so. But then you learn not only that Leigh was in front of a Cinnabon store when this happened, but that studies show people are much more likely to make change near a Cinnabon. It is likely that she wouldn’t have helped if she had been in front of J. Crew. Leigh’s helping behavior appears to be in no small measure due to the smells in the air, a fact about her environment that, let us suppose, she didn’t even notice. The smell of sweet treats gave her a “boost” to do a good thing. And just like situational features may provide boosts to do the right thing, they may also “boost” us in the other direction. Add white noise to the mix and voilà, she will be less likely to help. This is unsettling. If ambient smells and sounds secretly push us around, who is really in charge? Just how much of the springs of action are due to situational features affecting us in certain ways? It is especially worrying that the boosts happen below the level of awareness. If the smell of Mrs. Fields “moves” us to help others, this is not because we consciously act compassionately because of the smell in the air. We would not cite the smell of sweet treats as part of our reason for helping. In fact, we would likely deny that this played any role in our behavior at all. Furthermore, these situational features are not the kinds of things that make a difference as to what we should do. The smell of cookies is not a reason to help. White noise is not a reason not to help. These situational features are, as Carolina Sartorio puts it, “external factors of the environment that don’t make a difference to what we should do in the circumstances, but that still tend to have an effect on what we actually do” (2018: 796). Now recall the question with which we began: How should we understand our relationship to our situations? We saw that picture one—​situations as stage—​is not true to life. The situationist literature now casts doubt on the accuracy of picture two, the view that our situations are, except for occasional “traps,” benign. Much more might be happening “below the surface” than we realize. Perhaps so much is going on behind the scenes that we lack the necessary control over our behavior to be morally responsible for it. Perfectly ordinary features of our world might conspire against us—​either constantly or intermittently—​to prevent us from exercising the agency required to be morally responsible for our conduct. We are then faced with the possibility that we must accept a third picture of our situations: that our morally responsible agency is compromised much more often than we had thought, by perfectly “innocent” and ubiquitous features of life: our normal environments. On this third picture, our situations are perpetual threats. The very

status, and not, for example, merely by virtue of consequentialist or contractualist considerations” (Pereboom 2014: 2).

Responsibility and Situationism    473 things—​our situations—​that enable us to act may also threaten our ability to act as morally responsible agents.

2.  What Is the Threat? Situational features affect us in ways that appear to threaten our ability to be morally responsible for our conduct. But what exactly is the nature of the threat? This is not obvious. The empirical results are no doubt surprising. But this fact itself does not, at least in any straightforward way, impugn morally responsible agency. In this section, I explore how we should understand the claim that the situationist literature threatens moral responsibility. One way to do this is to see whether the situationist threat reduces to some other alleged threat to moral responsibility.11 Let’s consider a few candidates.

2.1  Determinism Incompatibilists think that determinism threatens moral responsibility. If our actions are the necessary result of the laws of nature and the facts of the past—​facts outside of our control—​then we cannot be responsible for our behavior because it too is out of our control. Perhaps situational features undermine responsibility by determining our behavior. On this construal of the threat of situationism, the situationist threat is just a specific instance of the threat of determinism.12 These studies reveal that situational features like the smell of cookies affect us by determining how we respond to them. And since determinism is a threat to moral responsibility, these situational features, when they affect us deterministically, also threaten moral responsibility. We can put the situationist threat as follows: 1. The situationist literature shows that situational features outside our control determine (much of) our behavior.13

11  Nelkin (2005) explores the possibility that the situationist threat reduces to issues regarding (1) characterological views of freedom and responsibility; (2) the fundamental attribution error; (3) determinism; (4) weakness of will; and (5) real self views of freedom and responsibility. In my view, Nelkin persuasively argues that none of these accurately captures the situationist threat. I will discuss only one of these—​determinism—​in addition to two new possible explanations. Miller (2017) also considers characterological views and determinism and rejects them as ways to understand the situationist threat. 12  For present purposes, I understand determinism to be the thesis that at any time, the universe has only one possible physical future. 13  Nelkin (2005) rightly notes that one need not think that situational features affect everyone uniformly to think that situational features undermine freedom and responsibility due to their determinative character (cf. Miller 2017). Nor would one need to think that such features operate on

474   Brandon Warmke 2. Behavior that is determined by factors outside our control is not free and we cannot be responsible for it. 3. Therefore, the situationist literature shows that we are not free and responsible for (much of) our behavior. On this way of framing the situationist threat, perfectly ordinary features of the world outside our control determine much of our behavior, even though we have no idea this is going on. This is reminiscent of Spinoza’s reason for denying free will and responsibility, more recently developed by Derk Pereboom (2001, 2014). As Spinoza put it, “men believe themselves to be free, because they are conscious of their own actions and are ignorant of the causes by which they are determined” (1766/​1985: 496). The situationist literature, it is alleged, vindicates Spinoza. However, two considerations count against reducing the situationist threat to the threat of determinism. First, situational features need not be taken to affect us deterministically. Perhaps they affect us this way. But this conclusion is not required by the current empirical evidence. The evidence does not show that agents in these studies could not have done otherwise. If situational features affect us indeterministically, the threat cannot be determinism. Second, notice that even though compatibilists do not think determinism as such threatens freedom and responsibility, this does not mean they must think that any way our behavior could be determined is compatible with determinism. In other words, the compatibilist is not committed to the claim that any deterministic production of action counts as free and responsible. Compatibilists typically regard action resulting from coercion, compulsion, or ignorance as unfree, even if such action is produced deterministically. So even if we assume situational features affect us deterministically, there may be some fact about how situational features affect us that threatens free and responsible agency. The threat would not be determinism as such. The threat could therefore be in principle something recognized by both compatibilists and incompatibilists. Tentatively, then, we can conclude that if the present situationist literature reveals a threat to free and responsible agency, that threat is not simply determinism.

2.2  Coercion or Manipulation Perhaps the situationist literature reveals that we are often manipulated or coerced to act. Consider David Brink’s (2013) notion of “situational control.” To have situational control is to be free from a kind of manipulation or coercion that would leave one without any everyone deterministically to get the deterministic worry off the ground. Suppose that some but not all reactions to situational features are brought about deterministically. Knowing this might be enough to undermine our confidence that, in any given situation, we are acting freely (and so not deterministically, if being non-​derivatively free and responsible for x-​ing precludes x-​ing being deterministically caused by factors outside one’s control).

Responsibility and Situationism    475 reasonable alternative courses of action. Although a lack of situational control does not itself compromise one’s status as a morally responsible agent, it does provide an excuse that challenges whether one is responsible for what one does. Perhaps the way to think about the situationist threat, then, is as revealing that we often lack situational control: Situational features manipulate or coerce us so that we lack any reasonable alternatives. But as Brink himself notes, none of the situationist studies involve coercion or threats that leave an agent without a reasonable alternative. The closest are the requests of the confederates in the Milgram studies. But even here, there is no reasonable interpretation of coercion such that the requests of the Milgram confederates count as coercive. Further, their requests certainly left open reasonable and permissible alternatives. Even if the Milgram subjects were not coerced, might they still have been manipulated? We can think of coercion and manipulation as external forces that lie along a continuum.14 Like coercion, manipulation undermines moral responsibility for conduct. As Allen Wood puts it, manipulation “influences people’s choices in ways that circumvent or subvert their rational decision-​making processes, and that undermine or disrupt the ways of choosing that they would themselves would critically endorse if they considered the matter in a way that is lucid and free of error” (2014: 35). Does the threat of situationism reduce to the threat of manipulation? You might think so. Manipulation often involves various forms of deception, pressures to acquiesce, and playing upon emotions, emotional needs, or weaknesses of character (Baron 2003). The Milgram experiments involve all these. Insofar as manipulative influence can undermine free and morally responsible agency, perhaps the threat of situationism is just the threat of manipulation. So consider: 1. The situationist literature shows that (much of) our behavior is the result of manipulative situational influences. 2. Behavior that is the result of manipulative situational influences is not free and we cannot be responsible for it. 3. Therefore, the situationist literature shows that we are not free and are not responsible for (much of) our behavior. There are at least two problems with this reduction of the situationist threat. The first concerns premise 1: Not all situationist studies involve manipulation, at least if by “manipulation” we mean something done by agents. Though it’s true that the situational features in the experiments are designed by agents, there is nothing about those designs that is inherently artificial—​qualitatively similar situations can and do arise in

14  Allen Wood observes: “Being manipulated into doing something is different from being coerced into doing it. The two seem to me to form a kind of continuum, with manipulation occupying the subtler end and coercion occupying the more heavy-​handed end” (2014: 31). The notion of manipulation here differs from that commonly employed in the free will and responsibility literature, where moment by moment control by an intervener can count as manipulation (Pereboom 2001). Wood is working with a more colloquial sense of the term, as will I.

476   Brandon Warmke non-​experimental settings. The second and more serious problem concerns premise 2. Recall Wood’s remark that manipulation characteristically involves circumventing or subverting one’s rational decision-​making processes and undermining or disrupting the ways of choosing. This suggests that the situationist threat is something more fundamental than manipulation as such: Situational features, whether the result of manipulative interference or not, appear to undermine or degrade our ability to make decisions in accord with the good reasons there are. We will return to this thought later.

2.3  Luck One situationist study found that subjects were twice as likely to help deliver documents to someone 40 meters away if the request was made after the subject exited a public bathroom (Cann and Blackwelder 1984: 224). It might appear, then, that whether someone engages in helping behavior is largely a matter of luck. Helping appears highly correlated with whether you were asked to help after leaving the bathroom. Whether you are asked to help after leaving the bathroom is a matter of luck. How, then, could someone be morally praiseworthy or blameworthy for helping when so much depends on whether you had just left the bathroom? Whether you help is largely, if not wholly, a matter of situational luck. To make matters worse, situational luck, much like situations themselves, is impossible to eliminate. As Michael Zimmerman puts the point: One is never in complete control of the situations that one faces, either with respect to “external” matters such as being born, being of a certain physical constitution, being distracted by a loud noise, being in a certain geographical location, and so on, or with respect to “internal” matters such as being irascible, suffering from an Oedipus complex, having a kindly disposition, and so on. And all of these matters affect what one does. It is against them as a background that one makes the decisions that one does; indeed, without such a background, no decisions could be made. (1987: 384)

So consider: 1. The situationist literature shows that (much of) our behavior is the result of situational luck. 2. Behavior that is the result of situational luck is not free and therefore we cannot be responsible for it. 3. Therefore, the situationist literature shows that we are not free and responsible for (much of) our behavior. However, there are problems with this reduction, too. Even if we grant premise 1, premise 2 is too quick. While it may be true that whether one is faced with a certain situation is often largely a matter of luck, this does not straightforwardly entail that one’s action in

Responsibility and Situationism    477 such a situation is not an exercise of free and responsible agency. Zimmerman himself makes this point, for he follows up the above passage immediately by noting that: Nevertheless, as long as the decision, for example, to collaborate [in wrongdoing] is made freely, then one is surely, ceteris paribus, to blame for such collaboration. (1987: 384)

What we want to know, then, is whether subjects in situationist studies (and their counterparts in real-​life situations) are acting freely and responsibly. The mere fact of situational luck does not obviously settle that question.

3.  Situationism and Reasons-​R esponsiveness I have challenged the claim that the situationist threat to moral responsibility can be reduced to the threat of determinism, coercion and manipulation, or luck. How, then, are we to understand it? To be frank, I am not entirely sure. One response is to conclude that there is no situationist threat and move along with our day. Or we might explore the thought that there is no single way of reducing the situationist threat. If situational features threaten free will and responsibility, they do so in a diverse and piecemeal way: sometimes via determinism, sometimes via luck, and so on, perhaps even in combination with one another. In the rest of this chapter, however, I will continue to proceed on the assumption that situational features pose a threat to free and responsible agency, and that this threat can be reduced to a more fundamental threat. Like others who have written on this issue, I understand the situationist threat as targeting our capacities of reasons-​responsiveness.15 So first let me say a few words about what reasons-​ responsiveness is and its importance for free and responsible agency. Then we will move on to see how situational features might threaten our ability to respond to reasons.16

3.1  Reasons-​Responsiveness: The Basics Many philosophers writing on free will and moral responsibility account for freedom at least partly in terms of an agent’s responsiveness to reasons.17 That is, to be free and morally responsible requires an ability to respond to reasons in certain kinds of ways. 15 

See, e.g., Nelkin (2005), Doris and Murphy (2007), Herdova and Kearns (2017), Miller (2017), McKenna and Warmke (2017), Sartorio (2018). 16  In this section, I draw from McKenna and Warmke (2017). 17  See, for example, Gert and Duggan (1979), Dennett (1984), Wolf (1990), Fischer and Ravizza (1998), Haji (1998), Nelkin (2011), Brink and Nelkin (2013), McKenna (2013), Vargas (2013a), and Sartorio (2016).

478   Brandon Warmke Crucially, on most reasons-​responsive theories, for an agent to be morally responsible for what she does, she must, in some sense of “able,” be able to respond to specifically moral reasons (Wolf 1990; Fischer and Ravizza 1998; and Nelkin 2011). Because many theorists understand free will in terms of the control condition(s) necessary for moral responsibility, reasons-​responsiveness theories are also frequently framed in terms of theories of (at least part of) the control condition for moral responsibility. Therefore, if situational features undermine or degrade one’s ability to be reasons-​responsive, then, according to these views, those features would also undermine or degrade one’s freedom and responsibility. What does reasons-​responsiveness require? This is a bit contentious, but many theorists agree that reasons-​responsiveness has two components: one cognitive, one volitional (Fischer and Ravizza 1998; Brink and Nelkin 2013; and McKenna 2013). One way to flesh this out is to say, as Fischer and Ravizza (1998) put it, that reasons-​responsiveness has a receptivity component and a reactivity component.18 An agent must first be able to recognize what good reasons there are and assess them for whether they are sufficient for pursing a course of action. This is the cognitive dimension whereby reasons are “received.” We are, for example, inclined to excuse an agent who cannot distinguish right from wrong, or who cannot recognize sufficient moral reasons for action at all. One’s cognitive abilities to recognize such reasons may simply be underdeveloped (as in very young children) or degraded (temporarily or permanently) because one is suffering from a delusional disorder. The second, volitional, component says that an agent must also be able to guide her actions in light of her recognition of the good reasons that there are. This is the dimension whereby an agent reacts to the reasons after she receives them. Irresistible desires, paralyzing fears, severe clinical depression, and certain kinds of brain damage are examples of obstacles that can eliminate or severely impair one’s ability to conform one’s behavior to one’s judgments (Brink 2013). Again, in these cases we are inclined not to hold such agents morally responsible for their conduct. I have described the situationist threat as potentially “undermining or degrading” one’s free and responsible agency. This is because one’s morally responsible agency—​ and therefore one’s ability to respond to good moral reasons—​is not an all or nothing affair. To have one’s reasons-​responsive capacities fully undermined would render one wholly incapable of responding to moral reasons. Such a person is not free and responsible. But one might instead retain the ability to be receptive and reactive to some good moral reasons while still falling below the threshold of responsiveness required for moral responsibility. Although not fully undermined, one’s capacities would still be severely impaired. A person rendered unfree and not responsible for an act of compulsive handwashing might still be responsive to some range of reasons while being unresponsive to many good reasons for not handwashing. If, for instance, she was told not to

18  Brink identifies similar components. On his view, moral responsibility requires both “cognitive capacities to distinguish right from wrong” as well as “volitional capacities to conform one’s conduct to that normative knowledge” (2013).

Responsibility and Situationism    479 wash her hands while being threatened at gun point, she might respond to this reason by refraining from washing. On the other hand, one might be receptive and reactive to many, but not all the good moral reasons in some situations, and still meet the requirements for free and responsible agency, whatever those happen to be. For example, a person who is praiseworthy for giving to a charity for starving children might be moved by many equally weighty moral reasons to give to different kinds of charities. And yet she might be blind to some similarly weighty moral reasons. Perhaps she doesn’t see helping victims of AIDS as a similarly weighty moral reason and so wouldn’t give to such charities. That she would not respond to all good (and sufficient) reasons to give to a worthy charity does not mean that she fails to act freely when she actually gives to a charity for starving children. The point is that one need not be able to be perfectly responsive to reasons to be morally responsible, and one need not be wholly unresponsive to reasons to fail to be morally responsible.19 What is required, then, on a reasons-​responsive account of freedom is the specification of a spectrum of responsiveness to a sufficiently rich range of reasons. Doing so establishes that the agent is a sane, morally competent person. What determines the relevant spectrum? This is a vexed question to which we will return. For now, we can simply say that there is some spectrum of relevant reasons to which one must be responsive to meet the threshold for free and responsible agency. We could say much more, but this is enough to see how situational features might undermine or degrade reasons-​responsiveness.20

3.2  The Threat to Reasons-​Responsiveness With the notion of reasons-​responsiveness in hand, let’s turn to how situational features might affect this capacity. Here is how Dana Nelkin frames the potential threat: One way of seeing the situationist cases—​or at least some of them—​as troubling is this: simply put, the subjects seem to be acting for bad reasons, or at least not acting for good reasons, and they seem stuck doing so. . . the way in which the subjects 19 

Some writers correlate degrees of reasons-​responsiveness with degrees of moral responsibility (at least once a certain threshold is met) (e.g., Brink 2013; Coates and Swenson 2013: 636; and Herdova and Kearns 2017: 3). But one need not do this. One can deny, as I do, that “being morally responsible for x” is something that can come in degrees. This does not entail denying one may be more or less deserving of blame or praise, or denying that one may deserve different kinds or blame or praise, or denying that one can be causally responsible for more or less of an action or a consequence. But this is perfectly compatible with thinking that one’s moral agency can come in degrees of quality or robustness or whatever. Being morally responsible for x “gets one in the game” of our moral responsibility practices as they concern x-​ ing. Then it can be determined what, if any, kind of praising or blaming response is appropriate. And at this latter state one’s degree of responsiveness is relevant. This “gateway” theory of moral responsibility is inspired by Fischer (2006: 233). 20  Fischer and Ravizza (1998) offer a specific proposal cast in terms of moderate reasons-​ responsiveness, MRR. Herdova and Kearns carefully discuss MRR with respect to the situationist threat in their (2017). For some reservations and qualifications about MRR in general, see McKenna (2013). I will pass over these more specific details.

480   Brandon Warmke seem to proceed raises a question about whether they can act for good reasons—​in some important sense of “can.” (2005: 199, 201)

Nelkin’s point is not just that subjects in these studies act for bad reasons—​this is nothing new and is not much of a reason for thinking one is not free and responsible. Rather, the worry is that the subjects appear to be “stuck”—​prevented from acting on the good moral reasons. Carolina Sartorio uses similar language: “the situationist threat arises from the fact that we can apparently be ‘gripped’ by some aspects of the actual circumstances (in many cases, without our even realizing this) in ways that can undermine our control and responsibility” (2018: 798). Even so, how exactly are situational features supposed to undermine one’s ability to respond to good reasons? How are we gripped or stuck? Here is one way: T1. A situational feature prevents you, in some way, from becoming aware of a fact that provides a reason to act.

Consider, for example, the Darley and Batson Good Samaritan study. Let us suppose that a sense of urgency does degrade certain of the seminarian’s capacities, such as the ability to notice certain things in her environment that she might otherwise notice. The seminarian therefore fails to recognize a person slumped in a doorway, seemingly in need of medical attention. (Of course, there may be lots of other things she doesn’t notice either, such as the marching band in the distance playing themes from Thelonious Monk.) According to this construal of the threat, situational features shield you from becoming aware of some fact that provides a moral reason to act in a particular way. Even if situational features can degrade one’s perceptive capacities in this way, this does not obviously threaten one’s reasons-​responsiveness. However, this is complicated. On the one hand, if you build into reason-​responsiveness certain capacities to perceive your environment, and if those capacities are undermined or degraded in scenarios like Good Samaritan, then there may be some reasons-​responsive-​relevant capacity that is impaired. On the other hand, reasons-​responsiveness should not require that one have the ability to be aware of all the facts that provide reasons in any given situation. It does not, I think, count against your reasons-​responsive capacities that you don’t dig the gold bar out of your bathroom wall because it is covered by drywall that prevents you from seeing it. Similarly, suppose that a large metal shield prevented you from seeing (or otherwise becoming aware of) a man slumped, hurting in a doorway. The fact that he is hurting is a reason to stop and help, but you are shielded from seeing this fact. It would be odd to see this as a threat to your free and responsible agency. Naturally, you are excused for not stopping to help, but this is not because your morally responsible agency has been undermined or degraded. It is because a situational feature prevented you from seeing a reason to help. Why, then, would your reasons-​responsiveness be threatened by other, less obvious kinds of situational “shields?” The point is that if situationist studies involve subjects who, like in Good Samaritan, do not help hurting strangers, they may not be responsible for failing to do so, but not

Responsibility and Situationism    481 because of any failure of reasons-​responsiveness, even though there is a sense in which they do not receive or react to the sufficient reasons there are.21 Rather, one plausible reason why they are not responsible for not helping is simply that they were not aware (non-​culpably, let us assume) of the fact that provided sufficient reason to help. But suppose situationist studies are threatening in a less innocent way. One way they might do so is by degrading our powers of receptivity. Consider: T2. A situational feature prevents you, in some way, from recognizing a fact (of which you are aware) as a reason to act.

On this construal of the threat, one is aware of the relevant reason-​providing fact, but cannot see it or register it as a reason. Recall that reasons-​responsive theorists typically require that an agent be able to recognize good moral reasons, or to see what is right or wrong. If situational features operated in the manner of T2, they would affect reasons-​ responsiveness by undermining or degrading one’s receptivity capabilities. Perhaps in the Milgram experiments, the “teachers” saw that the “students” were in pain, but due to situational pressures, they did not see this as reason to stop shocking the students. The second component of reasons-​responsiveness gives us another way of construing the situationist threat: T3. A situational feature prevents you, in some way, from reacting to a reason that you recognize.

On T3, one is aware of the fact that provides a reason and one recognizes this as a reason, but yet one is unable to translate this recognition into action. If situational features operated in this manner, they would affect reasons-​responsiveness by virtue of undermining or degrading one’s reactivity capabilities. In the Milgram studies, for example, it is not unlikely that the subjects saw that they should stop giving “shocks,” but that the situational pressures to obey authority undermined or degraded their capacities to translate that knowledge into action. There are other ways of construing the situationist threat to reasons-​responsiveness (and more precise ways of stating the ones we have discussed), but like life, this chapter is short.22 These will do for our purposes. Since 21  Nelkin (2005) make a similar observation. Here, it is tempting to say that such effects on our ability to recognize our surroundings are innocent and ungeneralizable as a threat to moral responsibility. While such effects on us can excuse a person by way of nonculpable ignorance, the excusing here is just a normal part of life. People are often excused for innocently not seeing something which, if they saw it, would give them reason to help another in need. In correspondence, however, Dana Nelkin asks that we suppose it turns out that situationist experiments show that such normal everyday excuses are just more widespread than we thought. If this is how things shake out, then we are responsible less often than many presume, but not because of any threat to morally responsible agency as such. Because I am presently focusing on situationist threats to reasons-​responsiveness, I set this possibility aside, but note that this is one other line of inquiry for the situationist to pursue. 22  For example, Herdova and Kearns (2017) suggest that situational features may cause agents to respond to non-​reasons (as opposed to disabling a capacity to respond to good ones).

482   Brandon Warmke T1 need not be understood as a threat to reasons-​responsiveness as such, we will focus our attention on T2 and T3.

4.  Responding to the Situationist Threat 4.1  The Modal Strategy We now have an initial grip on how situational features could undermine or degrade one’s responsiveness to reasons and therefore one’s free and responsible agency. We are “stuck” in these situations responding to bad reasons, or no reasons at all, because situational features (like the smell in the room or the presence of a confederate) prevent us from receiving (T2) or reacting (T3) to good moral reasons. Must we therefore concede that agents are generally unfree and so not responsible? Of those who have weighed in on the issue, many have argued that the evidence does not in fact show this. One common strategy for deflecting the situationist threat to reasons-​responsiveness has been defended by Brink (2013), Herdova and Kearns (2017), and McKenna and Warmke (2017). Call it the modal strategy. It proceeds in two steps.23 The first step brings to the fore a feature of many theories of reasons-​responsiveness that we have so far largely ignored: their modal dimension. In many cases, an agent’s acting from reasons-​responsive resources will involve her actually responding to pertinent reasons. If, for instance, she is blameworthy and acts wrongly from a motive of selfishness, this might involve her responding to a reason that some course of action will increase her wealth. Nevertheless, according to many reasons-​responsiveness theorists, it is not a requirement of an agent being reasons-​responsive in acting as she does that she in fact respond to any reasons at all. She might be acting from pure impulse or brute desire.24 Crucially, on most reasons-​responsive theories, for an agent to be morally responsible for what she does, she must be able, in some sense of “able,” to respond to moral reasons, regardless of whether she actually does so respond. In this way, a free agent will be able to respond to reasons even if, when she acts, she does not actually do so.25 An agent incapable of responding to moral reasons is not free to act in ways that would make her morally responsible for her conduct.26

23 

Here I draw from McKenna and Warmke (2017). On certain theories of action all actions involve reasons as causes. On such a view this claim would need modifying. 25  Accommodations must be made for actual-​sequence theorists who contend that the ability to do otherwise is not required (e.g., Fischer and Ravizza 1998; McKenna 2013; and Sartorio 2016). For a brief explanation see McKenna (2017). 26  See Talbert (2012) for an opposing view. 24 

Responsibility and Situationism    483 Because of this modal aspect of reasons-​responsiveness, an agent’s being reasons-​ responsive supports the truth of counterfactuals regarding how an agent would respond in a range of non-​actual but possible scenarios that are relevantly like the actual scenarios in which she finds herself. Suppose an agent acts on impulse when she responds hurtfully to a friend. She might nevertheless be reasons-​responsive if there is a range of similar conditions in which, were different reasons present, she would alter her conduct and not respond hurtfully. For instance, if she saw the friend was injured or depressed at the time, or if she was prompted by something in her environment to consider the likely effects of a hurtful remark, she would respond suitably. The second step of the modal response uses this distinction between one’s capacity for reason-​responsiveness and one’s actual exercise of that capacity to show that while situational features may affect the exercise of one’s capacity to respond to reasons, they need not affect the capacity itself (or at least they need not degrade that capacity below the threshold for responsibility). And since the capacity is what is required for free and responsible agency, situational features, so construed, do not undermine responsibility. Brink explains: It is important to frame this approach to responsibility in terms of normative competence and the possession of these capacities for reasons-​responsiveness. In particular, responsibility must be predicated on the possession, rather than the use, of such capacities. We do excuse for lack of competence. We do not excuse for failures to exercise these capacities properly. Provided the agent had the relevant cognitive and volitional capacities, we do not excuse the weak-​willed or the willful wrongdoer for failing to recognize or respond appropriately to reasons. (2013)

While it might be true that situational features make a crucial causal contribution to how a pertinent capacity is exercised, we cannot take this as a basis for exculpation, at least as it concerns threats to reasons-​responsiveness. What is required is that the contribution causally undermines or bypasses the pertinent cognitive and volitional (receptive and reactive) capacities. That is what is really at issue.

4.2  Pessimistic Realism This is the modal strategy for responding to the situationist threat. To what conclusion does the modal strategy lead us? I think we are led to a view called pessimistic realism.27 This view has two parts, the realism and the pessimism. Let’s take each in turn. First, the modal strategist concludes that situationist literature does not show that agents’ capacities are generally undermined or degraded, thereby rendering one unfree. 27 

McKenna and Warmke (2017). For more detailed discussion of a similar view, see Brink (2013), Herdova and Kearns (2017), and Miller (2017).

484   Brandon Warmke How could we know this? Part of the problem is that in actual cases, it is very difficult (perhaps humanly impossible) to know what is truly going on with one’s underlying capacities. However, on the modal strategy, we can point to a spectrum of counterfactual scenarios where it seems likely that the agent in question would respond to sufficiently good reasons. For example, we can ask, would the seminarians in the Good Samaritan study have stopped to help a person with an axe in their skull? What about one who was set ablaze or instead holding a child bleeding from the mouth? Likewise, what about the Milgram experiments? Would the “teachers” have treated the “students” so badly if the students were given the same first names as the teachers’ parents, or if the suggested forms of punishment were degrading in even more extreme ways? White noise appears to make one less likely to help someone pick up papers. But what if the confederate who dropped the papers was visibly disabled and in desperate need of help? The smell of sweet treats boosts the likelihood of helping make change. But would that smell have had such an effect if the person needed, not four quarters, but someone to call 911? Thinking through these studies, it appears that we have very little evidence that situational factors eliminate competence, and pretty good reason to think that sufficient competence is actually retained. These counterfactual scenarios reveal an important difference between an agent’s capacity for responding to a suitable pattern of reasons and the agent’s actual behavior. That situational features play a crucial role in an agent’s not actually responding to the good moral reasons does not show that an agent loses a more general capacity to respond to good reasons. And it is this general capacity, the modal strategist claims, that is required for free and responsible agency. We lack sufficient evidence that subjects in the situationist studies lose the capacity to respond to a large and reasonable range of moral reasons. Absent this evidence, we don’t have good reason to think that situational features like those identified in the studies undermine or degrade our reasons-​ responsive capacities. The first conclusion drawn by the modal strategist, then, is that at present, the situationist literature does not show that situational features generally render agents unfree and not responsible. Since these modal strategists (like myself) typically affirm that we have free will in the first place, we can call this the realist part of the conclusion, since the modal strategist remains a realist about free will and moral responsibility in the face of the situationist threat.28

28 

As Dana Nelkin and Derk Pereboom both pointed out in correspondence, the term “realist” is potentially misleading when used to describe the negative claim that the situationist threat, as presently understood, does not undermine free and responsible agency. After all, a free will skeptic like Derk Pereboom could maintain that situationism doesn’t undermine free and responsible agency, yet consistently deny that we are free and responsible, just for other reasons. Agnostics about free will could deny the situationist threat, too. A further problem with the label is that one might think that the situationist threat alone does not undermine free and responsible agency, but that the situationist threat,

Responsibility and Situationism    485 While the current situationist literature doesn’t show that situational features generally undermine our freedom and responsibility, these studies do cast what agency we do have in a relatively unflattering light. This is the pessimistic part of the view. The pessimism is induced in one of two ways (and perhaps both). First, even if our reasons-​ responsive capabilities are not regularly affected by situational features, the power of our own agency is considerably diminished if, for example, as it happens our moral concern for others can so easily be shifted about depending on utterly trivial alterations in conditions, say, involving colors or smells in the room. The pessimistic realist concedes that situational features can affect how we exercise our reasons-​responsiveness capacities. Situational features may distract us, or otherwise shift our attention or priorities in ways that influence us away from acting on the good reasons there are. We retain the relevant capacities, but fail to exercise them properly. This is Brink’s (2013) view. Yet simply failing to exercise properly one’s normative capacities does not entail that one does not act freely and responsibly. Failures to act on good reasons are common. The situationist challenge shows these failures may be more common than supposed. But the situationist literature may lead us to be pessimistic in another way. Situational features may indeed shrink the spectrum of reasons to which an agent is reasons-​ responsive, just not below the threshold required for morally responsible agency. On this view, our capacities are sometimes negatively affected, but not to the point that we generally lack free will (McKenna and Warmke 2017). Because reasons-​responsiveness capacities are not all or nothing, a mere negative effect on our capacities need not undermine free will if those capacities remain responsive to a sufficiently large set of moral reasons. The pessimistic realist continues to affirm that morally mature adults are generally free and responsible. And yet situational features sometimes either (1) “gum up” the proper exercise of those capacities in accordance with the good moral reasons there are, or (2) slightly shrink the spectrum of reasons to which we are capable of being reasons-​ responsive. Although this should not generally lead us to regard agents as unfree, these facts should perhaps lead us to mitigate our blame and praise, knowing how susceptible we can be to irrelevant situational features. The situationist challenge may also provide reasons to be more forgiving to those who fail to treat us rightly. Their situations may have made it more difficult for them to do the right thing.29

along with other threats, pushes our agency below the threshold sufficient for moral responsibility. I concede these points. I use the term here, however, for two reasons. First, McKenna and I used it in our (2017) and even recent habits are hard to break. (I now regret using this label; I won’t speak for McKenna.) Second, the label does capture the general consensus among those who have taken a stand on the situationist threat: We are generally free and responsible and the situationist literature doesn’t (yet) show otherwise. 29 

Nelkin (2016).

486   Brandon Warmke

5.  Objections to the Modal Strategy I have claimed that the modal strategy outlined earlier should lead us to respond to the situationist threat with pessimistic realism. I’ll now register some objections to this proposal. Crucial to the modal strategy is the claim that a failure to exercise a capacity to respond to reasons due to situational influence does not mean that one’s reasons-​ responsive capacities have been degraded by situational influence. The modal strategy must therefore be able to distinguish: (1) an agent’s retaining the requisite capacity for reasons-​responsiveness yet failing to properly exercise that capacity due to situational influence from (2) an agent’s failing to respond to good reasons due to situational influence that degrades or undermines one’s capacity for reasons-​responsiveness. In Brink’s terminology, (1) describes a performance error and (2) describes a competence error. Yet how do we determine whether situational features have negatively affected one’s reasons-​responsive capacities, or instead negatively affected only how those capacities are exercised? Brink appears committed to the view that we can distinguish performance errors from competence errors. But how? Here is a proposal. First, we ask this: Did the situational features in the actual scenario negatively affect the functioning of a relevant reasons-​responsive capacity? If not, situational features do not threaten reasons-​responsiveness. If, conversely, the situational features have negatively affected the capacity, then we identify the situationally affected capacity for responsiveness, hold it fixed, and then look at the relevant set of counterfactual scenarios where the agent is presented with diverse sets of good reasons to act and see if the agent reveals a sufficiently sane and understandable pattern of responses with those affected capacities. If she does reveal such a pattern, then the situational features, even if they have negatively affected the capacity, have not done so in a freedom and responsibility-​undermining way. If she does not reveal such a pattern, then (presumably) the situational factors have sufficiently undermined her reasons-​responsive capacity so as to render her unfree and not responsible. This response is sound so far as it goes. But problems remain. One is that we may have just kicked the bump in the rug. We wanted to know what distinguishes a competence error from a performance error. Our answer began by determining whether a situational feature affected the relevant capacities. But we may now ask this: How do we know when a situational feature has affected the capacity or instead affected only its exercise? We must know this before we can hold the relevant capacities fixed and look

Responsibility and Situationism    487 upon the counterfactual scenarios. It will be difficult enough to answer this question as a theoretical exercise. It will be even more difficult to discern how exactly situations affect us in real-​world cases. Consider, for example, the difficult question whether soldiers in wartime situations are morally responsible and blameworthy for war crimes or other immoral acts committed under the stress and pressure of combat.30 How do we determine whether such agents have had their capacities undermined or instead only failed to exercise those capacities appropriately? Here’s another objection to the modal response to the situationist threat. Suppose we discovered that situational features rarely undermine the relevant capacity, but that such features very often play a crucial role in the agent’s conduct. In other words, suppose we found that situational features were commonly causally contributing to our performance errors, and yet our underlying capacities remained intact (whatever that might mean). In such a scenario, morally irrelevant situational features constantly contribute to the production of conduct that is not a response to good reasons. Smells in the room, colors on the wall, and the mere presence of others often play a crucial role in producing our behavior. What should the modal theorist say? She could say that since the capacity is preserved, there is no situationist threat to reasons-​responsiveness, full stop. But this strikes me as tin-​eared. At a certain point, it strains credulity to say that an agent remains reasons-​responsiveness even though she very often fails to respond to good reasons because of how those situational features affect her exercise of those capacities. How many “performance errors” must we tolerate before we admit that a capacity itself has been undermined? Here’s another case. Suppose, unbeknown to me, someone slips me a drug that for half an hour makes me much less attentive to others, but not completely inattentive. In an ordinary sense, I still possess my usual capacities. But couldn’t the fact that I was slipped the drug still count as an excuse?31 Adverting to the preserved capacities is not sufficiently responsive to the objection. Suppose the modal theorist maintains that in the scenario just envisioned, the agent retains reason-​responsive capacities. It bears pointing out that she need not conclude that the situationist threat is therefore totally dissolved. She may deny that the situationist threat generally challenges agents’ reasons-​responsiveness, but then point out that there are other ways to undermine morally responsible agency. However, if the situationist threat is not to reasons-​responsiveness as the modal theorist construes it, what is the threat? Here, the modal theorist might suggest that the threat is determinism: The situational features operate in such a way as to deterministically bring us to fail to act on good reasons. Or instead, she might suggest that the threat is situational luck. Given the outsized role that irrelevant and largely unpredictable situational features play in the production of action, whether we act in a

30  31 

See, e.g., Doris and Murphy (2007) and Talbert and Wolfendale (2019). I thank Derk Pereboom for this example.

488   Brandon Warmke certain way is largely a matter of luck: A kind of luck that agents situated in the world cannot easily escape, if at all.32 But to construe the situationist threat in this way is to go back on our earlier claim that the best way to understand it is as a challenge to reasons-​responsiveness. On the kind of scenario we are imagining, one’s reasons-​ responsiveness is still intact because the relevant capacities have not been undermined. So if there still looks to be a threat to morally responsible agency, what is it? I leave this question for others. Carolina Sartorio and Manuel Vargas offer a different objection to the modal strategy. Sartorio (2018) argues that modalist responses are subject to a “demarcation problem.” The problem arises once we ask the following: To what counterfactual scenarios do we look to assess whether an agent is properly responsive? On the one hand, if we want to know whether someone in a hurry is appropriately reasons-​ responsive, we should not assess their responsiveness in, for example, counterfactual situations where one is not in a hurry. Those situations are irrelevant because ostensibly, what we want to know is whether given being hurried, one is morally responsible for failing to help. This suggests that we must hold fixed at least some of the situational features, otherwise those possible scenarios look irrelevant. On the other hand, we cannot simply take the relevant capacities and then reinsert them back into the actual situation. That is, we cannot hold all of the situational features fixed. This would tell us nothing about the spectrum of reasons to which the agent with those capacities would respond. Here is the puzzle, as Sartorio herself puts it: “we need some principled criterion to set apart, in each case, the aspects of the circumstances that we should hold fixed from those that we can legitimately vary. Unless we can offer such a criterion, reasons-​responsiveness views rest on shaky foundations” (2018: 800). Manuel Vargas (2013b) raises a similar objection to standard modal responses like those inspired by the Fischer and Ravizza paradigm. It is not enough, Vargas argues, to just find any set of counterfactual scenarios where the agent in question responds to good reasons. Rather, one must find a suitable portion of worlds relevantly similar to the actual situation where the agent is responsive. After all, if you want to know whether being hurried undermines your responsiveness, what help would it do to go to scenarios in which you aren’t in a hurry? Vargas argues for his own modal version of reasons-​ responsiveness that differs from the Fischer and Ravizza view. On his view, responsiveness to reasons is relativized to an agent’s circumstance in the actual scenario. We test for responsiveness by looking at how the agent responds in similar other hurried situations. Fischer and Ravizza’s version makes no such restriction. As Herdova and Kearns (2017) point out, this will have the result that situational features are more likely to undermine or degrade agency.33

32 

See Herdova and Kearns (2015). See Herdova and Kearns (2017) for a defense of the Fischer and Ravizza paradigm against Vargas’s critique. 33 

Responsibility and Situationism    489

6.  Concluding Remarks Let’s take stock. I have argued that while the situationist literature puts our agency in an unflattering light, it does not at present show that agents’ capacities for responsiveness to reasons are generally undermined by situational features. If we are already free, our situations make us look bad, but they don’t generally undermine free and responsible agency. This is pessimistic realism. We began the paper with three “pictures” of how situations could relate to our agency—​as stage, as occasional trap, and as threat. Pessimistic realism, I think, places us somewhere between pictures two and three. I wish I had more to offer, but at present, I can’t see the issues any clearer than this. In conclusion, I leave the reader with a few thoughts about future directions. I have assumed that the right way to think of reasons-​responsiveness is in modal terms. But suppose, as Sartorio (2016) argues, what matters for reasons-​responsiveness is not whether some capacity, defined by a set of counterfactual responses to reasons, is retained. Rather, what matters is simply the actual sequence of events leading up to and involving the agent’s acting and the actual reasons to which the agent does or does not respond. According to this actualist version of reasons-​responsiveness, we need not advert to the possession of capacities and their operation in counterfactual scenarios. All we need to know is whether, in the actual scenario, the agent was appropriately sensitive to the reasons there were. If situational features prevent an agent from being sensitive to a sufficient spectrum of the good reasons in the actual sequence of events, then her responsiveness to reasons is compromised and so with it moral responsibility. We don’t have to look at counterfactual scenarios to establish this.34 So should we be modalists or actualists about reasons-​responsiveness? Or should we adopt another framework, such as Vargas’s? And if so, would this alter the picture of how situations affect our agency? Second, are we right to construe the situationist threat as fundamentally one about reasons-​responsiveness in the first place? I have proceeded, like many others, to think of matters this way. But as noted earlier, perhaps the problem really does fundamentally concern determinism or situational luck after all. Yet another possibility: Current empirical evidence underdetermines which, if any, way of construing the threat is fundamental. Depending on how the facts on the ground turn out, a different threat may be fundamental. Or perhaps there is no singular “situational” threat: depending on the case, the threat is determinism, luck, or something else. Third, I have focused our attention on the negative effects that situational features have on our agency. Sometimes, however, situations apparently improve our agency. After all, smelling sweet treats made people more likely to respond to reasons to help. Recently, some philosophers have argued that knowing more about our situations

34  Crucial to Sartorio’s actualist view is the metaphysical claim that agents respond to both reasons and absences of reasons. Response to absences of reasons replaces the relevance of counterfactual scenarios in standard modal theories. I’ll set aside this complication here.

490   Brandon Warmke and the agential boosts they provide can therefore improve our capacities for seeing and responding to good moral reasons.35 How much can such knowledge help? And do these positive effects in any way counterbalance the negative situational influences? There may be room for a more optimistic realism, after all.36

References Alfano, Mark. 2013. Character as Moral Fiction. Cambridge: Cambridge University Press. Annas, Julia. 2003. “Virtue Ethics and Social Psychology,” A Priori 2: 20–​34. Ballantyne, Nathan. 2015. “Debunking Biased Thinkers (Including Ourselves),” Journal of the American Philosophical Association 1: 141–​162. Baron, Marcia. 2003. “Manipulativeness,” Proceedings and Addresses of the American Philosophical Association 77: 37–​54. Baron, R. 1997. “The Sweet Smell of . . . Helping: Effects of Pleasant Ambient Fragrance on Prosocial behavior in Shopping Malls,” Personality and Social Psychology Bulletin 23: 498–​503. Brink, David. 2013. “Situationism, Responsibility, and Fair Opportunity,” Social Philosophy and Policy 30: 121–​149. Brink, David, and Dana Nelkin. 2013. “Fairness and the Architecture of Responsibility.” In David Shoemaker, ed., Oxford Studies in Agency and Responsibility, vol. 1 Oxford: Oxford University Press: 284–​313. Cann, A., and J. Blackwelder. 1984. “Compliance and Mood: A Field Investigation of the Impact of Embarrassment,” The Journal of Psychology 117: 221–​226. Ciurria, Michelle. 2013. “Situationism, Moral Responsibility and Blame,” Philosophia 41: 179–​193. Coates, D. Justin, and Philip Swenson. 2013. “Reasons-​ Responsiveness and Degrees of Responsibility,” Philosophical Studies 165: 629–​645. Darley, J., and C. Batson. 1973. “‘From Jerusalem to Jericho:’ A Study of Situational and Dispositional Variables in Helping Behavior,” Journal of Personality and Social Psychology 27: 100–​108. Dennett, Daniel. 1984. Elbow Room: Varieties of Free Will Worth Wanting. Cambridge, MA: MIT Press. Doris, John. 1998. “Persons, Situations, and Virtue Ethics,” Noûs 32: 504–​530. Doris, John. 2002. Lack of Character. Cambridge, MA: Cambridge University Press. Doris, J., and D. Murphy. 2007. “From My Lai to Abu Ghraib: The Moral Psychology of Atrocity,” Midwest Studies in Philosophy 31: 25–​55. Fischer, John Martin. 2006. My Way: Essays on Moral Responsibility. New York: Oxford University Press. Fischer, John Martin, and Mark Ravizza. 1998. Responsibility and Control: An Essay on Moral Responsibility. Cambridge: Cambridge University Press. Flanagan, Owen. 1991. Varieties of Moral Personality. Cambridge, MA: Harvard University Press.

35 

See Mele and Shepherd (2013), Miller (2016), and McKenna and Warmke (2017). I thank Dana Nelkin and Derk Pereboom for their generous invitation to contribute to this fine volume and for their thoughtful comments on a previous draft. Remaining errors are not my fault and are to be explained by my neighbor’s very annoying lawn mower. 36 

Responsibility and Situationism    491 Gert, Bernard, and Tim Duggan. 1979. “Free Will as the Ability to Will,” Noûs 13: 197–​217. Haji, Ishtiyaque. 1998. Moral Appraisability: Puzzles, Proposals, and Perplexities. New York: Oxford University Press. Harman, Gilbert. 1999. “Moral Psychology Meets Social Psychology: Virtue Ethics and the Fundamental Attribution Error,” Proceedings of the Aristotelian Society 99: 315–​331. Herdova, M., and S. Kearns. 2015. “Get Lucky: Situationism and Circumstantial Moral Luck,” Philosophical Explorations 18: 362–​377. Herdova, M., and S. Kearns. 2017. “This is a Tricky Situation: Situationism and Reasons-​ Responsiveness,” Journal of Ethics 21: 151–​183. Isen, A., and P. Levin. 1972. “Effect of Feeling Good on Helping: Cookies and Kindness,” Journal of Personality and Social Psychology 21: 384–​388. Isen, A., and J. Reeve. 2005. “The Influence of Positive Affect on Intrinsic and Extrinsic Motivation: Facilitating Enjoyment of Play, Responsible Work Behavior, and Self-​Control,” Motivation and Emotion 29: 297–​325. Kamtekar, Rachana. 2004. “Situationism and Virtue Ethics on the Content of Our Character,” Ethics 114: 458–​491. Kupperman, Joel J. 2001. “The Indispensability of Character,” Philosophy 76: 239–​250. Latané, B., and J. Darley. 1968. “Group Inhibition of Bystander Intervention in Emergencies,” Journal of Personality and Social Psychology 22: 287–​295. Latané, B., and J. Rodin. 1969. “A Lady in Distress: Inhibiting Effects of Friends and Strangers on Bystander Intervention,” Journal of Experimental Social Psychology 5: 189–​202. Lieberman, Matthew. 2005. “Principles, Processes, and Puzzles of Social Cognition,” NeuroImage 28: 746–​756. Matthews, K., and L. Canon. 1975. “Environmental Noise Level as a Determinant of Helping Behavior,” Journal of Personality and Social Psychology 32: 571–​577. McKenna, Michael. 2012. Conversation and Responsibility. New York: Oxford University Press. McKenna, Michael. 2013. “Reasons-​Responsiveness, Agents, and Mechanisms.” In David Shoemaker, ed., 2013. Oxford Studies in Agency and Responsibility, vol. 1. Oxford: Oxford University Press: 151–​184. McKenna, Michael. 2017. “Reasons-​Responsive Theories of Freedom.” Routledge Handbook on Free Will. Routledge Press: 27–​40. McKenna, Michael, and Brandon Warmke. 2017. “Does Situationism Threaten Free Will and Moral Responsibility?,” Journal of Moral Philosophy 14: 698–​733. Mele, Alfred. Free Will and Luck. New York: Oxford University Press. Mele, Alfred, and Joshua Shepherd. 2013. “Situation and Agency,” Journal of Practical Ethics 1: 62–​83. Merritt, Maria. 2000. “Virtue Ethics and Situationist Personality Psychology,” Ethical Theory and Moral Practice 3: 355–​401. Milgram, S. 1969. Obedience to Authority. New York: Harper & Row. Miller, Christian B. 2003. “Social Psychology and Virtue Ethics,” The Journal of Ethics 7: 365–​392. Miller, Christian B. 2013. Moral Character: An Empirical Theory. Oxford: Oxford University Press. Miller, Christian B. 2014. Character and Moral Psychology. Oxford: Oxford University Press. Miller, Christian B. 2016. “Virtue Cultivation in Light of Situationism.” In Julia Annas, Nancy Snow, and Darcia Narvaez, eds., Developing the Virtues: Integrating Perspectives. Oxford: Oxford University Press: 157–​183.

492   Brandon Warmke Miller, Christian B. 2017. “Situationism and Free Will,” In Meghan Griffith, Kevin Timpe, and Neil Levy, eds., Routledge Companion to Free Will. Routledge. Montmarquet, James. 2003. “Moral Character and Social Science Research,” Philosophy 78: 355–​368. Nahmias, Eddy. 2007. “Autonomous Agency and Social Psychology.” In M. Maragga, M. De Caro, and F. Ferretti, eds., Cartographies of the Mind. Berlin: Springer. Nelkin, Dana. 2005. “Freedom, Responsibility, and the Challenge of Situationism,” Midwest Studies in Philosophy 29: 181–​206. Nelkin, Dana. 2011. Making Sense of Freedom and Responsibility. Oxford: Oxford University Press. Nelkin, Dana. 2016. “Difficulty and Degrees of Moral Praiseworthiness and Blameworthiness,” Noûs 50: 356–​378. Pereboom, Derk. 2001. Living Without Free Will. Cambridge: Cambridge University Press. Pereboom, Derk. 2014. Free Will, Agency, and Meaning in Life. Oxford: Oxford University Press. Rodgers, Travis, and Brandon Warmke. 2015. “Situationism versus Situationism,” Ethical Theory and Moral Practice 18: 9–​26. Ross, L. and R. Nisbett. 1991. The Person and the Situation: Perspectives on Social Psychology. New York: McGraw Hill. Russell, Daniel. 2009. Practical Intelligence and the Virtues. Oxford: Clarendon Press. Sabini J., and M. Silver. 2005. “Lack of Character? Situationism Critiqued,” Ethics 115: 41–​48. Sarkissian, Hagop. 2010. “Minor Tweaks, Major Payoffs: The Problem and Promise of Situationism in Moral Philosophy, Philosophers’ Imprint 10: 1–​15. Sartorio, Carolina. 2016. Causation and Free Will. Oxford: Oxford University Press. Sartorio, Carolina. 2018. “Situationism and Responsiveness to Reasons,” Noûs 52: 796–​807. Slingerland, Edward. 2011. “The Situationist Critique and Early Confucian Virtue Ethics,” Ethics 121: 390–​419. Spinoza, Baruch. 1677/​1985. “Ethics.” In Edwin Curley, ed. and trans., The Collected Works of Spinoza, vol. 1. Princeton, NJ: Princeton University Press. Sreenivasan, Gopal. 2002. “Errors about Errors: Virtue Theory and Trait Attribution,” Mind 111: 47–​68. Talbert, Matthew. 2009. “Situationism, Normative Competence, and Responsibility for Wartime Behavior,” Journal of Value Inquiry 43: 415–​432. Talbert, Matthew. 2012. “Moral Competence, Moral Blame, and Protest,” Journal of Ethics 16: 89–​109. Talbert, Matthew, and Jessica Wolfendale. 2019. War Crimes: Causes, Excuses, and Blame. New York: Oxford University Press. Tucker, Chris. 2004. “Harman vs. Virtue Theory: Do Character Traits Explain Behavior?,” Southwest Philosophy Review 21: 137–​145. Vargas, Manuel. 2013a. Building Better Beings. Oxford: Oxford University Press. Vargas, Manuel. 2013b. “Situationism and Moral Responsibility: Free Will in Fragments.” In Andy Clark, Julian Kiverstein, and Tillmann Vierkant, eds., Decomposing the Will. Oxford: Oxford University Press: 325–​349. Vranas, Peter. 2005. “The Indeterminacy Paradox: Character Evaluations and Human Psychology,” Noûs 39: 1–​42. Webber, Jonathan. 2006. “Virtue, Character and Situation,” Journal of Moral Philosophy 3: 193–​213. Wielenberg, Erik. 2006. “Saving Character,” Ethical Theory and Moral Practice 9: 461–​491.

Responsibility and Situationism    493 Wolf, Susan. 1990. Freedom Within Reason. Oxford: Oxford University Press. Wood, Allen. 2014. “Coercion, Manipulation, Exploitation.” In Christian Coons and Michael Weber, eds., Manipulation: Theory and Practice. New York: Oxford University Press, 17–​50. Zimmerman, Michael J. 1987. “Luck and Moral Responsibility,” Ethics 97: 374–​386.

CHAPTER 24

Ex perimental Ph i l o s oph y and Moral Resp onsi bi l i t y Gunnar Björnsson

Experimental philosophy, as I understand it here, is the attempt to help answer ontological, semantic, or epistemological questions about some matter of philosophical concern by quantitatively measuring judgments about that subject matter, typically among non-​philosophers. The last couple of decades has seen a flurry of such attempts in various subdisciplines, including ethics, epistemology, philosophy of mind, philosophy of language, and action theory. Philosophers have of course always relied on judgments about their subject matters, but have less often relied on non-​philosopher’s judgments, and rarely in ways involving quantitative assessments of the distribution of such judgments. Philosophers have also from time to time relied on quantitative empirical data, perhaps most prominently in philosophy of mind and applied ethics, but rarely on data on folk judgments about the subject matter of concern. Experimental philosophy contrasts with both these more traditional philosophical practices. The ambitions of experimental philosophers vary. Some modestly hope to correct misapprehensions about how certain expressions are ordinarily used. Others think that empirical methods might radically reshape philosophical inquiry. The revolutionary potential has prompted interesting broad questions: Might experimental philosophy show that attempts to solve philosophical problems “from the armchair” are fundamentally mistaken, or are data on judgments largely irrelevant for central philosophical problems?1 Under what conceptions of philosophical inquiry might empirical methods be more or less relevant? Are some philosophical questions particularly amenable to these empirical methods, whereas others are not? Should experimental philosophy both become the sustained concern of a substantial group of specialists

1 

For discussion, see, e.g., Kauppinen 2007; Knobe 2007; Alexander et al. 2010; and contributions to Sytsma and Buckwalter 2016.

Experimental Philosophy and Moral Responsibility    495 and a source of relatively accessible tools used throughout philosophy, much as philosophical logic?2 Some of what is said in this chapter is relevant to these broader issues. Our question is narrower, however, in keeping with the focus of this volume: How, if at all, might experimental philosophy help us answer central questions about the nature of moral responsibility, such as the question of whether moral responsibility is compatible with determinism?3 Might folk judgments in line with one answer provide support for that answer? This might seem implausible, as a non-specialist is at best marginally more likely than not to be correct. But even meagre reliability of suitably independent judgments accumulate, yielding strong support for the dominant judgment under the right conditions. We ask whether those conditions are satisfied for this question. Throughout these discussions, I will assume a broad familiarity with philosophical research on moral responsibility. Understanding the philosophical relevance of experimental results involves the same difficulties and requires the same familiarity with distinctions and possible theoretical options as does understanding the relevance of arguments that rely, in familiar ways, on philosophers’ verdicts about cases.

1.  Wisdom of the Crowd? Why should philosophers care about what nonspecialists think about a complex philosophical question? A straightforward answer would be this: If one position on this issue is supported by significantly more people, this provides some evidence for that position. In line with this, some of the early experimental work on whether non-​philosophers are compatibilists or incompatibilists was indeed a response to claims by some philosophers that their position had common sense on its side, claims taken (correctly or not) to imply that the burden of proof would be on the other side (Nahmias et al. 2005: 563–​564). Some will balk at this answer. Given the difficulty of the questions about incompatibilism and given how easily people get confused when thinking about it, one might find it obvious that folk judgments carry no evidential weight whatsoever. But there is reason to at least take the suggestion seriously and be clear about why we reject it. Generally speaking, and for most of the questions we consider in our lives, our ordinary ways of answering these questions are somewhat reliable, or better than chance at getting things right. But if people in general are slightly more likely than not to get things

2 

For the last suggestion, see Weinberg 2015. For some recent overviews of the problems and prospects of experimental philosophy, see Mallon 2016; Knobe and Nichols 2017; Ludwig 2017. For data on the reproducibility of results in experimental philosophy, see Cova et al. 2018; for discussion of the use of online subjects, see, e.g., Paolacci and Chandler 2014. 3  For some earlier overviews of the relevance of experimental philosophy to the philosophy of moral responsibility specifically, see Vargas 2006; Nelkin 2007; Sommers 2010; Björnsson and Pereboom 2016; Chan et al. 2016.

496   Gunnar Björnsson right, and if their chances of doing so are independent, then if a significant majority of a large number of people agree on a given issue, the majority would most likely be correct. Or at least they would be in cases where there is one correct answer, no answer is antecedently more likely than the other, and judges would be equally likely of getting things right independently of which answer is in fact correct.4 Under those conditions, if the chance for each to arrive at the correct answer is a modest .55, then if 61 out of 100 such judges come down on one side, there is nevertheless over a .99 chance that they are correct. If the chance for each is a mere .505, then the same is true if it is supported by 605 out of 1,000 judges.5 The sheer difference in number of supporting judgments make up for the modest or meager quality of each judgment. How is this possible? It helps to think of marginally reliable judges as being mostly affected by factors that do not track the truth of the matter—​by noise—​and only to some small degree by factors that do. While noise is random and cancels out over a large number of independent judgments, the rare instances where judgments are guided by the truth always pull in the same direction. This sort of appeal to the wisdom of the crowd has many applications, but its strength depends on the extent to which: (1) No answer is antecedently more likely than others in a way that lets us significantly discount the reliability of judges arriving at some other answer. (2) People’s chances of getting things right are independent. Suppose that each is more likely than not to get things right but is so by leaning on the same piece of evidence as everyone else, a piece which is slightly more likely than not to reveal the truth. Then even if noise cancels out, uniformity in judgment will primarily reflect what that piece of evidence suggests. However large the group and uniform their agreement, the likelihood that the group gets things right will not exceed the likelihood that the evidence in question is veridical. (3) People are reliable when judging this particular issue, being somewhat better than chance in getting things right. (We obviously cannot assume that people are highly reliable when it comes to difficult philosophical questions. But given that even very low reliability can ground a wisdom of the crowd argument, we cannot reject this reliability assumption without further consideration.) (4) The question has one correct answer, independently of who is answering it. If judgments are made correct by different facts (or lack truth-conditions), large numbers cannot be expected to accumulate truth-​tracking while canceling out noise. 4  The reasoning that is merely roughly indicated here is made more precise in discussions of the Condorcet Jury Theorem. For the relevance of CJT to experimental philosophy and a more formal treatment of both reliability and independence, see discussions in Talbot 2014; Ludwig 2017. Cf. Goodin and Spiekermann 2018. 5  These numbers, and corresponding numbers provided later on are calculated based on Theorem 9 in List 2004: 528. Cf. Talbot 2014. I move freely between talk of probabilities, likelihoods, and chances. What I have in mind in each case is something like fitting credences given accessible information; I take the discussion to be robust relative to various ways of spelling out that notion.

Experimental Philosophy and Moral Responsibility    497 (5) A significantly larger number of judgments comes down on one side of this issue compared to the other. If the difference in numbers is very slim, there is little accumulation of evidential weight. Can such an argument be had for central questions about responsibility? In what follows, I will discuss conditions (2) through (5), in reverse order. To constrain the discussion, I will continue to focus primarily on the question of whether incompatibilism is correct and restrict myself to a limited range of studies telling us whether and under what circumstances people tend to make compatibilist and incompatibilist judgments, respectively. Though this will mean ignoring other interesting studies, I take large parts of what is said here to generalize.6 For the same reason, I will also assume (unless otherwise stated) that compatibilism and incompatibilism are antecedently equally likely.

2.  How Many Accept or Reject Compatibilism? Several studies have tried to determine whether non-​philosophers take determinism to be incompatible with moral responsibility; whether they are “natural incompatibilists” (e.g. Nichols 2004; Nahmias et al. 2005; Nahmias et al. 2007; Nichols and Knobe 2007; Roskies and Nichols 2008; Sarkissian et al. 2010). In these studies, subjects are typically presented with either an indeterministic or a deterministic scenario and asked about whether agents can be morally responsible for their decisions or actions in that scenario. If judges are willing to attribute responsibility in the former but not the latter condition, this would suggest that they are incompatibilists. What has been the conclusion of these studies? Unfortunately, results have varied, depending on how the (in)deterministic aspect of the scenarios were described and on how questions about responsibility were asked. For example, looking at two of the most cited studies, the proportion of subjects that are willing to attribute responsibility to agents in deterministic scenarios vary between 14% (Nichols and Knobe 2007: 670) and 89% (Nahmias et al. 2007: 227). A number of reasons for this variation have been identified: Abstract vs. Concrete: When subjects are asked to attribute responsibility to a specified individual for a concrete bad action in a deterministic universe (“Is it possible that Bill is fully morally responsible for killing his family?”), they attribute more responsibility than when asked, in the abstract, whether someone could be fully

6  Among the many studies that I set to the side are those concerned with the extent to which people acknowledge moral luck (e.g., Cushman 2008; Kneer and Machery 2019), how people understand the relation between consciousness and responsibility (e.g., Shepherd 2015), and whether people take there to be historical conditions on responsibility (Taylor and Maranges 2020).

498   Gunnar Björnsson morally responsible in such a universe (“Is it possible for a person to be fully morally responsible for their actions?”). In one study, 72% of subjects who were asked the concrete question gave a positive answer whereas only 14% of subjects who were asked the abstract question (Nichols and Knobe 2007); in another, the proportion of positive answers fell from 79% to 52% (Nahmias et al. 2007). Actual vs. Possible: When subjects are asked to consider whether agents would be fully morally responsible if the actual world is deterministic, they are also more inclined to give a positive answer than when considering agents located in some merely imagined deterministic universe. In one study, 89% gave positive answers to the former question, but only 72% to the latter (Nahmias et al. 2007; cf. Roskies and Nichols 2008). Neurology vs. Psychology: Subjects seem considerably more inclined to attribute responsibility when the deterministic scenario describes the deterministic processes in psychological terms (minds, thoughts, desires, plans) than when they are described in neurological terms (brains, chemical reactions, neural processes): 87% versus 41% in one study (Nahmias et al. 2007; but see De Brigard et al. 2009; Chan et al. 2016). Framing of Determinism: Subjects are significantly more inclined to attribute responsibility when the idea of determinism is conveyed as the idea that the same events always follow causally from the same initial conditions than when put in strongly modal terms (“given the past, each decision has to happen in the way that it does”): 52% versus 27% in one study (Murray and Nahmias 2014: 447).

Some of these variations suggest that certain ways of asking or framing the question change what answers seem reasonable without therefore changing the question (see, e.g., Nelkin 2007; Björnsson and Persson 2013, in addition to discussion in papers presenting the original studies). We will return to the matter of whether these variations affect the reliability of the answers given. Other variations might suggest that subjects have been asked questions that have different answers. Most obviously, it might be that different ways of conveying the deterministic aspect of the scenarios lead subjects to consider different versions of determinism, only some of which are incompatible with responsibility. More subtly, it might be that the nature of responsibility depends on how things are in the actual world, and that this might explain the actual vs. possible contrast (Roskies and Nichols 2008: 382–​ 385). For example, it could be that responsibility should be understood as requiring a kind of freedom ruled out by determinism if, as most people seem to believe (Nichols and Knobe 2007; Sarkissian et al. 2010; Deery et al. 2013), human decisions and actions are actually free in that way. But if the world is deterministic, as subjects are asked to assume in questions about the actual world, then perhaps responsibility should be understood in a way that still takes us to be responsible: The responsibility assumption would trump incompatibilist commitments (cf. van Inwagen 1983: 219ff). Another equivocation worry stems from the well-​known ambiguity of the term “responsibility.” Even if we restrict ourselves to retrospective moral responsibility, some have argued that it comes in several importantly different varieties, sometimes labeled “accountability,” “attributability,” and “answerability” (see, e.g., Watson 1996; Shoemaker

Experimental Philosophy and Moral Responsibility    499 2015; for criticism, see Smith 2015; Jeppsson 2016), where accountability for an action would be what is required for an agent to deserve retributive emotional and other responses. One possible explanation of diverging responsibility judgments about deterministic cases might then be that whereas accountability is generally taken to be incompatible with determinism, some subjects have another, less demanding, notion of responsibility in mind, one that merely requires that it makes sense to ask the agent why they did what they did, demand that they abstain from such actions, or assess their character based on their action (Björnsson and Pereboom 2016: 152). There is some reason to think that this cannot explain more than part of the variation in judgment, however. Several studies of folk judgments ask not only about responsibility but also whether agents deserve blame. Since answers to these two questions tend to be fairly strongly correlated, there is reason to think that subjects are concerned with responsibility understood exactly as a precondition for deserved blame. This does not settle the matter, of course: perhaps a significant proportion of subjects understand “deserves blame” differently (ibid.). But we currently lack data suggesting that this accounts for much of the variations in judgment listed earlier. Suppose that we could get around equivocation worries, and collected a set of answers from a large group of subjects with a significant majority coming down on one side of the issue (thus satisfying conditions [4]‌and [5] above). Suppose further that our antecedent reasons do not very strongly favor either compatibilism or incompatibilism (thus satisfying condition [1]). Then the extent to which the answers gives us reasons to accept the majority answer depends on the reliability of the subjects, and the extent to which their judgments track the truth independently. I discuss issues of reliability in sections 3–​5 and issues of independence in section 6.

3.  Folk Reliability, Disagreement, and Identifiable Bias Why should we think that experimental subjects are more likely than not to correctly assess whether agents in deterministic scenarios are morally responsible? A quick answer would be that even nonspecialists have enough conceptual competence to form beliefs about responsibility and, if the studies are set up right, some understanding of the deterministic aspects of the scenarios they consider. This answer might seem as misguided as it is quick. Human beings are famously subject to various perceptual and cognitive illusions, and there are issues that are simply beyond most people’s ken. Based on this, one might think that reliability has to be established for the judgments at hand, and that this requires independently answering the very question that folk judgments were supposed to help us with (cf. Ludwig 2017: 395–​396). But as a general demand, this is too strong. Before knowing the answer to a question, we can often determine that people have enough cognitive capacities to

500   Gunnar Björnsson ground a presumption of positive reliability. Pending specific reason to the contrary, reasons having to do with the matter at hand and the judge’s access to it, we should expect someone with such capacities to get things right more often than not. The question, then, is whether there are specific reasons to reject such a presumption of reliability with respect to compatibilist or incompatibilist judgments. One reason comes from disagreement about the issue at hand. For example, if 57% of a large number of judgments agreed with compatibilism and 43% disagrees with it, and the judgments are independently tracking the truth, the average reliability of these judgments is unlikely to much exceed 57.7 Given that most studies of folk incompatibilism reveal a significant minority, this might in itself restrict what degree of reliability we can plausibly attribute to subjects given some modest amount of independence. But that restriction would not itself prevent the totality of judgments from strongly supporting one answer. If the distribution of answers is compatible with a very modest reliability, the number of subjects is large enough, and each tracks the truth independently enough of the others, the upshot might still be a very high likelihood that the majority is correct. (As noted before, if 605 of 1,000 people who independently track truth with a meager .505 reliability agreed on an answer, this would make that answer 99% likely to be correct!) There are, however, more direct reasons for doubting the reliability of folk judgments about incompatibilism. One is the complexity and difficulty of the issue at hand; another the frequent missteps displayed when non-​philosophers begin to think about these issues, and the possibility that people’s judgments are sytematically biased in a way that swamps any general reliability. In section 5, we consider whether the nature of the issue gives us reason to reject folk reliability; in this and the next section, we look closer at evidence for specific sources of bias and error. One sign that non-​philosophers are unreliable judges of incompatibilism is provided by the confusion they display when first confronted with the issue. As those who have taught classes on incompatibilism know, students not only find many of these issues difficult, but also frequently make specific basic mistakes, conflating determinism and fatalism or conflating the question about whether people who do bad things can deserve blame under determinism with the question of whether there might still be forward-looking reasons to blame them. The point here is not that incompatibilist judgments are mistaken, nor that compatibilist judgments are—​we will soon discuss such specific claims—​but that non-​experts often make what are uncontroversially mistakes about matters that seem crucial for settling the question. Based on this, it might seem very reasonable not to expect their acceptance or rejection of incompatibilism to be reliably truth-​tracking. The fact that nonspecialists make these kinds of errors plausibly lowers their degree of reliability. But it does not necessarily mean that they will not do better than chance. As long as these various mistakes constitute noise—​as long as their effects are randomly distributed—​and as long as we could expect people’s judgments to have a slight tendency 7  For discussion of how the distribution of answers might affect plausible attributions of reliability, see Talbot 2014: 3867–​3871.

Experimental Philosophy and Moral Responsibility    501 to track the truth, and independently so, a sufficiently large majority of subjects coming down on one side might still give us strong reasons to accept that side. Things change, however, if we know that large proportions of people are swayed by biases that prompt judgments falling specifically on one side, biases that are unlikely to be canceled out. After discounting such judgments, the wisdom of the remaining crowd might point in the opposite direction. Here, experimental philosophers have proposed and provided empirical support for a number of hypothesis about factors that (i) drive folk judgments toward one kind of judgment and (ii) seem clearly irrelevant to the truth of the matter. In what follows, I outline and discuss three of the most important proposals.8 The first hypothesis is that many who make compatibilist judgments are willing to attribute responsibility no matter what, even to agents who are definitely not responsible. Specifically, Adam Feltz and Melissa Millan (2013) suggest, based on experimental data, that: No Matter What: Many who attribute responsibility to agents in deterministic scenarios also attribute responsibility to agents who should not be judged responsible because their actions are fated.

An action is “fated” in the relevant sense if it is bound to happen independently of the sorts of states—​character, values, preferences—​that compatibilists see as grounding responsibility. Fatalism is thus different from determinism: even if our agential states are determined by factors in the distant past, our actions are still counterfactually or causally dependent on these states. Through a series of experiments, Feltz and Millan showed that those who were willing to attribute free will and responsibility to agents in various “fatalistic” scenarios (between 30% and 60% of subjects, depending on experiment) were also much more willing to do so to agents in a deterministic scenario. (Most fatalistic scenarios involved a book where the agent’s decisions and actions are written beforehand and where everything written in the book will necessarily take place.) On the assumption that fatalism does undermine free will and responsibility, this would indeed suggest that a considerable proportion of compatibilist judgments are made by unreliable judges.9 If we could discount these judgments while taking the remaining 8  I set aside the suggestion that people’s thinking about free will and moral responsibility rely on (potentially suspect) ideas about a soul (a suggestion largely debunked in Mele 2014; Vonasch et al. 2018) and the idea that judgments relied upon in manipulation arguments for incompatibilism or responsibility (or accountability) nihilism should be interpreted in compatibilist-​friendly ways (proposed by Sripada 2012; debunked in Björnsson 2016). 9  By what processes would people erroneously attribute responsibility to fated agents? One hypothesis would be that subjects are in the grip of retributive emotions (cf. Nichols and Knobe 2007: 671–​672). However, hypotheses appealing to emotional effects seem to be in conflict with the available data, judging by a recent meta-​study (Feltz and Cova 2014). But there are other possibilities, not appealing to affect. One suggestion has been that we are operating with a general assumption that if something bad happens, someone is responsible (Mandelbaum and Ripley 2012). Another is that we want to believe that people are responsible, and adjust our other commitments accordingly in acts of motivated reasoning (Clark et al. 2019).

502   Gunnar Björnsson folk judgments to be at least somewhat reliable, the overall pattern of judgments might strongly support incompatibilism. Unfortunately for the No Matter What hypothesis, the kinds of fatalism that Feltz and Millan tested reactions to are not uncontroversially incompatible with responsibility. First, some of the descriptions of fated actions offered in their studies might well have been understood along the lines of determinism. Second, all their descriptions seem compatible with responsibility given popular “sourcehood” forms of compatibilism. According to these forms, inspired by Harry Frankfurt (1969), responsibility for an action only requires that it resulted from the right aspects of the agent, not that the agent could have avoided performing the action. Third, a study by James Andow and Florian Cova (2016) strongly suggests that most subjects who attribute responsibility to an agent in a fatalistic scenario take the agent’s actions to be the result of such agential aspects. Though attributions of responsibility in deterministic and fatalistic scenarios of a certain kind tend to go hand in hand, then, the process at work is not one that would necessarily make the judgments in question unreliable. To show that it is, we would need independent grounds for rejecting compatibilist reasoning. A different reason to discount apparently compatibilist judgments comes from Thomas Nadelhoffer, David Rose, Wesley Buckwalter, and Shaun Nichols (2019). They have suggested that: Indeterminist Intrusion: Many who attribute responsibility to agents in deterministic scenarios do so because they do not fully understand these scenarios as deterministic.

In experiments supporting this hypothesis, subjects were presented with deterministic scenarios and answered a number of questions, including questions about the free will and responsibility of agents in these scenarios. They were then asked what chance there was, on a scale from 0% to 100%, that these agents would do something other than what had been stipulated to follow given the past and deterministic laws. Restricting ourselves to scenarios involving human agents, well over half the subjects said that there was some non-​zero chance that this would happen, apparently thinking of the scenario as involving indeterministic elements. Moreover, whereas 63% of these subjects agreed or strongly agreed with the claim that the agents were fully morally responsible for their actions, only 46% did so among those who said that there was zero chance that the agent would act differently.10 Apparently, a significant proportion of attributions of responsibility to agents in these scenarios were made without full acceptance of their key deterministic aspects.

10  Number of subjects were 282 and 426, respectively. (Based on data set made public by the authors.) Responses could be “strongly disagree,” “disagree,” “somewhat disagree,” “neither agree nor disagree,” “somewhat agree,” “agree,” and “strongly agree.” Studies of folk compatibilism standardly exclude subjects failing comprehension tests for determinism, but questions used for these tests tend to involve close to verbatim repetition of original description of deterministic scenario; the test used by Nadelhoffer at al. arguably required reasoning from deterministic premises.

Experimental Philosophy and Moral Responsibility    503 If this is the correct interpretation of these attributions, they cannot plausibly increase the likelihood of compatibilism, however reliable these judges are in other regards.11 The No Matter What and Indeterminist Intrusion hypotheses targeted compatibilist judgments. Another recent proposal, by Eddy Nahmias and Dylan Murray (2010; Murray, Dylan and Nahmias, Eddy 2014; cf. Nelkin 2007: 255–​256), instead targets incompatibilist judgments. Nahmias and Murray suggest that: Bypassing: A large proportion of people who think that agents in deterministic scenarios lack responsibility do so because they take these agents’ rational agency to be bypassed, playing no role in determining what they do.

Since determinism does not in fact imply that rational agency is bypassed, it follows from Bypassing that a large proportion of incompatibilist judgments are based on a clearly identifiable mistake. The Bypassing hypothesis could potentially explain the Abstract vs. Concrete, the Neurology vs. Psychology, the Actual vs. Possible, and the Framing of Determinism effects already mentioned. Cases that concern the actual world and are framed in concrete psychological terms might naturally recruit ordinary psychological explanatory models in which the effects of antecedent events on actions passes through rather than bypasses rational agency, thus counteracting the mistake. In addition, Nahmias and Murray presented data showing that in experiments varying the ways determinism was framed, incompatibilist judgments were quite strongly correlated with the acceptance of claims like: bypass: In [deterministic scenario], [what a person wants /​what they believe /​their decisions] have no effect on what they end up doing.

Given this, it is natural to assume that variations in ways of presenting determinism give rise to varying degrees of incompatibilist judgments by prompting the bypass mistake to corresponding degrees. Though both striking and replicated in several later experiments, the correlation between incompatibilist and bypass judgments does not in the end support the Bypassing hypothesis.12 First, acceptance of bypass statements does not well explain the acceptance of incompatibilism: The data turns out to be better accounted for by statistical 11  Why do people make this mistake? Interestingly, subjects thinking about deterministic scenarios involving robots rather than human agents were less inclined to make this mistake. Based on statistical analysis of this and subjects’ attributions of free will to human agents and robots, Nadelhoffer et al. suggest that subjects make the error because indeterminism is part of their conception of human agency. This is in line with evidence that indeterminism is an integral part of our understanding of human conscious agency (see, e.g., Deery et al. 2013; Björnsson and Shepherd 2020). 12  Even if it were correct, Bypassing would need to be reconciled with the fact that many subjects seem to be impressed by concerns of determinism when these are distinguished from concerns of conscious voluntary control. See Shepherd 2017.

504   Gunnar Björnsson models taking incompatibilist judgments to explain bypass judgments rather the other way around, or taking both to have a common cause (Rose and Nichols 2013; Björnsson 2014; Björnsson and Pereboom 2014). This undermines the claim that incompatibilist judgments are based on mistaken bypass judgments. Second, later experiments (Björnsson 2014) showed that there was no negative correlation between incompatibilist judgments and the acceptance of claims like: throughpass: In [deterministic scenario], when earlier events cause an agent’s action, they typically do so by affecting what the agent believes and wants, which in turn causes the agent to decide and act in a certain way.

In fact, there was a weak positive correlation between the acceptance of bypass and throughpass statements. This undermines the idea that subjects who make incompatibilist judgments really think that rational agency is bypassed.13 But why, then, do these subjects accept bypass statements? My own suggestion has been that they are gripped by the (arguably correct) thought that determinism implies that the person’s desires, beliefs, and decisions, or more generally their deliberation, has no independent effect on their actions, as it is itself the mere consequence of earlier events: No Independent Effect: Subjects who (a) take determined agents’ deliberation to have no independent effect on their actions and (b) take this to undermine responsibility tend to (c) interpret bypass statements as denying such independent causal influence because such a denial strikes them as particularly interesting (Björnsson 2014; Björnsson and Pereboom 2014).

Since bypass statements are (arguably) true under this interpretation, the No Independent Effect hypothesis gives us no reason to reject incompatibilist judgments as confused. Another suggestion, by David Rose and Shaun Nichols, is in line with the intuitive thought that someone whose deliberation is predetermined to result in a specific intention did not really have a choice and did not really decide the already predetermined outcome: No Choice: People who accept the bypass statements that determined agents’ decisions do not affect their actions do so because they take determinism to imply that agents

13 

It also undermines Fischer’s (2013) suggestion that people make incompatibilist judgments because they confusedly take determinism to undermine guidance control when it only threatens regulative control. (The former sort of control requires the right involvement in the actual causal sequence leading to the object of responsibility; the latter requires that one decides the unfolding of an open future.) Given this suggestion, and given that those who accept throughpass likely think that the agent has guidance control, one would expect a significant negative correlation between incompatibilist judgments and the acceptance of throughpass. The data instead reveal a weak positive correlation, and the acceptance of throughpass among a large majority of those making incompatibilist judgments (Björnsson 2014).

Experimental Philosophy and Moral Responsibility    505 make no choices or decisions. They further take this to imply that these agents are not morally responsible for what they do. (Rose and Nichols 2013)

In support of the No Choice hypothesis, Rose and Nichols’s present data suggesting that subjects’ rejection of the former of the following claims can explain their acceptance of the latter: decisions: In this [deterministic] universe, people make decisions. no effect: In this [deterministic] universe, when people make decisions, what they think and want has no effect on what actions they end up performing.

Potentially, this could ground another error theory for incompatibilist intuitions, given the assumption that there can be such a thing as predetermined decisions. The No Choice explanation is problematic, however. First, most bypass statements in the studies cited before merely deny the effects of what people believe or desire, and do not even mention their decisions. (Even no effect, though concerned with cases involving decisions, is denying effects of what people “think and want,” not of their decisions.) And while there is something intuitive about denying that deterministic agents really have a choice, the thought that they lack beliefs or desires is much less compelling.14 Second, subjects who make incompatibilist judgments are no less willing to accept throughpass than those making compatibilist judgments. This suggests that they generally and easily recognize a sense in which determined agents can make decisions and believe and want things—a compatibilist sense, we might say. Moreover, if they recognize such a sense, they presumably take no effect to be concerned with decisions, beliefs, and desires understood in that sense. The alternative would be that they understand no effect as absurdly concerned with decisions, beliefs, and desires that they think cannot exist in the deterministic universe. (So understood, no effect would be either vacuously true or suffer from a presupposition failure, being akin to “In 2018, when people met Adolf Hitler, the questions they asked had no effect on what answers he gave.” Subjects would presumably prefer an easily available interpretation not suffering from those defects. But if they understand no effect in the compatibilist sense, their rejection of decisions cannot plausibly explain why they accept NO EFFECT. By contrast, the No Independent Effect hypothesis readily explains why incompatibilists tend to accept no effect as well as bypass.15

14 

Chan et al. 2016 ask bypass questions about beliefs and values separately from questions about decisions. Unfortunately, they only provide composite results. Moreover, the fairly weak correlational data presented in the article make it hard to see how bypass judgments can be largely explained by responsibility judgments. 15  It can also be extended to explain why the rejection of decision is correlated with the acceptance of no effect and the rejection of responsibility. Assume, as seems likely, that to “make a decision” can be understood strongly, as requiring that one plays a causally independent role in forming an

506   Gunnar Björnsson The No Matter What, Indeterminist Intrusion, and Bypassing hypotheses represent what I take to be the most promising recent attempts to experimentally establish that a large proportion of either compatibilist or incompatibilist verdicts are epistemically defective. No Matter What failed to identify a process of judgment formation that is uncontroversially epistemically suspect. Bypassing identified such a process, but data contradicted the claim that incompatibilist judgments result from it. Indeterminist Intrusion, finally, identified a process rendering some attributions of responsibility to determined agents suspect, though a large proportion of judgments not suspect in this way were still compatibilist judgments.

4.  Reliability and Conditional Likelihoods Though not discussed in the literature, there is another way in which hypotheses about epistemic processes might be important, one that does not rely on these processes being uncontroversially suspect. Reliability in judgment, remember, is understood as the likelihood of making a correct judgment: of judging that P if P is true and judging that not-​P if not-​P is true. For a claim about an epistemic process to affect the relevant sort of reliability, it is thus enough to make plausible that the process would affect the likelihood of a certain kind of judgment given that a certain answer is true. This makes it easier to avoid begging the question against one answer. Nevertheless, conditional likelihoods can be epistemically highly relevant. Suppose, for example, that we have a plausible account of why many people would reject compatibilism even if it were true, but no plausible account of why many people would reject incompatibilism if it were true. Then the likelihood of a compatibilist judgment given that compatibilism is true might now be .5, whereas the likelihood of an incompatibilist judgment given incompatibilism might be .75. Suppose also that the antecedent likelihood of incompatibilism is .75, that judges are tracking truth independently, and that 58 out of 100 judgments are incompatibilist. Then in spite of the preponderance of incompatibilist judgments and the above chance antecedent likelihood that an incompatibilist judgment will be correct, the compatibilist position would be .99 likely to be correct! This might look surprising, but the explanation is simple: The distribution of answers would be significantly more likely given compatibilism than given incompatibilism, for while it would involve more erroneous judgments, each error would be much more likely.

otherwise undetermined intention, as well as weakly, as merely requiring that one’s deliberation figures in a certain way in the causal sequence leading to that intention. Just as with the “no independent effect” interpretation of the bypass and no effect statements, the stronger interpretation will be more interesting and thus tend to be more salient for subjects who take determinism to rule out independent causal influence and take this to undermine responsibility.

Experimental Philosophy and Moral Responsibility    507 Conditional likelihoods can in principle help provide non-question begging evidence for a position. But to assess the likelihood that judges would get things right if a particular position were true is itself complex. To do so, we need a sense of how judges have access to truth in this domain. This in turn might require as a sense of the facts that would make a certain position rather than another true. On one familiar view, the relevant facts are a priori or conceptual truths, grasped through a priori reflection or conceptual inferential dispositions. On another, stemming from Strawson (1962), the relevant facts involve our practices of holding responsible, primarily understood through participation in and analysis of these practices. Each view suggests different possible sources of error. While experimental philosophy can have a role in assessing these views, it would be in the context of systematic philosophical work.16

5.  Folk Reliability, Difficult Philosophical Questions, and Conceptual Analysis Suppose that we lacked evidence of any particular systematic mistake that would make people no better than chance in their responsibility judgments about deterministic scenarios. Suppose also that we lacked evidence that compatibilist judgments are more likely to be mistaken than incompatibilist judgments, or vice versa. We could still think that nonspecialists ‘judgments are no more likely than chance to get things right. Generally, the compatibility issue seems no different from other central philosophical topics, with its striking yet hard-​to-​pin-​down relations to nontrivial matters such as causation, explanation, laws, control, freedom, rationality, agency, moral wrongness, blame, credit, reactive attitudes, punishment, and desert. Moreover, nonspecialists lack access to the kinds of arguments that philosophers take to be importantly revealing, such as Peter van Inwagen’s (1983) consequence argument, Harry Frankfurt’s (1969) argument against the requirement of alternative possibilities, or Peter Strawson’s (1962) account of the systematic similarities between personal reactive attitudes and practices of holding responsible, to mention some of the most influential (cf. Sommers 2010: 205–​206). Given this complexity of the compatibility question, it might seem implausible that folk judgments would be more reliable than chance.

16 

For my own attempt to explain why, if compatibilism is correct, it would be true in a way given which we should expect the messy distribution of judgments canvassed in section 2, see Björnsson and Persson 2012, 2013; Björnsson 2014, 2016, 2017. For other systematic attempts to incorporate epistemic and semantic reflection in theorizing about free will and moral responsibility, see Double 1996; McCormick 2013; Vargas 2013; Nichols 2015; Deery 2019; Kumar 2019.

508   Gunnar Björnsson Perhaps, though, this conclusion depends on a mistaken view about the subject matter and our access to it. On one understanding, judgments about imagined deterministic cases are based on direct, intuitive knowledge of relatively simple necessary truths about responsibility and understanding of basic features of determinism, rather than knowledge about complex ethical, metaphysical, or empirical matters. Because confusions about the latter factors and an insecure grasp of the relevant concepts downgrade our intuitive access to the relevant facts, philosophical arguments can help us see more clearly. But they are not strictly speaking necessary for epistemic access. (Contrast knowledge of whether incompatibilism is true with knowledge of whether human decision-​making is a deterministic process. The latter is clearly an empirical matter to which nonspecialists have no independent access.17 Incompatibilism does not seem to be inaccessible in that way.) On one popular version of this view, the facts are conceptual truths, and they are accessed by the employment of the concepts involved, concepts possessed by nonspecialists and specialists alike.18 In either case, nonspecialists can access the relevant facts, even if their susceptibility to various sources of noise make their reliability low.19 On many a​ ccounts of how words or elements of thought have their meanings and referents determined—​on many “metasemantic” accunts—we lack privileged access to the nature or necessary features of these referents merely in virtue of our rational capacities and possession of the relevant concepts.20 But even on such accounts, it might seem plausible that people in general can have access to the truth conditions of responsibility judgments. For a recent example, consider Laura and François Schroeter’s (2014) explanation of how people can be concerned with the very same subject matter even if they have very different conceptions of it. The key, they suggest, is that the parties take themselves to be concerned with what they and others have been concerned with at other times, and thus as continuing one and the same “representational tradition.” What property this tradition has been concerned with, if any, is then a matter of what interpretation makes best sense of the tradition as a whole, not about individual judgments based on different conceptions of the subject matter. Conflicting conceptions might in turn be sustained by conflicting attempts to make sense of these traditions, fueled when traditions involve tensions that might be resolved in different ways depending on which of their commitments are seen as more central.21 At least judging from the philosophical

17 

It is of course also an issue not yet resolved by science (see, e.g., Balaguer 2010). For a defense of the potential usefulness of experimental philosophy in revealing conceptual truths, see Balaguer 2016. 19  It is also unclear to what extent experts do better, and with regard to what judgments. For an introductory discussion, see, e.g., Nado 2014. 20  Classical discussions of how our most fundamental ideas about a thing can come apart from what that idea is about include Putnam 1973, 1975; Kripke 1980; Lewis 1984; Millikan 2000. 21  As the Schroeters stress (2014: 16–​18), the best interpretation will at times require disambiguation. Moreover, as Mark Balaguer insists (2016: 2379–​2380), we should allow that people might have inconsistent and thus uninstantiated concepts in mind. But this does not itself mean that every inconsistency in dispositions to apply a concept calls for disambiguation, or mean that the concept to which judges relate is itself inconsistent. 18 

Experimental Philosophy and Moral Responsibility    509 literature, this is very much true about the tradition of attributing responsibility in the context of guiding reactive attitudes and practices of holding responsible, with its notorious tension between compatibilist and incompatibilist or skeptical elements.22 But even with all this variation, it seems that individual judges, qua members of the relevant tradition, will tend to have access to the very aspects of it that a best interpretation should make sense of. It would be odd to antecedently deny that this access makes members better than chance at getting right whether determinism would rule out responsibility.23

6.  Are People Independently Tracking the Truth? Suppose that we can accord folk judgments some reliability—​perhaps a very modest one. Suppose further that considerably more judgments are in line with one theory than with its rival. Then these judgments could provide very strong evidence for that theory in the way sketched in section 1. To do so, however, they must be independently tracking the truth of the matter, in the following sense: Independence: One judgment whether P is (completely) independent of other judgments whether P if and only if (i) the likelihood that it will yield the verdict that P on the assumption that P and (ii) the likelihood that it will yield the verdict that not-​P on the assumption that not-​P are unaffected by what the verdicts of the other judgments are.

Independence, so understood, would most obviously be undermined if we knew that some judgments would causally influence others, perhaps because some judges took others to be authorities on the issue. For an extreme example, suppose that we knew that only one judge would make a judgment based on inconclusive evidence and everyone else would merely try to copy that judgment (with fairly high reliability). Then the likelihood that one particular copycat judgment, J, would yield the verdict that P on the assumption that P would depend on what we assume that the other copycat judgments are. 22 

Cf. Vargas 2006; 2013, part I. A possibility here is that different ways of resolving tensions might be more or less appropriate depending on context, leaving room for even more variations in judgment (cf. Nichols 2015). 23  Not all metasemantic theories are neutral with respect to compatibilism and incompatibilism. For example, consider teleosemantic accounts like that of Ruth Millikan’s (1984, 2017), which take the referent of a concept to be a matter of what the mechanisms involved in employing that concept have as their etiological function to recognize. For mechanisms that produce judgments that regularly guide action and communication in what people take to be successful ways, that function is not plausibly to track some metaphysically esoteric condition without systematic effects on the practice in which those mechanisms are employed. Cf. Björnsson and Persson 2012: 345–​346; Vargas 2013, ch. 3; Deery 2019.

510   Gunnar Björnsson For these judgments, especially in aggregate, would tell us something about the original judgment from which it would be copied and about the evidence on which the original judgment was based. Thus, if we assume that P and that all the other copycats judged that P, it would be very likely that J would yield the verdict that P, but if we assumed that P and that all the other copycats judged that not-​P, it would be very likely that J would yield the verdict that not-​P. Moreover, it is clear that even if all judgments agreed and the original judgment was reasonable in light of the evidence, the aggregate reliability of all judgments would be no higher than the reliability of the evidence, and the aggregate reliability of the copycat judgments would be no higher than the reliability of the original judgment.24 Admittedly, causal influence of some judgments on others might not be an urgent worry for the sorts of studies that we have looked at. Subjects are often recruited from large online pools rather than from small socially connected groups, and even when subjects are socially connected, the sorts of judgments reported in the experiments are not characteristically socially shared. But it seems highly likely that many judges have access to the same kind of evidence. Many might rely on intuitive reactions to the cases that confront them in the studies prompted by similar cognitive processes, and many will have a similar experience of holding and being held responsible, and of various appeals to excuses and exemptions. Even if there is a richness and variation in the evidence on which people base their judgments, there is bound to be a great deal of overlap. It could be, then, that although more judgments will tend to bring higher accumulated reliability, the marginal epistemic utility might fall quickly, as more judges would be increasingly unlikely to bring new sources of reliability to the table rather than more of the same. In light of this, and in light of the fact that professional philosophers have access to evidence in addition to what they share with laypeople, the wisdom of large crowds of nonspecialists might add little to the judgments of specialists.

7.  Concluding Discussion Depending on the subject matter, a preponderance of judgments in favor of one answer might mean that it is extremely likely to be correct. We have seen several reasons to doubt that it applies to the central philosophical question discussed here, namely that of compatibilism. We can perhaps assume, as most people in the literature do, that the issue of whether compatibilism is true has one correct answer, and we might well accord people some antecedent reliability. Pending further arguments, however, the reliability is presumably low, as the questions are difficult and people seem to be fairly evenly divided in their judgments. More importantly, it is unclear to what extent we should

24 

For discussion of the independence requirement, see, e.g., Estlund 1994; Dietrich 2008; Talbot 2014; Goodin and Spiekermann 2018, ch. 5.

Experimental Philosophy and Moral Responsibility    511 assume that people’s judgments are independently tracking the truth. If they are not, their individual reliabilities fail to aggregate. This mostly negative conclusion is admittedly concerned with one specific and notoriously thorny philosophical issue, but it is easy to see how it generalizes to other similarly difficult questions. Moreover, much of what is said also applies if we turn from appeals to the wisdom of crowds of nonspecialists to corresponding appeals to the judgments of specialists. Going by the raw numbers of compatibilists among specialists in action theory (see Bourget and Chalmers 2009, 2014) and assuming a modest reliability and independence and an antecedently equal likelihood for compatibilism and incompatibilism, compatibilism would be highly likely to be correct, whereas libertarianism and free will skepticism would be extremely unlikely. In response, one might raise doubts about whether the questions behind the numbers track compatibilism proper.25 One might also resist the conclusions by pessimistically denying that specialists are even modestly reliable. And, less pessimistically, one can deny the independence of their judgments. As with laypeople, philosophers’ reliability plausibly depends on their access to largely overlapping sources of evidence mostly falling into a few large groups whose reliabilities are internally strongly interdependent.26 The worries raised here have concerned the idea that when a preponderance of judgments by nonspecialists fall in line with one philosophical position, this can provide strong aggregative evidence for that position. But studies of folk judgments might help us understand responsibility in less direct ways. First, they might serve as a check on philosophical group think. If one suspects that the field ignores an important strand of ordinary thinking about responsibility, experiments involving non-​philosophers can tell us whether it is indeed an important part of the practice that needs to be accounted for or explained away. Second, although the No Matter What and Bypassing hypotheses did not hold up in further studies, Indeterminist Intrusion currently seems well supported, and new experiments might reveal further reasons why certain judgments are particularly unreliable and ways in which compatibilists or incompatibilists can explain away judgments that do not accord with their theory. (Studies of philosophers’ judgments could reveal the same, calling for methodological self-​reflection.) Third, studies might reveal widespread dispositions of judgment and relations between different kinds of judgments that tell us something about the nature of the practice.27 But, as noted in the introduction, neither of these ways for experimental philosophy to be helpful bypasses

25  Subjects were asked about free will rather than responsibility, and some philosophers who call themselves compatibilists have in mind a kind of responsibility that few self-identified incompatibilists deny. 26  For discussion of how dependence undermines aggregation of evidential weight among philosophers, see Talbot 2014. 27  See, e.g., (a) Victor Kumar’s (2019) account of responsibility and associated retributive practices, based in part on Cushman 2008; (b) Karl Persson’s and my appeal to empirical studies in support of a certain account of the nature of responsibility judgments (Björnsson and Persson 2012, 2013); and (c) Shaun Nichol’s (2015) account of retributive practices and deterministic and incompatibilist intuitions.

512   Gunnar Björnsson the need to attend to more ordinary philosophical arguments and systematizing efforts. Though it can be a useful tool, it is no shortcut.

Acknowledgments I thank Manuel Vargas, Christopher Franklin, Shaun Nichols, Mark Balaguer, Martin Peterson, Kai Spiekermann, Dana Nelkin, Derk Pereboom, Bob Hartman, Björn Eriksson, and participants at the Stockholm Higher Seminar in Practical Philosophy for valuable input. Work on this paper was funded by the Swedish Research Council [2015-​01488].

References Alexander, Joshua, Mallon, Ronald, and Weinberg, Jonathan. 2010. “Accentuate the Negative.” Review of Philosophy and Psychology 1, pp. 297–​314. Andow, James and Cova, Florian. 2016. “Why Compatibilist Intuitions Are Not Mistaken: A Reply to Feltz and Millan.” Philosophical Psychology 29, pp. 550–​566. Balaguer, Mark. 2010. Free Will as an Open Scientific Problem. Cambridge, MA: MIT Press. Balaguer, Mark. 2016: “Conceptual Analysis and X-​Phi.” Synthese 193, pp. 2367–​2388. Björnsson, Gunnar 2014. “Incompatibilism and ‘Bypassed’ Agency.” In Surrounding Free Will, Alfred Mele (ed.). New York: Oxford University Press, pp. 95–​122. Björnsson, Gunnar 2016. “Outsourcing the Deep Self: Deep Self Discordance Does Not Explain Away Intuitions in Manipulation Arguments.” Philosophical Psychology 29, pp. 637–​653. Björnsson, Gunnar. 2017. “Explaining Away Epistemic Skepticism About Culpability.” In Oxford Studies in Agency and Responsibility, David Shoemaker (ed.). Oxford: Oxford University Press, pp. 141–​164. Björnsson, Gunnar and Pereboom, Derk 2014: “Free Will Skepticism and Bypassing.” In Moral Psychology: Free Will and Moral Responsibility, Walter Sinnott-​Armstrong (ed.). Cambridge, MA: MIT Press, pp. 27–​35. Björnsson, Gunnar and Pereboom, Derk 2016. “Traditional and Experimental Approaches to Free Will and Moral Responsibility.” In Justin Sytsma & Wesley Buckwalter (eds.), Companion to Experimental Philosophy. Blackwell: Wiley, pp. 142–​157. Björnsson, Gunnar and Persson, Karl 2012. “The Explanatory Component of Moral Responsibility.” Noûs 46, pp. 326–​354. Björnsson, Gunnar and Persson, Karl 2013. “A Unified Empirical Account of Responsibility Judgments.” Philosophy and Phenomenological Research 87, pp. 611–​639. Björnsson, Gunnar and Shepherd, Joshua. 2020. “Determinism and Attributions of Consciousness.” Philosophical Psychology 33, pp. 549–​568. Bourget, David and Chalmers, David J. 2009. “The Philpapers Survey,”. https://​philpapers.org/​ surveys/​. Bourget, David and Chalmers, David J. 2014. “What Do Philosophers Believe?” Philosophical Studies 170, pp. 465–​500. Chan, Hoi‐Yee, Deutsch, Max, and Nichols, Shaun 2016. “Free Will and Experimental Philosophy.” In A Companion to Experimental Philosophy. Justin Sytsma, Wesley Buckwalter (eds.). New York, pp. 158–​172.

Experimental Philosophy and Moral Responsibility    513 Clark, Cory J., Winegard, Bo M., and Baumeister, Roy F. 2019. “Forget the Folk: Moral Responsibility Preservation Motives and Other Conditions for Compatibilism.” Frontiers in Psychology 10. Cova, Florian, Strickland, Brent, Abatista, Angela, Allard, Aurélien, Andow, James, Attie, Mario, Beebe, James, Berniūnas, Renatas, Boudesseul, Jordane, Colombo, Matteo, Cushman, Fiery, Diaz, Rodrigo, N’Djaye Nikolai van Dongen, Noah, Dranseika, Vilius, Earp, Brian, Torres, Antonio Gaitán, Hannikainen, Ivar, Hernández-​Conde, José V., Hu, Wenjia, Jaquet, François, Khalifa, Kareem, Kim, Hanna, Kneer, Markus, Knobe, Joshua, Kurthy, Miklos, Lantian, Anthony, Liao, Shen-​yi, Machery, Edouard, Moerenhout, Tania, Mott, Christian, Phelan, Mark, Phillips, Jonathan, Rambharose, Navin, Reuter, Kevin, Romero, Felipe, Sousa, Paulo, Sprenger, Jan, Thalabard, Emile, Tobia, Kevin, Viciana, Hugo, Wilkenfeld, Daniel, and Zhou, Xiang. 2018. “Estimating the Reproducibility of Experimental Philosophy.” Review of Philosophy and Psychology (1):1–​36. Cushman, Fiery. 2008. “Crime and Punishment: Distinguishing the Roles of Causal and Intentional Analyses in Moral Judgment.” Cognition 108, pp. 353–​380. De Brigard, Felipe, Mandelbaum, Eric, and Ripley, David. 2009. “Responsibility and the Brain Sciences.” Ethical Theory and Moral Practice 12, pp. 511–​524. Deery, Oisín. 2019. “Free Actions as a Natural Kind.” Synthese 198 (1):823–​843. Deery, Oisín, Bedke, Matthew S., and Nichols, Shaun. 2013. “Phenomenal Abilities: Incompatibilism and the Experience of Agency.” In Oxford Studies in Agency and Responsibility, David Shoemaker (ed.). Oxford: Oxford University Press, pp. 126–​150. Dietrich, Franz. 2008. “The Premises of Condorcet’s Jury Theorem Are Not Simultaneously Justified.” Episteme 5, pp. 56–​73. Double, Richard. 1996. Metaphilosophy and Free Will. Oxford: Oxford University Press. Estlund, David M. 1994. “Opinion Leaders, Independence, and Condorcet’s Jury Theorem.” Theory and Decision 36, pp. 131–​162. Feltz, Adam and Cova, Florian. 2014. “Moral Responsibility and Free Will: A Meta-​Analysis.” Consciousness and Cognition 30, pp. 234–​246. Feltz, Adam and Millan, Melissa. 2015. “An Error Theory for Compatibilist Intuitions.” Philosophical Psychology, 28(4): pp. 529–​555. Fischer, John Martin. 2013.“The Frankfurt Style Cases: Philosophical Lightning Rods.” In In Free Will and Moral Responsibility, Ish Haji and Justin Caouette, Justin (eds.). Newcastle: Cambridge Scholars Publishing, pp. 43–​57. Frankfurt, Harry G. 1969. “Alternate Possibilities and Moral Responsibility.” Journal of Philosophy 66, pp. 829–​839. Goodin, Robert E. and Spiekermann, Kai. 2018. An Epistemic Theory of Democracy. Oxford: Oxford University Press. Jeppsson, Sofia. 2016. “Accountability, Answerability, and Freedom.” Social Theory & Practice 42, pp. 681–​705. Kauppinen, Antti. 2007. “The Rise and Fall of Experimental Philosophy.” Philosophical Explorations 10, pp. 95–​118. Kneer, Markus and Machery, Edouard. 2019. “No Luck for Moral Luck.” Cognition 182, pp. 331–​348. Knobe, Joshua. 2007. “Experimental Philosophy and Philosophical Significance.” Philosophical Explorations: An International Journal for the Philosophy of Mind and Action 10, pp. 119–​121. Knobe, Joshua and Nichols, Shaun. 2017. “Experimental Philosophy.” In The Stanford Encyclopedia of Philosophy, Edward N. Zalta (ed.). URL= https://​plato.stanford.edu/​ archives/​win2017/​entries/​experimental-​philosophy/​.

514   Gunnar Björnsson Kripke, Saul A. 1980. Naming and Necessity. Cambridge, MA: Harvard University Press. Kumar, Victor. 2019.” Empirical Vindication of Moral Luck.” Noûs 53 (4):987–​1007. Lewis, David. 1984.”Putnam’s Paradox.” Australasian Journal of Philosophy 62, pp. 221–​236. List, Christian. 2004. “On the Significance of the Absolute Margin.” British Journal for the Philosophy of Science 55, pp. 521–​544. Ludwig, Kirk. 2017. “Thought Experiments in Experimental Philosophy.” In Routledge Companion to Thought Experiments, Mike Stuart, James Robert Brown, and Yiftach J. H. Fehige (eds.), New York: Routledge, pp. 385–​405. Mallon, Ron. 2016. “Experimental Philosophy.” Oxford Handbook of Philosophical Methodology, ed. Herman Cappelen, John Hawthorne, and Tamar Szabó Gendler., pp. 410–​433. Mandelbaum, Eric and Ripley, David. 2012.”Explaining the Abstract/​Concrete Paradoxes in Moral Psychology: The Nbar Hypothesis.” Review of Philosophy and Psychology 3, pp. 351–​368. McCormick, Kelly Anne. 2013. “Anchoring a Revisionist Account of Moral Responsibility.” Journal of Ethics and Social Philosophy 7, pp. 1–​20. Mele, Alfred. 2014. “Free Will and Substance Dualism: The Real Scientific Threat to Free Will?.” In Moral Psychology, vol. 4: Free Will and Responsibility. Walter Sinnott-​Armstrong (ed.). Cambridge, MA: MIT Press, pp. 195–​207. Millikan, Ruth Garrett. 1984. Language, Thought, and Other Biological Categories: New Foundations for Realism. Cambridge, MA: MIT Press. Millikan, Ruth Garrett. 2000. On Clear and Confused Ideas: An Essay About Substance Concepts. New York: Cambridge University Press. Millikan, Ruth Garrett. 2017. Beyond Concepts: Unicepts, Language, and Natural Information. Oxford: Oxford University Press. Murray, Dylan and Nahmias, Eddy. 2014. “Explaining Away Incompatibilist Intuitions.” Philosophy and Phenomenological Research 88, pp. 434–​467. Nadelhoffer, Thomas, Rose, David, Buckwalter, Wesley and Nichols, Shaun. 2020. “Natural Compatibilism, Indeterminism, and Intrusive Metaphysics.” Cognitive Science 44 (8). Nado, Jennifer. 2014. “Philosophical Expertise.” Philosophy Compass 9, pp. 631–​641. Nahmias, Eddy, Coates, D. Justin, and Kvaran, Trevor. 2007.”Free Will, Moral Responsibility, and Mechanism: Experiments on Folk Intuitions.” Midwest Studies in Philosophy 31, pp. 214–​242. Nahmias, Eddy, Morris, Stephen, Nadelhoffer, Thomas, and Turner, Jason. 2005. “Surveying Freedom: Folk Intuitions About Free Will and Moral Responsibility.” Philosophical Psychology 18, pp. 561–​584. Nahmias, Eddy and Murray, Dylan. 2010. “Experimental Philosophy on Free Will: An Error Theory for Incompatibilist Intuitions.” In New Waves in Philosophy of Action, Jesús Aguilar, Andrei Buckareff, and Keith Frankish (eds.). UK, London: Palgrave Macmillan, pp. 189–​216. Nelkin, Dana K. 2007. “Do We Have a Coherent Set of Intuitions About Moral Responsibility?” Midwest Studies in Philosophy 31, pp. 243–​259. Nichols, Shaun. 2004. “The Folk Psychology of Free Will: Fits and Starts.” Mind & Language 19, pp. 473–​502. Nichols, Shaun. 2015. Bound: Essays on Free Will and Responsibility. Oxford: Oxford University Press. Nichols, Shaun and Knobe, Joshua. 2007. “Moral Responsibility and Determinism: The Cognitive Science of Folk Intuitions.” Noûs 41, pp. 663–​685. Paolacci, Gabriele and Chandler, Jesse. 2014. “Inside the Turk: Understanding Mechanical Turk as a Participant Pool.” Current Directions in Psychological Science 23, pp. 184–​188.

Experimental Philosophy and Moral Responsibility    515 Putnam, Hilary. 1973. “Meaning and Reference.” Journal of Philosophy 70, pp. 699–​7 11. Putnam, Hilary. 1975. “The Meaning of ‘Meaning.” In Mind, Language and Reality. Cambridge: Cambridge University Press, pp. 215–​271. Rose, David and Nichols, Shaun. 2013. “The Lesson of Bypassing.” Review of Philosophy and Psychology 4, pp. 599–​619. Roskies, Adina L. and Nichols, Shaun. 2008. “Bringing Responsibility Down to Earth.” Journal of Philosophy 105, pp. 371–​388. Sarkissian, Hagop, Chatterjee, Amita, Brigard, Felipe De, Knobe, Joshua, Nichols, Shaun, and Sirker, Smita. 2010. “Is Belief in Free Will a Cultural Universal?” Mind & Language 25, pp. 346–​358. Schroeter, Laura and Schroeter, François. 2014. “Normative Concepts: A Connectedness Model.” Philosophers’ Imprint 14, pp. 1–​26. Shepherd, Joshua. 2015. “Consciousness, Free Will, and Moral Responsibility: Taking the Folk Seriously.” Philosophical Psychology 28, pp. 929–​946. Shepherd, Joshua. 2017. “The Folk Psychological Roots of Free Will.” In Experimental Metaphysics, David Rose (ed.). London: Bloomsbury Academic, pp. 95–​116. Shoemaker, David. 2015. Responsibility from the Margins. Oxford: Oxford University Press. Smith, Angela M. 2015. “Responsibility as Answerability.” Inquiry: An Interdisciplinary. Journal of Philosophy 58 (2):99–​126. Sommers, Tamler. 2010. “Experimental Philosophy and Free Will.” Philosophy Compass 5, pp. 199–​212. Sripada, Chandra Sekhar. 2012. “What Makes a Manipulated Agent Unfree?” Philosophy and Phenomenological Research,85, pp. 563–​593. Strawson, Peter F. 1962. “Freedom and Resentment.” Proceedings of the British Academy 48, pp. 187–​211. Sytsma, Justin and Buckwalter, Wesley. 2016. A Companion to Experimental Philosophy. John Wiley & Sons. Talbot, Brian. 2014. “Why So Negative? Evidence Aggregation and Armchair Philosophy.” Synthese,191, pp. 3865–​3896. Taylor, Matthew C. and Maranges, Heather M. 2020. “Are the Folk Historicists About Moral Responsibility?” Philosophical Psychology, 33 (1):1–​22. van Inwagen, Peter. 1983. An Essay on Free Will. Oxford: Clarendon. Vargas, Manuel. 2006. “Philosophy and the Folk: On Some Implications of Experimental Work for Philosophical Debates on Free Will.” Journal of Cognition and Culture 6, pp. 239–​254. Vargas, Manuel. 2013. Building Better Beings: A Theory of Moral Responsibility. Oxford: Oxford University Press. Vonasch, Andrew J., Baumeister, Roy F., and Mele, Alfred R. 2018. “Ordinary People Think Free Will Is a Lack of Constraint, Not the Presence of a Soul.” Consciousness and Cognition 60, pp. 133–​151. Watson, Gary. 1996. “Two Faces of Responsibility.” Philosophical Topics 24, pp. 227–​248. Weinberg, Jonathan. 2015.” The Methodological Necessity of Experimental Philosophy.” Discipline Filosofiche 25, pp. 23–​42.

PA RT X

R E SP ON SI B I L I T Y, R E L AT ION SH I P S , AND MEANING IN LIFE

CHAPTER 25

Mor a l Resp onsi bi l i t y a nd E x istential At t i t u de s Paul Russell

Philosophy, and in particular moral philosophy, is still deeply attached to giving good news.—​Bernard Williams

1.  Existential Attitudes and the Human Predicament Our understanding and interpretation of human agency and moral responsibility is evidently not just a matter of theoretical interest and curiosity. It affects and modifies our attitude to our lives and how we may lead them. In this respect the philosophical issue of human freedom and moral responsibility is similar to the problems of the existence of God and the immortality of the soul. These are, as Kant noted, issues that concern not just what we know but what we ought to do and what we may hope for.1 We care about these matters because they serve to guide and inform how we lead our lives and why we may want to continue them. With respect to these matters, therefore, it is understandable that we may seek some measure or degree of philosophical consolation and reassurance. It is, nevertheless, entirely possible that the truth about our predicament will disappoint these hopes and aspirations. We may refer to metaphysical problems that carry significance for the human predicament as existential metaphysical problems. In confronting problems of this sort, we seek to provide some sort of vindication of our lives and existence—​both as 1 

Kant, Immanuel. 1781/​1965: II, ii, ch.2 (esp. p. 635).

520   Paul Russell individuals and collectively. The challenge is “to make sense of human life” within a credible worldview. These problems share several important features. First, they generally arouse attitudes that are either positive or negative in quality. They are positive insofar as they affirm or sustain some feature or aspect of our existence that we value or care about. Conversely, they are negative when they have the opposite significance. We might refer to this as the valence of our existential attitudes. Second, this valence may also vary as a matter of degree—​being more or less intense. The intensity of a given attitude will vary with either the weight and significance of what it is that we attach value to or with the extent to which it is perceived to be threatened (e.g., in light of skeptical arguments). Finally, when we consider existential metaphysical issues of this kind, we may distinguish between an orthodox and unorthodox interpretation of the relationship between skepticism and pessimism. Generally speaking, we assume that skepticism about the feature in question justifies some measure or degree of pessimism, on the ground that the feature in question is required to secure something of particular value or worth for us. It is clear, nevertheless, that skeptical conclusions do not always generate pessimism, as some may accept a given skeptical conclusion but go on to deny that this licenses any form of pessimism. As we shall see, when we consider the issue of moral responsibility, the general structure of our available options falls into this general structure or pattern, as found in various other existential metaphysical problems.

2.  Strawson’s Taxonomy Our primary task, in this chapter, is to provide an overview of the available positions relating to existential attitudes and moral responsibility. For this purpose, as it concerns the contemporary debate, the right place to begin is with P. F. Strawson’s enormously influential discussion in “Freedom and Resentment” (first published in 1962). In this paper Strawson introduces the labels of “optimism” and “pessimism” as a way of categorizing the opposing parties in the debate concerning the relationship between moral responsibility and the thesis of determinism. He describes the “Pessimists” as those who hold that if the thesis of determinism is true, then the concepts of moral obligation and responsibility really have no application, and the practices of punishing and blaming, of expressing moral condemnation and approval are really unjustified (Strawson 1962/​2013: 63). The “Optimists,” however, “hold that these concepts and practices in no way lose their raison d’être if the thesis of determinism is true” (Strawson 1962/​2013: 63). (Hereafter I will use capital letters when citing the specific labels that Strawson employs.) Strawson also distinguishes a third party in this debate, “the genuine moral skeptic,” who takes the view “that the notions of moral guilt, of blame, of moral responsibility are inherently confused and that we can see this to be so if we consider the consequences of either the truth of determinism or its falsity” (Strawson 1962/​2013: 63).

Moral Responsibility and Existential Attitudes    521 Strawson’s tripartite analysis is in some respects confusing and problematic. First, there is the question concerning the relationship between the Pessimist and the Skeptic. The Pessimist, as Strawson usually presents this position, is a libertarian Pessimist, who believes that determinism is false and that we do have free will. Both the libertarian Pessimist and the Skeptic are incompatibilists. The Skeptic, however, denies that we have free will, whether determinism is true or false. The Skeptic, therefore, rejects libertarian metaphysics (e.g., “contra-​causal freedom” of some kind) and maintains that there are no free and responsible moral agents—​which is certainly a dramatic conclusion. Since the libertarian Pessimist is not a Skeptic—​and believes that humans are moral agents who have free will—​there is an important sense in which the libertarian Pessimist is actually optimistic. Clearly, then, both the libertarian Pessimist and the (compatibilist) Optimist share the (optimistic) view that we enjoy free will of the relevant kind that supports moral responsibility. Where they differ, is that the Optimist believes that this does not require the falsity of determinism, whereas the libertarian Pessimist believes that it does. Where, then, does Strawson stand in relation to the Optimist/​Pessimist divide? Contrary to the Pessimist, Strawson rejects the suggestion that the truth of determinism implies skepticism about moral responsibility. Strawson’s argument for this is complex and not all of its details need concern us here, but the crucial element he relies on to defeat skepticism is that of the “reactive attitudes” or “moral sentiments,” along with a theory of excuses and exemptions. According to Strawson’s account, the foundations of responsibility rest with the fact that we care about the attitudes and intentions of other agents and the extent to which they do or do not manifest “some degree of good will or regard” (Strawson 1962/​2013: 66–​67). Reactive attitudes are, it is argued, “natural human reactions to the good or ill will or indifference of others towards us, as displayed in their attitudes and actions” (Strawson 1962/​2013: 70—​ Strawson’s emphasis). Our proneness to such reactions is exactly what is involved in making these demands—​such that when they are violated we naturally respond with feelings of indignation, resentment, and blame. It is, according to Strawson, these fundamental features of human moral psychology that are essential if we are to make sense of the conditions of moral responsibility and the kind of justification that it may require (Strawson 1962/​2013: 64). Understanding responsibility in terms of the reactive attitudes serves to explain why skepticism of this general kind would be both unlivable and unbearable in practice. Strawson’s key arguments against skepticism turn on the rationale of excuses and exemptions. Nothing about the (assumed) truth of determinism implies that our reactive attitudes are never justified. In respect of excuses, such as ignorance, accidents, or physical force, none of these considerations apply generally simply because determinism is true. Similarly, in the case of exemptions, where we regard a person as an inappropriate target of reactive attitudes, there is nothing about determinism that suggests that all agents are somehow insane, immature, or severely disabled. In circumstances of these kinds, where agents are morally incapacitated, we are required to systematically withdraw our reactive attitudes. This involves dropping the participant

522   Paul Russell stance and adopting a sustained attitude of “objectivity” (Strawson 1962/​2013: 69–​72). An attitude of this kind, Strawson maintains, requires us to view the agent “as excluded from ordinary adult human relationships” and as an individual who is to be “managed or handled or cured or trained”—​someone we may regard in a more detached manner and as not belonging to “the moral community” (Strawson 1962/​2013: 76, 79, 80). If the truth of the thesis of determinism implied that we should always, universally, adopt “a sustained objectivity of inter-​personal attitude” that would, indeed, be deeply disturbing. However, not only is this not required of us, any attempt to systematically adopt the objective attitude in this way is psychologically impossible for human beings (Strawson 1962/​2013: 71–​73). Let us now turn to Strawson’s particular understanding of the relationship between skepticism and pessimism.2 Although Strawson believes that the relevant source of pessimism rests with the threat of skepticism, and that skepticism can be discredited, it is also clear that he distances himself from Optimism (i.e., classical compatibilism) in crucial respects. The Optimist, as Strawson describes this position, places emphasis on “the efficacy of our practices of punishment, and of moral condemnation and approval, in regulating behaviour in socially desirable ways” (Strawson 1962/​2013: 64; see also 65, 78–​ 80). This pragmatic, utilitarian perspective is not, Strawson maintains, a sufficient basis or even the right sort of basis on which to regulate and justify these attitudes and practices (Strawson 1962/​2013: 65). What it leaves out is the “vital” role of the reactive attitudes. The Pessimist, Strawson argues, is correct in claiming that there is a “lacuna in the Optimist’s story” but mistakenly tries and plug this gap with the extravagant and implausible metaphysics of libertarianism. In this respect, both Optimists and Pessimists are guilty of the shared mistake of “over-​intellectualizing the facts” and overlooking or ignoring the crucial role of reactive attitudes or moral sentiments in describing the real basis of these problematic concepts and practices. Viewed this way, Strawson’s basic objection to the Optimist’s position is that the stance they adopt is in important respects like that of the Skeptic—​viewing agents in a detached manner and as mere objects for treatment, control, and manipulation. Strawson agrees with the Pessimist that the Optimist outlook, with its “one-​eyed utilitarian” emphasis on controlling future behavior, is dehumanizing and to a considerable extent shares the unattractive features of the Skeptic’s outlook (Strawson 1962/​2013: 80). From the perspective of Strawson’s critique, an exclusively objective stance is not a basis for optimism of any kind but rather a bleak and humanly impoverished form of existence. Even if we could embrace it—​and Strawson questions this assumption—​we have every reason to resist it (Strawson 1962/​2013: 72–​73). In this way, Strawson’s modified optimism rests with his effort to preserve a central role for reactive attitudes in human

2  Strawson’s arguments have, of course, been criticized in a number of different respects which cannot all be reviewed here. One place to begin is McKenna and Russell 2008. See also Russell 1992; and Russell 2010.

Moral Responsibility and Existential Attitudes    523 moral life, as opposed to a detached stance of controlling and manipulating agents using the techniques of (behavioral) science.3 There are, as already noted, many possible objections and criticisms that could be leveled against Strawson’s own brand of (compatibilist-​friendly) optimism. Two criticisms are, however, of particular interest for this discussion. The first is that Strawson both exaggerates the pessimistic implications of skepticism and, at the same time, underestimates the significant sources of pessimism that persist for his own account. With respect to the first of these, Strawson maintains that if we were to accept the case for skepticism, and cease entirely to entertain reactive attitudes, this would result in the universal adoption of the objective attitude. A world that operated exclusively on the basis of the objective attitude would, he claims, be a bleak, inhuman world, stripped of all personal relations and emotional ties. Strawson’s critics have argued that his understanding of the reactive attitudes, and their relation to the moral and personal emotions, distorts and exaggerates the problems that we may face here.4 It is not obvious, for example, that the reactive attitudes should be taken to comprehend all the various emotions involved in personal relationships (e.g., love and friendship). If this is correct, then abandoning the reactive attitudes need not imply the loss of all personal relations and emotional attachments. Beyond this, it may be further argued that not all reactive attitudes would be discredited or undermined by skeptical arguments, since some may be retained or amended, consistent with skeptical commitments. In a similar vein, it may be argued that not only are there some moral emotions that are not reactive attitudes (e.g., moral admiration), there may also be reactive attitudes that do not involve moral expectations or standards. In all these ways, it may be argued that Strawson’s effort to link skepticism with a pessimism based on the consequences of universally adopting the objective attitude is too simple and exaggerates the challenges that the skeptic faces. On the other side of the coin, critics have argued that Strawson’s effort to “reconcile” the Optimist and the Pessimist (Strawson 1962/​2013: 63), and provide a suitably qualified defense of optimism on this subject, underestimates persisting sources of pessimism for his own position. In particular, there are issues of “history” and “luck” that lie at the heart of the Pessimist outlook, which Strawson fails to properly address and engage with. This line of criticism has been strongly pressed by, among others, Gary Watson. Although Watson is not entirely unsympathetic to Strawson’s (“expressivist”) approach, he argues that Strawson neglects what may be described as “the historical dimension of the concept of responsibility” (Watson 1987/​2013: 106–​107). We 3 

The Optimist who Strawson specifically cites is Nowell-​Smith (1948) but there is good reason to suppose that his most obvious target, in this context, is Moritz Schlick (1939), who presented a particularly influential statement of the classical compatibilist position. It was Schlick’s view that free will is “a pseudo-​problem” and that responsibility can be reduced to the question of when rewards and punishments are effective “educative measures”—​a view that falls squarely into Strawson’s sights. Another important figure in the context is B. F. Skinner (1948), a leading figure in behavioral psychology. Skinner’s views are briefly discussed further in note 8. 4  See, e.g., Wallace 1994: esp. ch. 2; Pereboom 2001: esp. 199–​207; and Sommers 2007.

524   Paul Russell can appreciate this dimension of the problem, Watson suggests, if we examine some real and specific cases, where an individual’s upbringing and formative experiences have brutalized them and turned them into vicious criminals. When we consider the crimes of these individuals our reactive attitudes will, no doubt, be strongly aroused—​ especially in cases where none of the usual excusing or exempting considerations that Strawson has mentioned seem to apply. Nevertheless, when we turn our attention to the background conditions that shaped the criminal’s character and conduct, Watson suggests, we feel ambivalence and conflict, as we come to see the criminal not just as a victimizer but also as a victim (Watson 1987/​2013: 101). Our awareness of moral luck, in cases of this kind, can be generalized to all cases, showing that our reactive attitudes may be more sensitive to historical considerations—​and skeptical pressures—​than Strawson acknowledges.

3.  Dispelling the Bugbears and Bogeymen—​Dennett’s Revisionary Optimism A very different approach to the problem of moral responsibility is presented in Daniel Dennett’s Elbow Room, a work that was published more than two decades after Strawson’s “Freedom and Resentment.” Dennett’s work has little to say about reactive attitudes and, instead, offers an account of the importance of rational self-​control for our understanding of these matters. There are two closely related themes woven together in Dennett’s discussion. First, he aims to provide an account of what it is with respect to free will that we value and care about and to show that we do possess those abilities and qualities that are worth wanting. Second, he aims to unmask and debunk the various skeptical arguments and analogies that attempt to show that our self-​image as free and responsible agents is in some way illusory or mistaken. The first task supports his optimism and the second discredits the pessimism that lurks behind these skeptical arguments. His conclusions, he says, are “neither revolutionary nor pessimistic.” “They are,” he continues, “only moderately revisionary: the common wisdom about our place in the universe is roughly right. We do have free will. We can have free will and science too” (Dennett 1984: 19). Dennett’s optimistic conclusion that free will is not an illusion rests on a clear statement of what it is that we care about or “deem worth wanting” (Dennett 1984: 153). What we want when we want free will is the power to decide our courses of action, and to decide them wisely, in light of our expectations and desires. We want to be in control of ourselves, and not under the control of others. We want to be agents, capable of initiating, and taking responsibility for, projects and deeds. All this is ours [Dennett has] tried to show, as a natural product of our biological endowment,

Moral Responsibility and Existential Attitudes    525 extended and enhanced by our initiation into society. (Dennett 1984: 169—​my emphasis)

The real threats to human freedom, Dennett maintains, are not metaphysical (e.g., determinism) but threats we face due to control, manipulation, and coercion by others. These real threats to human freedom do limit and erode our responsibility. We can, nevertheless, still distinguish between “responsible moral agents and beings with diminished or no responsibility”—​a distinction that “plays a crucial role in the quality and meaning of our lives” (Dennett 1984: 155). When it comes to holding others responsible, or taking responsibility for our own conduct and activities, we do not need to rely on some absolutist conception of (God-​like) “Cosmic Responsibility” (Dennett 1984: 158). On the contrary, we can make perfect sense of our associated attitudes and practices (e.g., punishment) in terms of the way they serve to discourage actions and activities that cause unwanted harms and injuries (Dennett 1984: 158–​64). In this respect, Dennett’s system remains firmly committed to most of the pragmatic, forward-​looking features of the Optimistic model that Strawson had earlier criticized.5 What is distinctive about Dennett’s approach, and constitutes a substantial advance on earlier statements of Optimism or classical compatibilism, is his account of freedom in terms of our ability to recognize and respond to reasons (Dennett 1984: ch. 2). Once we appreciate the way in which complexity matters, Dennett suggests, we are better placed to unmask and expose the various “bugbears” and “bogeymen” that torment incompatibilists and their worries about the consequences of determinism. One issue that incompatibilists have worried about is that if we accept compatibilist conditions of responsibility we might still be manipulated by other agents. In reply, Dennett argues that our “power of reflexive monitoring” enables us to detect “sneaky manipulators” or “evil tricksters” (Dennett 1984: 37-​–​8). Nothing about the thesis of determinism implies that we do not possess and exercise abilities of this kind. Along with the various “bogeymen” of manipulation cases (e.g., “the Nefarious Neurosurgeon”) there are a number of other “bugbears” that Dennett seeks to discredit. It has been argued, for example, that if determinism is true then we would be controlled by nature or controlled by the past—​and could not, therefore, be in control of ourselves. Dennett objects that claims of this kind misleadingly turn the environment and the past into “superhuman intelligences,” when obviously they are nothing of the kind. “Far from it being the case that we are completely under the control of our ancestors or our evolutionary past,” he says, “it is rather the case that that heritage has tended to set us up as self-​controllers—​ lucky us” (Dennett 1984: 72). The opening chapter of Dennett’s Elbow Room presents a more general overview of the way in which incompatibilist pessimism has acquired such a strong 5  In his review of Elbow Room, Gary Watson suggests that “Dennett’s treatment of responsibility seems the least instructive part of the book” and suggests, in particular, that Dennett’s account could benefit if it made more use of Strawson’s discussion of “personal relations” in “Freedom and Resentment” from (Watson 1986: 522).

526   Paul Russell psychological grip over us. Worries and anxieties about determinism, he suggests, are the product of “fearmongery” by philosophers, who have “conjured up a host of truly frightening bugbears” (Dennett 1984: 4). These pessimistic “gloomleaders” have relied, Dennett claims, on various thought experiments that serve as “intuition pumps” designed to arouse some relevant negative response in us (Dennett 1984: 12, 18). The fact is, however, that all these “intuition pumps” rely on misleading analogies and metaphors that do not deserve the respect and influence they have enjoyed (Dennett 1984: 7). Dennett’s fundamental task in Elbow Room is to “exorcise” these bugbears and eradicate this groundless, “self-​induced panic” (Dennett 1984: 169–​170). Although Dennett is no doubt successful in discrediting some of the cruder and more misleading features of the various “bugbears” that he considers, critics may argue that he does not succeed in dispelling all the worries and anxieties that may grip us. Among the most important of the “bugbears” that Dennett examines are the issues of fate and luck. As we have noted, much of what Dennett says on these subjects repeats the familiar repertoire of classical compatibilism. Dennett takes the doctrine of fatalism to be the claim that all our deliberations and actions are causally ineffective and make no difference to the course of events (Dennett 1984: 104–​105). He goes on to point out that nothing about the thesis of determinism implies that this is the universal condition. The problem with this reply, as his critics see it, is that it is evasive and fails to engage with the deeper difficulties we encounter here. The relevant issue is not about the causal influence of the agent but rather the causal influences on the agent. The incompatibilist argues that there is another mode of fatalistic concern that arises from a backward-​looking perspective. The question that concerns us, from this perspective, is whether or not the agent has ultimate control over her character and conduct. Do the qualities and features of her will for which we may praise or blame her originate with her? The essence of the incompatibilist view is that if determinism is true then origination fate is the universal condition and responsibility is impossible.6 Dennett’s effort to deal with the bugbear of luck runs into a number of related difficulties. From an incompatibilist perspective, concern about origination fate is intimately connected with worries about “moral luck.” It is intuitively unjust, they argue, to hold agents responsible for aspects of their conduct and character they do not control. Agents subject to origination fate are vulnerable to “moral luck” of just this kind. Dennett aims to deflate the luck objection by arguing that human agents are “not just lucky” but are, on the contrary, “skilled self-​controllers” (Dennett 1984: 94). Here again, Dennett’s critics are not convinced. While it is true that human agents are sensitive to reasons and able to guide their conduct in light of their reasons, given determinism it is also true that the specific capacities we have, the circumstances in which they are engaged, and the way they are actually exercised depend, in the final analysis, on external factors that the agent has no control over. In sum, contrary to what Dennett claims,

6 

For another perspective on this matter see Russell 2000/​2013; and Russell 2002.

Moral Responsibility and Existential Attitudes    527 powers of reason-​responsiveness fail to provide immunity from the various forms of (moral) luck that incompatibilists are worried about.7 Dennett, as we have noted, believes that his account of our (compatibilist-​friendly) powers of rational self-​control are all we need in order to defuse worries about (evil) manipulators of various kinds (Dennett 1984: 8, 33–​34, 37–​38, 64–​65). At the same time, Dennett also advances a theory of responsibility that emphasizes the way in which society can deploy the stance of “designers” and “engineers,” with a view to controlling and directing an agent’s character and motivations (e.g., through the effective implementation of rewards and punishments) (Dennett 1984: 156–​165; see also 139–​141, on robot design). These core arguments in Dennett’s work, it may be argued, pull in opposite directions and reveal deep tensions in his overall project in Elbow Room. Although Dennett claims to distance himself from the “chilly” world of B. F. Skinner’s Walden Two and the “vision of behaviourism” that lies behind it, his general position on this subject might well be described as neo-​Skinnerian.8

4.  Hard Incompatibilism—​Optimistic Skepticism Both Strawson and Dennett, we have noted, aim to refute skepticism and offer compatibilist accounts of moral responsibility that are plainly more sophisticated than the simpler models proposed by classical compatibilism (Optimism). Whereas Strawson’s model rests on the role of reactive attitudes and excuses, Dennett emphasizes the importance of our powers of rational self-​control.9 Whatever their differences, they both agreed that, if we can discredit skepticism, then we can vindicate some reasonable degree of optimism about our general predicament. Failing this, skepticism implies 7  Nagel 1976/​2013 provides the classical statement of these worries. For other similar lines of criticism see, for example, Miles 2015: 58–​74; and also Levy 2011; Waller 2011; Harris 2012. 8 In Walden Two (1948) Skinner employs the apparatus of his “behavioural science” to describe a (dis) utopian world in which the benefits of the advances in the science of human behavior could be taken advantage of with a view to creating a superior ethical agent and citizen. Dennett is careful to reject Skinner’s language of “man being controlled by his environment” (Skinner 1971: 19–​20; Dennett 1984: 57–​59), and is certainly aware of the potential misuse of behavioral control (Dennett 2003: chs. 9 and 10). It may be argued, nevertheless, that like Skinner he welcomes the prospect of the state or authorities “engineering” society and controlling others by means of a “technology of behavior.” While we may be skeptical about the prospects of “behaviourism,” as Skinner conceives it, worries of this more general kind are not just science fiction but constitute a real threat. (Consider, for example, some of the policies and practices, employing new techniques and scientific methods relating to mind control with a view to creating “better citizens” that are now being implemented in China, among other places.) Belief in “autonomous man” may well be illusory, but when it is discarded we must confront some uncomfortable problems. 9  The respective strengths of each theory could be combined into a unified account. See, e.g., Wallace 1994.

528   Paul Russell a significant measure of pessimism. In recent years, orthodox views of this kind have been challenged by “Hard Incompatibilists” or “optimistic skeptics.”10 They argue that skepticism has none of the bleak or disturbing implications or consequences that are generally attributed to it by its critics (such as Strawson and Dennett). The most influential and substantive statement of this view is presented in the work of Derk Pereboom.11 Pereboom’s approach, as the (alternative) label “optimistic skepticism” suggests, also falls into two parts (Pereboom 2001: xxiii). The first involves the set of arguments advanced in support of skepticism about moral responsibility. The second concerns his effort to justify an optimistic attitude on the basis of the skeptical claims advanced. Our concerns in this context rest primarily with Pereboom’s effort to justify optimism. There are, however, two features of his skepticism that should be noted. First, Pereboom’s skepticism is not only incompatibilist with respect to determinism, it is also incompatibilist with respect to indeterminism. Pereboom is “agnostic” about the truth of causal determinism but maintains that we lack moral responsibility either way.12 It is for this reason he refers to this general position as “hard incompatibilism.” The second point about Pereboom’s skepticism that should be noted is that it is qualified or limited in its scope. Pereboom makes clear that there are several different senses of “moral responsibility” and that he is specifically concerned with the one “that has been at issue in the historical debate” (Pereboom 2001: xxx–​xiii; Pereboom 2007: 86–​87; Pereboom 2013: 421; Pereboom 2014: 2–​4, 6–​7; Pereboom 2017: 121–​122).13 This is, he suggests, the concept that involves “the notion of basic desert” (Pereboom 2007: 122–​123; Pereboom 2014: 2). The basic desert notion suggests that for an agent to deserve blame (or praise), she must “be able to exercise a certain kind of control in its production” (Pereboom 2001: 47; Pereboom 2007: 86–​87). In particular, the agent must be the source of her decision in the right way, and it is this sort of agency that we lack (whether determinism is true or false). What motivates these concerns with the (ultimate) source of our decisions and conduct is that basic desert serves as the foundation of our retributive dispositions and practices.14 The skeptic can allow, Pereboom maintains, that there are other forms of free 10 

These labels are used by Pereboom, and others in this school have followed him. Other important figures who also belong to this school include Waller 2011; Harris 2012; Miles 2015; Nadelhoffer 2010, and Caruso 2016, Caruso, forthcoming. For useful background see Caruso 2018. 12  Unlike some skeptics (e.g., G. Strawson 1994), Pereboom does not claim that moral responsibility is “impossible.” What he claims is that although being “undetermined agent causes” may be a coherent possibility, it is not credible given our best theories.” (Pereboom 2013: 422). 13  Given that the form of skepticism that Hard Incompatibilist defends is “targeted” in this way, we may describe it as a partial skepticism, to distinguish it from a more radical, total skepticism. However, in light of this distinction, it may be argued that there is a latent instability in the Hard Incompatibilist position about whether the skepticism that they defend is partial or total. To the extent that Hard Incompatibilists insist that the traditional, historical debate is (entirely) about basic desert responsibility, and that compatibilists are, therefore, really “skeptics” (e.g., Pereboom 2001: xxi–​xxii; Pereboom 2014: 2–​3; Caruso 2017: 200) they tend to blur this important distinction. For compatibilists it is crucial that this distinction should not be blurred in this manner, since partial skeptics are not committed to total skepticism. 14  Waller places particular emphasis on this point (Waller 2011: 2–​7). According to Waller, “the core concept of moral responsibility is what justifies blame and praise, punishment and reward; moral 11 

Moral Responsibility and Existential Attitudes    529 will and moral responsibility that agents may actually possess and that are obviously compatible with determinism (Pereboom 2001: xxi–​xxii; Pereboom 2007: 86; Pereboom 2013: 421–​22; Pereboom 2014: 2–​3, 131, 134). Nevertheless, efforts to defend compatibilism in these terms are, Pereboom maintains, irrelevant to the traditional debate about free will and moral responsibility (Pereboom 2001: xxi–​xxiii; Pereboom 2017: 121–​122).15 On the basis of arguments advanced to show that both compatibilist and libertarian views must be rejected, Pereboom goes on to show that the Hard Incompatibilist position that follows has few, if any, of the unattractive consequences that are generally attributed to it. In this respect, his method bears some resemblance to Dennett’s, in that he aims to eradicate various “bugbears” that critics of skepticism have conjured up. If this can be done, we can vindicate optimism on the basis of skepticism (as opposed to its refutation). The title of Pereboom’s first book—​Living Without Free Will—​aptly describes this program. According to Pereboom, the program that he proposes is not only feasible, it is an attractive ideal that we have good reason to implement and put into practice. There are three major sources of pessimism that are frequently said to follow from the skepticism of the kind that Hard Incompatibilists defend. The first is that skepticism about moral responsibility would have catastrophic consequences for morality and society. Pereboom denies this and argues that not only can we legitimately continue to judge actions as morally right or wrong, we still have sufficient resources to respond to wrongdoers in an effective and appropriate manner. Not only can we use “moral admonition and encouragement”—​stripped of any association with basic desert—​we can also respond to “crime” in ways that treat the wrongdoer as needing treatment and cure rather than punishment (Pereboom 2013: 435–​437).16 When an agent (criminal) is judged dangerous then it may be necessary to “quarantine” her—​but there is, Pereboom maintains, nothing morally objectionable about that.17 responsibility is the basic condition of giving and claiming both positive and negative deserts.” (See also Miles 2015: 42–​44; Harris 2012: 55–​60; and, especially, Caruso 2016). Caruso’s paper is devoted to providing a more detailed picture of “what the basic desert sense of moral responsibility amounts to.” He argues that it is tied very tightly with retributivist aims and values and suggests that it should be called “retributivist desert moral responsibility.” 15 

Sam Harris devotes a whole chapter of his (brief) book on free will to this general concern, arguing that “compatibilists” are guilty of “changing the subject” (Harris 2011: 15–​26). 16  It has been argued by some that it is not possible to separate wrongness from desert and retribution. For example, Mackie claims that “if we analyse the concept of a morally wrong action it will be seen to have much wider significance. In the central cases, the wrongness of an action is thought of as being made up of three elements, its being harmful, its being forbidden, and its calling for a hostile response. No one of these elements, if the others are somehow denied is sufficient to make an action wrong in the full sense” (Mackie, 1985: 213). 17  A related worry here is about what methods may be used to “cure” or “help” the offender and protect society? What limits are there on such methods once we abandon the idea that we need to distinguish between (truly) responsible agents and those who are simply “ill” or “sick”? The scope for (arbitrary) interference and mind control of various kinds would appear to be much more open-​ended once these skeptical principles and assumptions are adopted or put into practice. This may well also be a problem for compatibilist theories, such as Dennett’s, but that does not show that the skeptic can deal with these concerns or has found a way around them.

530   Paul Russell Another potential source of pessimism, critics suggest, is that in a world stripped of basic desert there would be no “genuine achievement,” since no agent would deserve credit for their accomplishments. Pereboom denies that achievement is as closely tied to praiseworthiness as the objection supposes. Nor does Pereboom accept that our sense of self-​worth rests entirely on features of ourselves that depend on deserved credit. In ordinary life we take pride in various features of ourselves that do not depend on our efforts in any way (e.g., intelligence, beauty, etc.). Moreover, qualities and features that do depend on our own will may still be valued even if it turns out that they are not freely willed. Nor is there any reason, according to Pereboom, to experience any sense of dismay even if our moral character is not something that we have produced or deserve praise or blame for. Given that these worries can be handled, we can be confident that our lives will continue to have worth and significance in the absence of deserved praise or blame. A final set of pessimistic worries (bugbears) that Pereboom seeks to discredit concerns the suggestion that if skepticism is correct it will result in the complete erosion and distortion of our personal relationships and emotional lives. This is a line of pessimistic concern that, as we noted, Strawson advanced in “Freedom and Resentment.” Contrary to what Strawson suggests, not all forms of reactive attitude will be discarded if we accept skepticism. There are some “analogues” that will be “sufficient to sustain relationships” (Pereboom 2007:119). Moreover, many of those that do not survive, on skeptical principles, we would be better off without—​such as (desert-​based) anger and resentment (Pereboom 2013: 443–​444). Included in the range of emotions that are not endangered by skeptical principles or that may survive in an “analogue” form, Pereboom argues, are forgiveness, gratitude, mature love, regret, and moral sadness. Taken together, all these emotions are sufficient to avoid anything resembling the bleak, dehumanized existence that troubles Strawson when he describes a world dominated exclusively by the objective attitude. At this juncture there may be some questions we may ask as to whether there is any real distance between Hard Incompatibilism and Dennett’s compatibilism, beyond the way in which they frame or label their respective positions. Clearly both Pereboom and Dennett are skeptical about moral responsibility in the “basic desert sense” (hereafter BDR) (Pereboom 2014: 2–​3; cp. Pereboom 2001: xxi–​xxii; Dennett 1984: 166). Pereboom also accepts that there are non-​BDR understandings of moral responsibility that are “compatible” with determinism—​including the specific forms of moral responsibility that Dennett aims to describe and defend. They also both agree that jettisoning BDR and retaining surviving modes of responsibility allows for a generally optimistic attitude. On all these key issues, therefore, they converge. There is, perhaps, some real disagreement about the extent to which these shared conclusions “leave everything where it is”—​especially as concerns the issue of punishment (although, even here, disagreement may be more apparent than real). Beyond this, however, the only remaining point of disagreement we are left with is that Pereboom, as we noted, objects to Dennett-​style compatibilism on the ground that it presents other kinds moral responsibility as a substitute for BDR, which he takes to be an attempt to “recast the debate” (Pereboom 2001: xxii; and see also Harris 2012: 20–​26). Pereboom objects, in particular, that presenting

Moral Responsibility and Existential Attitudes    531 some non-​BDR interpretation under the heading of “compatibilism” turns almost everyone into a “compatibilist” (Pereboom 2014: 2; similar criticisms of compatibilists are advanced in Nadelhofer 2010). It may be pointed out, in reply, that this line of criticism turns every compatibilist who rejects BDR into a “skeptic.”18 For many compatibilists, however, including Dennett, the crucial point is that partial (qualified) skepticism about “absolutist” or BDR conceptions of moral responsibility does not imply total (unqualified) skepticism. Insisting that compatibilists should be labeled “skeptics” (tout court) obscures this point.19 Setting this issue aside, the question still remains with us about how convincing it is to claim that skepticism about basic desert responsibility has none of the bleak and troubling implications or consequences that have been attributed to it? The basis for much of the optimism claimed by (partial) skeptics is that by pushing us to abandon the illusory self-​image required by BDR, we are encouraged to discard and repudiate cruel and destructive retributive emotions and practices. This is, no doubt, an attractive and humane aim of the Hard Incompatibilist position. It is not clear, however, that skepticism about BDR is needed to achieve this end. The underlying concern here is that the skeptic operates with a questionable picture of the relationship that holds between BDR and our retributive principles and practices. On the face of it, Hard Incompatibilists present BDR as committed to some form of positive retributivism, which suggests that blameworthy individuals who have committed a wrong ought to be punished or that it is good that they are made to suffer.20 There are, however, weaker and more restrained forms of retributivism, such as negative retribution, which requires only that we limit punishment to the guilty, and permissive retributivism, which allows that the guilty may be punished but does not demand this. If we take BDR to be committed to a positive retributive principle of some kind then it is, indeed, a severe and unattractive doctrine. It is, however, highly questionable that a notion of moral responsibility of this (narrow and extreme) kind represents our “ordinary” views—​which are, arguably, much more conflicted, fragmented, and unstable than this analysis suggests.21

18 

From the perspective of Pereboom, Harris, and other like-​minded skeptics, those who identify themselves as “compatibilists” fall into two classes. There are those who accept BDR and (mistakenly) believe that it can be secured within compatibilist constraints. There are also those who reject BDR and are, therefore, really skeptics—​since presenting themselves as compatibilists involves “changing the subject.” Suffice it to note, beyond this, that which camp compatibilists are assigned to will depend on how exactly we interpret the requirements of BDR (about which there is little consensus). 19  Accepting this, of course, is entirely consistent with rejecting any particular compatibilist account (including Dennett’s), which may well be judged inadequate or incomplete. Plainly, compatibilists disagree among themselves as to what kinds of alternative accounts are adequate to the (complex) facts and standards that we need to appeal to. 20  These distinctions concerning retributive principles are drawn and explained in Mackie 1985: 207. 21  For a rather different perspective on retributivism and the various distinctions that are required to make sense of it, see Hart 1959/​1968: esp. ch. 1. The deeper problem here is that Pereboom’s account of “basic desert” is described largely in negative terms, as being not derived from consequentialist or contractualist considerations (on this see McKenna 2012: 121). For this reason it remains unclear exactly how or in what way BDR is related to retributivist principles. McKenna provides an illuminating

532   Paul Russell It may be argued that Hard Incompatibilism not only exaggerates the benefits of (partial) skepticism, it also tends to downplay or minimize its costs. In support of their view, Hard Incompatibilists claim that there are available “analogues” and alternative to desert-​based reactive attitudes and that these other forms of moral emotion will suffice to support and sustain a (post-​BDR) morality. Critics may question how effective alternative or analogue versions of our reactive attitudes would be in supporting and sustaining (post-​BDR) moral life? Consider, for example, the proposal that (moral) “sadness,” “disappointment,” and “regret” can fill the void left by our desert-​based reactive attitudes. Responses of this kind seem too tepid and tame to do the work of their originals—​especially in more serious or grave contexts and situations.22 Even if we agree with the Hard Incompatibilist that criminal agents do not deserve blame or punishment—​and many will, obviously, contest this—​any surviving “analogues” that remain available will be an unconvincing substitute. When our moral resources are reduced and impoverished in this way, our commitment to the (moral) values will be correspondingly weakened and eroded. Justified or not, this is not a comforting or stabilizing message for morality.23

5.  Mixed Theories—​Affirmation and Illusionism Hard Incompatibilism is firmly committed to optimism and maintains that the skeptical conclusion that it advances should be enthusiastically welcomed. There are, however, some closely related positions that argue that the skeptical claims being advanced demand more concessions to pessimism. Influential views of this kind have been presented and defended by Ted Honderich and Saul Smilansky. They are both skeptics about BDR but they do not accept that skepticism is as benign as the Hard Incompatibilist would have us believe. The views that Honderich and Smilansky have advanced may be characterized as “mixed” theories, since they absorb both optimistic and pessimistic elements into their stance. Where they differ from each other is over the matter of the degree of pessimism that follows from skepticism about BDR and what practical measures need be taken in response to the skeptical assessment of our human predicament.

examination of a number of the issues arising in relation to this general problem (McKenna 2012: 4–​5, 114–​121, 150–​170). 22 

For criticism along these lines see Nichols 2007; but see Pereboom’s replies in Pereboom 2014: 146–​ 149, 180–​181. 23  It may be further argued that skepticism, if put into practice, is liable to result in collateral damage, for both our moral motivation and our appreciation of our various personal and moral relationships. The expectation that our moral and personal lives will carry on, largely undamaged, with few significant costs to bear for such a radical switch in our attitudes and self-​understanding, seems too optimistic. (On this see Shabo 2012; and Pereboom’s response in Pereboom 2014: 183–​186.)

Moral Responsibility and Existential Attitudes    533 Honderich’s “Affirmation” theory is embedded in his general “theory of determinism.”24 Determinism, as Honderich interprets it, rests on the claim that our choosing and deciding is subject to ordinary (necessitating) causal regularities. This is a theory, he argues, that is “strongly supported” and has certainly “not been shown to be false” (Honderich 2002: 90). The question he asks is what are the consequences of determinism, if it is indeed true? His answer to this question begins by pointing out that if determinism is true then the opposing view of “the philosophy of Free Will” is mistaken, since it presupposes indeterminism (Honderich 2002: 2, 4, 35, 41–​42, 49, 69, 76). A key aim behind the Free Will view is that we need to understand responsibility in a certain way. More specifically, it requires that responsible agents have “a kind of personal power to originate choices and decisions and their actions” (Honderich 2002: 2—​emphasis in original). Absolute responsibility of this kind not only requires the falsity of determinism, it also requires some sort of “ongoing entity” that possesses the “active power” required to produce or not produce a given action (Honderich 2002: 42,49). If we accept the truth of determinism, then we must reject the idea of Free Will, so understood. According to Honderich, accepting the truth of determinism and abandoning the idea of Free Will certainly damages a range of things that we care deeply about. This includes, among other things, our personal feelings and attitudes to other people (e.g., resentment and gratitude), to the extent that these depend on a notion of origination and Free Will. These personal feelings and reactions are closely bound up with our moral feelings and responses (Honderich 2002: 97–​100). If there are no active agents of the kind presupposed in our idea of Free Will, Honderich suggests, this will discredit our entire understanding of retributive justice and deserved punishment (Honderich 2002: 101–​103, 133-​141). It is in relation to this issue that our sense of dismay in confronting the truth of determinism will be felt most acutely. Honderich recognizes that compatibilists will resist these claims. The compatibilist will argue that the entire analysis, leading to dismay, rests on a mistaken understanding of what we mean by freedom. Freedom is not, they maintain, a matter of origination of some metaphysical kind but rather a matter of voluntariness (Honderich 2002: 98–​99). As long as we keep in mind that free actions have a certain kind of causal history, we can draw all the relevant distinctions that we need to make. Even if we accept the truth of determinism, we can still distinguish, for example, between voluntary and involuntary conduct; and distinguish conduct due to violence, ignorance, or merely accidental, and so on. The compatibilist concludes from this that determinism changes nothing and leaves everything as it was (Honderich 2002: 97, 99). The compatibilist, therefore, rejects dismay and instead adopts an attitude of “satisfied intransigence” (Honderich 2002: 97). Both these opposing responses to the likely truth of determinism, Honderich maintains, are unsatisfactory. The root mistake, common to both, is that they assume 24  Honderich’s theory of determinism was first presented in Honderich 1988 and then, in a revised and compressed form, in Honderich 2002. Although the later work “follows the same path” as the earlier work, it is intended to be more than a mere précis of it. Readers should consult Honderich’s first and larger work for more detailed arguments and discussions relating to his views.

534   Paul Russell that we have just one, single, ordinary idea of freedom and that any alternative account is either inadequate or irrelevant (or both). Contrary to this view, Honderich argues that we have two distinct ideas of freedom or free choice, one understood in terms of origination and the other in terms of voluntariness (Honderich 2002: 109, 120). Any effort to arrive at a proper understanding in terms of just one of these two concepts will be one-​sided and incomplete. The fact is that when we consider the consequences of determinism, we experience a dual response of both dismay and intransigence, which results in “attitudinarian instability and discomfort” (Honderich 2002: 125–​126). The solution to this problematic situation is not just “intellectual” or conceptual but a matter of reconciling our conflicted feelings and attitudes (Honderich 2002: 114). This requires some concessions on both sides. Contrary to compatibilist accounts, we must accept that some of the features and aspects of human agency that we care about cannot be retained or supported, as they are not consistent with the truth of determinism. It is, for this reason, a mistake to suppose that accepting determinism “changes nothing” or “leaves things just as they are.” However, while we must not deny that there are things we value that we must give up, we should not underestimate what we still have. Determinism does not justify any sense of defeat or resignation and to this extent our sense of dismay can be contained, if not eliminated. The upshot of all this is an attitude of Affirmation, which is a “philosophy of life” that encourages us to “accommodate ourselves to the situation we find ourselves in—​accommodate ourselves to just what we can really possess if determinism is true, accommodate ourselves to the part of our lives that does not rest on the illusion of Free Will” (Honderich 2002: 126). One important respect in which Affirmation is in agreement with Hard Incompatibilism is that it is skeptical about BDR—​which rests on the idea of origination and cannot be rendered consistent with the truth of determinism. Where Affirmation diverges from Hard Incompatibilism is that it remains firmly committed to a stance of dismay (or pessimism)—​an attitude that Hard Incompatibilists aim to discredit or minimize. The crucial point here is that Affirmation refuses to conceal the source of dismay that determinism generates, as grounded in skepticism about the idea of Free Will or origination. From this perspective, the undiluted optimism of Hard Incompatibilism involves a form of denial about the real losses that we must face up to. At the same time, Affirmation also insists that we can accept our condition, and live with it, without relying on any illusions, as there are plenty of considerations of a more reassuring kind for us to call on. Nothing about the consequences of determinism implies that our lives will “go cold” or suggests that “everything is wrecked” (Honderich 2002: 125). Accepting the (likely) truth of determinism does not leave everything where it was, and certainly involves the loss of aspects or dimensions of our lives that we have reason to care about. There are, nevertheless, no tragic or unbearable consequences that flow from this predicament. Another “mixed” theory that resembles Affirmation in several respects is presented by Saul Smilansky in the form of “Illusionism” (Smilansky 2001/​2013). Smilansky is, like Honderich, a “dualist” about free will in the sense that he holds that, while it is true that we enjoy compatibilist-​friendly forms of freedom (i.e., voluntariness and forms of

Moral Responsibility and Existential Attitudes    535 control associated with it), we also lack control in the sense of origination or ultimacy. Smilansky does not deny that compatibilism has “partial validity,” in that the forms of freedom it describes are indeed needed to support a “Community of Responsibility.” What he denies is that this sort of compatibilist freedom is sufficient for human ethical life (Smilansky 2000/​2013: 384–​385, 393). More specifically, agential resources of this compatibilist kind, he argues, cannot support and sustain a wide range of fundamental and essential ethical values and interests. Among the values and interests that Smilansky has particularly in mind are justice and respect (Smilansky 2000/​2013: 386–​388). The general difficulty we face is that while we need to believe in (libertarian) conceptions of freedom and responsibility, in order to maintain our confidence in and commitment to ethical life, this belief collapses under critical scrutiny. This theoretical problem, Smilansky maintains, presents us with a real and troubling practical problem. Although Smilansky accepts much of Honderich’s analysis, he takes a much more pessimistic view about the consequences of determinism and the accompanying loss of our belief in Free Will (qua ultimacy and origination). As Smilansky sees it, Affirmation is not a viable or credible solution to the predicament that we actually face. For this reason, he proposes a much more radical solution—​“illusion.” If we are to maintain a secure foundation for our core values and ethical practices this will require that the illusory belief in free will should be promoted and encouraged. Given that a belief of this general kind is “a condition of civilized morality and personal value” (Smilansky 2000/​2013: 88) this is not something that we can afford to abandon or give up. Clearly, then, Smilansky’s (deep) pessimism about the implications of (openly) embracing skepticism about origination and ultimate agency leads him to advocate this more radical proposal. Smilansky is, of course, well aware of both the considerable costs of this (e.g., with regard to truthfulness) and the challenging practical obstacles (e.g., maintaining patterns of deception by elites over the masses). He argues, nevertheless, that all this is necessary to avoid a catastrophic erosion of human ethical life and practice. When we compare Hard Incompatibilism, Affirmation, and Illusion with each other, they fall into a spectrum of attitudinal alternatives, and all of them are founded on skepticism about BDR. Hard Incompatibilism, as presented by Pereboom, Caruso, Waller, and others, embraces and welcomes skepticism as providing an opportunity for a radical transformation of our attitudes and practices, grounded on the skeptical arguments and reflections that have been advanced. Smilansky views these developments with a sense of dread and alarm. and he encourages us to conceal this truth about our predicament, in the hope of averting the (ethical) damage that it is liable to bring about. Honderich takes a middle path between these two poles, arguing that although there certainly are bleak and troubling losses that we must accept if we discard our belief in Free Will, we have, nevertheless, plenty of other resources to call upon to support ethical and meaningful lives. One important consideration that separates Affirmation and Illusionism, on one side, from Hard Incompatibilism on the other, is that for the former pair of views the primary source of their persisting pessimism is rooted in skepticism about BDR. In contrast with this, Hard Incompatibilism presents our situation in a very different light, holding that our mistaken and illusory belief is mostly a source

536   Paul Russell of human misery and irrationality, and one that we are better off without.25 Despite these differences, all three parties agree that some form of compatibilist freedom and associated forms of moral responsibility survive skepticism about BDR. Considered in their own terms, these compatibilist forms of responsibility are understood to be in no way troubling or problematic—​much less a source of pessimistic concern. According to these views, whatever sources of pessimism there may be, we do not encounter them here.

6.  Free Will Pessimism We have noted that both libertarians and compatibilists generally accept the orthodox view that pairs skepticism and pessimism and, conversely, suggests that the defeat of skepticism serves to vindicate (some measure of) optimism. We also considered the Hard Incompatibilist’s unorthodox view that skepticism may be construed in optimistic terms. Mixed views, such as Affirmation and Illusionism, maintain the orthodox linkage between skepticism and pessimism, locating sources of pessimism in skepticism about BDR, while retaining differing measures of optimism based on compatibilist considerations. What has not been considered, so far, is the suggestion that arguments and considerations advanced that defeat skepticism about moral responsibility may, nevertheless, serve as a basis for pessimism. Given the standard matrix of options, this may be received as a puzzling proposal. It is, however, a view of this kind that is advanced and defended by the Free Will Pessimist.26 The approach taken by Free Will Pessimists relies on distinguishing carefully between issues of skepticism and pessimism. (We have already noted, in relation to Hard Incompatibilism, that it is a mistake to conflate these issues.) Free will skepticism is the view that the lack of origination agency, along with vulnerability to moral luck, serves to discredit our view of ourselves as free and responsible agents (tout court). Free will pessimism rejects free will skepticism, since the basis of its pessimism rests with the assumption that we are free and responsible agents. All the major parties and positions in the (traditional) free will debate resist the suggestion that we are free and responsible agents who are, nevertheless, subject to significant modes of fate and luck. According to free will pessimism, all the major positions are, in different ways, modes of philosophical evasion that encourage an untruthful picture of the human predicament with respect to agency and ethical life.

25  Thomas Nadelhoffer has described the Hard Incompatibilists as “the revolutionaries” in this debate and suggests that this outlook “might helpfully be called free will disillusionism” (as an obvious point of contrast with Smilansky’s “illusionism”). Nadelhoffer is unconvinced by claims relating to “the alleged benefits of believing in libertarian free will and moral desert” and declares his allegiance to the revolutionary cause (Nadelhoffer 2010). 26  Russell 2017a; see also Russell 2000; Russell 2002; Russell 2008; and Cuypers 2013.

Moral Responsibility and Existential Attitudes    537 At the heart of the free will debate, structuring the standard options that we are presented with, there is a fundamental shared assumption. What all parties agree about is that any acceptable account of conditions of responsible agency must be such that they do not allow that agents who satisfy these conditions may also be subject to conditions of fate and luck. Strong resistance to this proposal is something that unites all the parties, whatever other disagreements there may be. Disagreement, across the compatibilist/​incompatibilist boundary, is rooted in different understandings or interpretations of what the relevant (contested) notions do or do not involve. Compatibilists and libertarians propose very different accounts of the powers and abilities that are required to exclude the presence of fate and luck. Both sides contest the adequacy of their opponent’s proposals. The skeptic maintains that no satisfactory or credible “solution” to this problem can be found—​but insists, nevertheless, that the relevant standard in place must still be respected and cannot be simply be discarded or dropped. The main focus of critical attention for the free will pessimist is the (shared) assumption that any credible account of freedom and responsibility must exclude conditions of fate and luck. It is this fundamental assumption that the free will pessimist challenges, maintaining that there are available conditions of free, responsible agency that fail this test but are, nevertheless, adequate for the purposes of sustaining and supporting ethical life and practice.27 Although accounts of this kind retain an appropriate degree of connection between conditions of control (e.g., as involving the capacity and exercise of rational self-​control, under some interpretation), they reject the ambition to exclude or eliminate all surviving modes of fate and luck. Among the forms of fate and luck that would still be admissible, on this view, are limits relating to agent’s control over the circumstances in which she decides and acts, the background features that may shape and condition her motivation and character, and the particular way that the agent may exercise her deliberative and agential capacities. In conditions of this kind, the free will pessimist maintains, agents may be both (fully) free and responsible and still subject to significant limits of control over the moral trajectory of their lives. All the established parties involved in the (traditional) free will debate simply rule out this possibility on the ground that it is a necessary condition of the adequacy of any proposed account of freedom and responsibility that conditions of this kind do not coexist. The core objection to any general proposal of this kind is rooted in concerns about “fairness” and justice. From this perspective, it is neither fair nor fitting to hold agents who are limited in these ways responsible for their conduct. This criticism has force, the free will pessimist may reply, only if we aspire to some form of absolute fairness, that “runs all the way down,” and distinguishes morality from other forms of agential evaluation that do not rely on this (metaphysically ambitious) standard (Russell 2008). This absolutist standard places impossible metaphysical weight on the requirements of responsible agency, which is why it collapses into (complete, unqualified) skepticism. The

27 

As I explain further later, this does not mean that “nothing changes” or that “the common wisdom about our place in the universe is roughly right” (Dennett 1985: 19).

538   Paul Russell free will pessimist rejects these aspirations along with the standards that they promote and encourage. The issue of fairness, on this alternative view, is internal to the fabric of ethical life and the norms and values that they track (e.g., as broadly described in Strawson 1962). These norms and values may well be modified, corrected, and amended over time, on the basis of critical reflection. However, any demand for fairness beyond or above this—​some mode of absolute fairness—​is a product of an illusory self-​image and an untruthful picture of our human ethical predicament.28 With respect to the related objection that free will pessimism, as presented, fails to defeat or discredit skepticism about BDR or provide a “solution” to the free will problem as understood in these terms, the free will pessimist can readily accept this. It is not the aim of the free will pessimist to offer a solution to the “traditional” problem, much less to respect this particular standard of success. On the contrary, it is a matter of fundamental importance for the free will pessimist that these aims and aspirations, which structure the traditional problem, should be rejected and abandoned. They are faulty because they rule out the very possibility that we may encounter (fully) responsible agents who are, nevertheless, subject to conditions of fate and luck. The only skepticism that the free will pessimist endorses is partial skepticism (i.e., as targeted on BDR and related concepts that aim to satisfy its demands). A skepticism of this limited nature can, however, still allow for the legitimacy of (robust) forms of moral responsibility that do not deny the possibility of responsibility in the face of conditions of persisting fate and luck.29 One way we can account for the pessimism involved in this stance is to consider free will pessimism in relation to what Bernard Williams calls “the morality system” (Williams 1985: ch. 10). The morality system, as Williams describes it, places particular emphasis on the (peculiar) concept of obligation, along with the closely related concepts of blame and voluntariness. Moral responsibility, as the morality system understands it, is taken to be primarily a matter of rational agents voluntarily violating their obligations and, thereby, being liable to blame and retribution. Williams refers to these sub-​features of morality as “the blame system” (Williams 1985: 217). Understood in these terms, the concept of BDR and of “accountability” may be viewed as specific modes or forms of “the blame system.” Building on these core concepts and features, the morality system demands that moral responsibility, understood in these terms, must somehow be capable of “transcending luck,” providing a “purity” that only genuine (rational) agency of some kind makes possible (Williams 1985: 217). Within this framework, the aspiration of libertarianism and its commitment to some form of absolute, ultimate origination is entirely intelligible. Although compatibilists plainly reject these metaphysical ambitions, they generally remain committed to many of the features of “the blame system” and

28 

For a valuable discussion and examination of the various modes of moral luck that may concern us in relation to these matters, see Hartman 2017. 29  As we have already noted, there may well be disagreement and debate about the adequacy of various proposed alternative accounts. There is, however, no basis for dismissing all such alternative accounts as inadequate, unless we impose the (contested) standard of BDR itself.

Moral Responsibility and Existential Attitudes    539 argue that it can be satisfied within compatibilist constraints.30 In contrast with views of this kind, free will pessimism rejects “the blame system” and the forms of moral responsibility that are essential to it. As the free will pessimist sees it, the aspirations of “the blame system” not only motivate the various philosophical conundrums that plague the (traditional) free will problem, but they also distort and misrepresent our understanding of the human ethical predicament (which plainly lacks the sort of “purity” that the blame system aspires to and demands). We are, as Williams suggests, “better off ” without “morality” and, in particular, better off without “the blame system.”31 Once these distinctions are drawn it is apparent that in abandoning “the blame system” we abandon significant optimistic features that are integral to it. In particular, “morality” offers us a comforting story to tell about the human predicament in respect to moral agency—​a story that runs deep in Western society. Morality puts up strong resistance to any disturbing or troubling view of human ethical life that suggests that the manner in which agents exercise and operate their moral capacities and abilities depends in large measure on features that are not controlled or governed by these same capacities and powers. As Thomas Nagel puts it, accepting an outlook of this kind would leave us “morally at the mercy of fate.” Nagel takes this to be a reductio of any outlook of this kind, whereas the free will pessimist maintains that it is the only truthful description of our predicament available to us. Trying to avoid this truth about our circumstances and conditions is exactly what generates the (intractable) conundrums of the free will problem and the various forms of philosophical evasion associated with it. We may conclude by noting, with Bernard Williams, that “philosophy, and in particular moral philosophy, is still deeply attached to giving good news” (Williams 1996/​ 2006: 49). This tendency is particularly apparent and heavily pronounced in the debate concerning free will and moral responsibility. Not only do almost all the parties aim to deliver “good news” or an optimistic conclusion of some kind, the very assumptions and aspirations that structure and motivate the whole debate manifest this prejudice. In contrast with this general outlook, what free will pessimism claims is that with respect to these issues, what we have is not an intractable (philosophical) problem that is still looking for a solution but rather a troubling and unsettling predicament that must be recognized and acknowledged.32

30  This aim is explicitly stated in Wallace 1994: 39–​40, 64–​66. Many would interpret Strawson as sharing these aims and aspirations (Strawson 1962). Even Dennett, who takes a self-​consciously pragmatic turn on this subject, continues to cater to central features of “the blame system”—​including the aspiration to “transcend luck.” 31  Williams 1985: 193. For reasons already touched on, it should be clear that abandoning “the blame system” and the (peculiar) forms of moral responsibility it seeks to impose on us, does not commit us to total or complete skepticism about moral responsibility. On the contrary, while adherents to the morality system tend to present the situation this way, it is generally recognized—​even by “morality’s” adherents—​ that the narrow conception of moral responsibility, as constructed around this system, is one that is both “local” (i.e., modern and Western) and widely contested—​even within our own modern, Western ethical community. (See, e.g., Wallace 1994: 39–​40, 64–​65.) 32  I am grateful to Derk Pereboom for his comments and suggestions on an earlier draft of this paper.

540   Paul Russell

References Caruso, Gregg. (2017). “Compatibilism and Retributivist Desert Moral Responsibility: On What is of Central Philosophical and Practical Importance.” Erkenntnis 82(4). Caruso, Gregg. (2017). “Origination, Moral Responsibility, and Life-​Hopes.” In G. Caruso, ed. Ted Honderich on Consciousness, Determinism and Humanity. London: Palgrave. Caruso, Gregg. (2018). “Skepticism About Moral Responsibility.” In Edward N. Zalta, ed. The Stanford Encyclopedia of Philosophy (Spring 2018 Edition), URL = . Caruso, Gregg. 2019. “Free Will Skepticism and Its Implications: An Argument for Optimism.” In Elizabeth Shaw (ed.), Free Will Skepticism in Law and Society. New York: pp. 43–​72.. Cuypers, Stefaan E. (2013). “Moral Shallowness, Metaphysical Megalomania, and Compatibilist Fatalism.” Ethical Theory and Moral Practice 16(1): 173–​188. Dennett, Daniel. (1984). Elbow Room: The Varieties of Free Will Worth Wanting. Oxford: Clarendon Press. Dennett, Daniel C. (1985). When does the intentional stance work? Behavioral and Brain Sciences 8 (4):763–​766. Dennett, Daniel. (2003). Freedom Evolves. New York: Viking. Fischer, John, Robert Kane, Derk Pereboom, and Manuel Vargas. (2007). Four Views on Free Will. Oxford: Blackwell. Hart, H. L. A. (1959/​1968). Punishment and Responsibility: Essays in the Philosophy of Law. Oxford: Clarendon Press. Harris, Sam. (2012). Free Will. New York: Free Press. Hartman, Robert J. (2017). In Defence of Moral Luck. London: Routledge. Honderich, Ted. (1988). A Theory of Determinism: The Mind, Neuroscience and Life-​Hopes. Oxford: Oxford University Press. Honderich, Ted. (2002). How Free Are You? 2nd ed. Oxford: Oxford University Press. Kant, Immanuel. (1781/​1965). Critique of Pure Reason, trans. Y. N. K. Smith. New York: St. Martin’s Press. Levy, Neil. (2011). Hard Luck: How Luck Undermines Free Will and Moral Responsibility. Oxford: Oxford University Press. Mackie, J. L. (1985). “Morality and the Retributive Emotions.” In Persons and Values: Selected Papers Volume II. Oxford: Clarendon Press. McKenna, M., & Russell, P. (2008). Free will and reactive attitudes: Perspectives on P.F. Strawson’s “freedom and resentment”. In Free Will and Reactive Attitudes: Perspectives on P.F. Strawson's 'Freedom and Resentment' (pp. 1–​328). Ashgate Publishing Ltd. McKenna, Michael (2012). Conversation and Responsibility. New York: Oxford University Press. Miles, James B. (2015). The Free Will Delusion: How We Settled for the Illusion of Morality. Kibworth: Matador. Nagel, Thomas. (1976/​2013). “Moral Luck.” Reprinted in Russell and Deery (2013), 31–​42. Nadelhoffer, Thomas. (2010). “The Threat of Shrinking Agency and Free Will Disillusionism.” In W. Sinnott-​Armstrong and L. Nadel, eds. Conscious Will and Responsibility. New York: Oxford University Press, 173–​188. Nichols, Shaun. (2007). “After Incompatibilism: A Naturalistic Defence of the Reactive Attitudes.” Philosophical Perspectives 21: 405–​428. Pereboom, Derk. (2001). Living Without Free Will. Cambridge, MA: Cambridge University Press. Pereboom, Derk. (2007). “Hard Incompatibilism.” In Fischer, Kane et al., 85–​125.

Moral Responsibility and Existential Attitudes    541 Pereboom, Derk. (2013). “Optimistic Skepticism About Free Will.” In Russell and Deery (2013), 421–​449. Oxford: Oxford University Press. Pereboom, Derk. (2014). Free Will, Agency, and Meaning in Life. Oxford: Oxford University Press. Pereboom, Derk. (2017). “Skeptical Views About Free Will.” In K. Tempe, M. Griffith, and N. Levy, eds. The Routledge Companion to Free Will. New York & London: Routledge. Russell, Paul (1992). Strawson's Way of Naturalizing Responsibility. Ethics 102 (2):287–​302. Russell, Paul. (2000/​2013). “Compatibilist Fatalism: Finitude, Pessimism and the Limits of Free Will.” Reprinted in Russell and Deery (2013), 450–​468. [Reprinted in Russell: 2017b.] Russell, Paul. (2002). “Pessimists, Pollyannas and the New Compatibilism.” In Robert Kane, ed. The Oxford Handbook of Free Will. New York: Oxford University Press, 229–​256. [Reprinted in Russell: 2017b.] Russell, Paul. (2008). “Free Will, Art and Morality.” Journal of Ethics 12: 307–​325. [Reprinted in Russell: 2017b.] Russell, Paul. (2017a). “Free Will Pessimism.” In David Shoemaker, ed. Oxford Studies in Agency and Responsibility, Volume 4. Oxford: Oxford University Press, 93–​120. [Reprinted in Russell: 2017b.] Russell, Paul. (2017b). The Limits of Free Will: Selected Essays. New York: Oxford University Press. Russell, Paul, and Oisin Deery, eds. (2013). The Philosophy of Free Will: Essential Readings from the Contemporary Debates. New York: Oxford University Press. Schlick, Moritz. (1939). Problems of Ethics. Trans. by D. Rynin. New York: Dover. Shabo, Seth (2012) “Where Love and Resentment Meet: Strawson’s Intrapersonal Defence of Compatibilism,” Philosophical Review 121: 95–​124. Skinner, B. F. (1948). Walden Two. Toronto: Macmillan. Skinner, B. F. 1971. Beyond Freedom and Dignity. Harmondsworth: Penguin. Smilansky, Saul. (2000). Free Will and Illusion. Oxford: Oxford University Press. Smilansky, Saul. (2001/​ 2013). “Free Will: From Nature to Illusion.” Proceedings of the Aristotelian Society 101: 71–​95. Reprinted in Russell and Deery, 379–​399. Sommers, Tamler. (2007). “The Objective Attitude.” The Philosophical Quarterly 57: 321–​341. Strawson, P.F. (1962/​2013). “Freedom and Resentment.” Reprinted in Russell and Deery (2013), 63–​83. Strawson, Galen. (1994/​ 2013). “The Impossibility of [Ultimate] Moral Responsibility.” Reprinted in Russell and Deery (2013), 363–​378 Strawson, Galen. (1998, June 26). “Luck Swallows Everything,” Times Literary Supplement. Wallace, R. Jay. (1994). Responsibility and the Moral Sentiments. Cambridge, MA: Harvard University Press. Waller, Bruce. (2011). Against Moral Responsibility. Cambridge, MA: MIT Press. Watson, Gary. (1986). “Review of Daniel Dennett, Elbow Room: The Varieties of Free Will Worth Wanting.” Journal of Philosophy 83(9): 517–​522. Watson, Gary. (1987/​2013). “Responsibility and the Limits of Evil.” Reprinted in Russell and Deery, 84–​113. Williams, Bernard. (1985). Ethics and the Limits of Philosophy. With a commentary by A. W. Moore and a foreword by Jonathan Lear. London and New York: Routledge (Routledge Classics edition 2011). Williams, Bernard. (1996/​ 2006). “The Women of Trachis: Fictions, Pessimism, Ethics.” Reprinted in The Sense of the Past. Ed. With an Introduction by M. Burnyeat. Princeton, NJ: Princeton University Press.

CHAPTER 26

Rel ationsh i ps a nd Resp onsi bi l i t y Dana Kay Nelkin

1.  Introduction Philosophical discussion of moral responsibility over time has included periods of sustained focus on the metaphysics of responsible agency and periods in which attention has broadened to include the role of responsible agency in larger social and political questions, among others. In recent decades, there has been a particular interest in the variety of ways in which responsibility appears inextricably linked with human relationships. While there is still much productive discussion of the metaphysics of responsible agency, Peter Strawson’s “Freedom and Resentment” (1962) is often taken as the launching point of the current wave of interest in the topic. In that influential article, Strawson recommends that in investigating the topics of freedom, punishment, moral condemnation and approval, we should: . . keep before our minds something it is easy to forget when we are engaged in philosophy, especially in our cool, contemporary style, viz. what it is actually like to be involved in ordinary interpersonal relationships, ranging from the most intimate to the most casual. (1962: 77)

While the rest of Strawson’s argument is a matter of intense interpretative debate, this prescription is very clear and seems to have been well followed. In this chapter, I examine three debates in which the connection between responsibility and relationships is at stake: the debate about whether valuable human relationships such as friendship in some way presuppose responsible agency on the part of their participants, the debate about whether the conditions of responsible agency must make essential reference to relationships, and the debate about the ethics of blame.

Relationships and Responsibility    543 Before getting into these debates in more detail, it will be helpful to clarify the relationships among the concepts of responsibility, free agency, and blameworthiness and praiseworthiness with a particular eye to the roles they play in the debates at hand. Free agency is often taken to be a necessary and constitutive condition for responsible agency and for being an appropriate candidate for blame and praise, and we will see this point presupposed in some lines of reasoning in what follows. Importantly, the notion of responsible agency at work in many of these debates is a backward-​looking one, associated with desert of negative or positive responses for an action or omission, and one that is associated with our practices of holding each other responsible or to account. It is also important to note that it is generally simply assumed that being responsible is necessary for blameworthiness and praiseworthiness, and sometimes responsible agency is itself understood in terms of being a candidate for appropriate blame and praise.

2.  Valuable Personal Relationships and Responsible Agency A striking claim that has been defended in a variety of ways is that without responsible agency we simply cannot engage in the kind of valuable interpersonal relationships we take for granted, such as true friendship and romantic loving relationships. This claim is sometimes put in terms of the kind of free agency associated with responsible agency; in other words, if we lack the free agency, or the abilities that are necessary, and indeed, on many views, constitutive of being responsible, then we will not be able to form genuine friendships or romantic partnerships.1 Writing about our interest in freedom and taking up Strawson’s thought, Susan Wolf, in describing the worries of many, writes that “the most gruesome difference between [a world in which we give up our practices that depend on the assumption that we are free agents] and our world would be reflected in our closest human relationships” (1981: 291). Although we could form “some sorts of associations that could be described as relationships of friendship and love,” the words “friendship” and “love” would take on “a hollow ring.” We would inhabit “a world of human isolation so cold and dreary that any but the most cynical must shudder at the idea of it” (291). Robert Kane also lists friendship, together with love, as one of ten items that hang in the balance in figuring out if we are free in the sense of being undetermined originators of our action.2 Interestingly, Wolf ’s and Kane’s reasoning to a similar conclusion are completely different. The reasoning Wolf lays out connects friendship and freedom via the disposition

1 

This section draws in part from material presented in Nelkin (2014a, 2014b, and 2015). Kane takes it that the same freedom that is required for friendship (or a particularly valuable sort) is the same as is required for moral responsibility. Throughout the chapter, I will assume that the notion of freedom at issue is the freedom required for moral responsibility, and desert of praise and blame. 2 

544   Dana Kay Nelkin to the reactive attitudes, such as resentment, indignation, gratitude, and forgiveness, which themselves presuppose free and responsible agency. Without our being susceptible to the reactive attitudes, we could not engage in true—​or at least particularly valuable kinds of—​friendships or reciprocal loving relationships. And since the reactive attitudes themselves are appropriate only if their targets are free agents, true friendship that does not rest on an illusion requires free agency. In contrast, Kane reasons that the relationship itself is more valuable, or valuable in a special way, in virtue of its being freely chosen. Let us call the first way of connecting responsibility-​relevant freedom and friendship “the Reactive Attitudes Connection.” The idea here is that without a susceptibility to the reactive attitudes, we could not have a truly personal relationship with another. Let us call the second way of connecting responsibility-​relevant freedom and friendship “the Free Choice Connection.” The idea is simply that freely chosen relationships have a value that can’t be realized otherwise, where that free choice is the same sort of free choice that is required for, and indeed partly constitutive of, robustly responsible agency. Recently, there has been discussion of a third way of connecting the freedom required for responsible agency to friendship, among other special relationships: Friendship is partly defined by the special obligations friends have to one another. Insofar as having obligations requires an important kind of free will, control, or ability that is constitutive of responsible agency, then, friendship itself will require that sort of free will, control, or ability. Call this the “Special Obligations Connection”. Challenges have been raised to each of these lines of reasoning, and a positive case for the preservation of valuable relationships even in the face of skepticism about free will and moral responsibility has emerged. In the remainder of this section, I set out in more detail and assess each of the three lines of reasoning below.

2.1  The Reactive Attitudes Connection At the same time that Strawson introduced the term “reactive attitudes,” he argued for its importance to the debate about free will precisely by claiming their essential connection to personal relationships (61–​62). Without these attitudes that are reactions to the good or ill will of others and through which we demand good will from those with whom we stand in a variety of personal relationships, we are left, on Strawson’s view, to adopt the “objective” attitude, to “see him, perhaps as an object of social policy; as a subject for what, in a wide range of sense, might be called treatment; as something certainly to be taken account, perhaps precautionary account, of; to be managed or handled or cured or trained; perhaps simply to be avoided” (66). If these are our only alternatives, then it is a short step to the conclusion that we need the sort of free—​and responsible—​agency presupposed by the reactive attitudes to have genuinely valuable relationships such as friendship. (It might be that we could be under an illusion that we are responsible and yet have the attitudes that presuppose that we are; but in this case,

Relationships and Responsibility    545 our relationships would be at best in a precarious situation and based on an illusion, decreasing their value in a different way.) But that the objective attitude and the reactive attitudes that presuppose responsibility are exhaustive alternatives is an assumption that has not held up well. In fact, Derk Pereboom has offered a rich and subtle account of a whole set of other options, or what he calls “analogues” of the reactive attitudes.3 Even if we were to set aside the reactive attitudes that presuppose responsibility, many other personal attitudes remain, including moral sadness, disappointment, resolve, and analogues to forgiveness and gratitude that retain many positive features of each while eliminating any presupposition of responsibility. And yet, this is not the end of the debate. At this point, there remains a question of whether we can eliminate our susceptibility to, or even regularly disavow, the reactive attitudes and still maintain our personal relationships. If we cannot—​even if only because of our human psychological limitations—​this suggests that such susceptibility has significant value, contingent though it may be. And, even if we settle this (at least partly) empirical question, there is a further question: Are the admittedly personal relationships that would remain without the specifically reactive attitudes as valuable as relationships in which we are susceptible to such attitudes? I address each of these questions in turn. Taking up the Reactive Attitudes Connection, Seth Shabo has recently addressed the first question with a qualified “no” (2012).4 While there may be exceptional people who can disavow the reactive attitudes and still maintain caring personal relationships, most of us cannot do so. Shabo argues that genuine loving relationships, such as romantic ones, are ones in which the participants care not only for the other person, but also care about the other’s attitudes and feelings about them. That is, if one is not susceptible to hurt feelings if, say, one’s partner is unfaithful, or shows one disrespect or unkindness, then this is a sure sign that one does not have the kind of caring relationship that is characteristic of romantic love. One has to care about the other’s feelings and attitudes displayed in his or her behavior, and to do this one must be at least susceptible to hurt feelings. As Shabo puts it, one must be susceptible to “taking things personally,” and having a certain degree of “emotional vulnerability” (112). In turn, hurt feelings “often beget resentment” (114). While the first connection—​between a caring relationship and taking things personally—​is claimed to be conceptual, the second—​between taking things personally and the reactive attitudes—​is a psychological generalization. An example is meant to support this: Learning from a trusted friend that her husband openly scoffs at her artistic ambitions behind her back, Caitlin’s feelings are badly hurt . . . As her initial 3 

See, for example (2001, 2007, and 2009 and 2014). See also Sommers (2007) and Milam (2016). See also Justin Coates (2013) for a related defense of the Reactive Attitudes Connection that takes the reactive attitudes to be essential to personal relationships because they embody normative expectations. See Pereboom (2018 and forthcoming) for a reply. 4 

546   Dana Kay Nelkin embarrassment subsides, the first stirrings of resentment set in. He has been so supportive of their friends’ creative endeavors, why not hers? When she thinks that this is her husband, her resentment shades into moral indignation. (115)

We are then asked to imagine that Caitlin wants to disavow such feelings of resentment and indignation because she believes them to be morally objectionable. But it seems that she cannot do so without “emotionally divesting” in a way that not only sets aside the feelings of resentment and indignation, but also the vulnerability to the hurt feelings (115). This case invokes strong intuitions that Caitlin’s feelings of hurt and resentment are intertwined, and that what she would attempt in order to remove resentment would end up with a more global emotional divestment. One response, given by Pereboom (2014a, 2014b), is to provide reason to think that our psychological flexibility is greater than Strawson and Shabo believe, because in many cases we have been able to disavow and even eliminate our reactive attitudes to criminals, the insane, and children, among others. These cases from history are powerful, though notably these are each in a way special cases of relationships that differ from friendship and romantic love. It remains an open empirical question to what extent we could either eliminate or disavow the reactive attitudes when wronged in the way that Caitlin seems to be, while keeping our vulnerability to hurt feelings and our ability to take things personally (assuming that this is essential to a meaningful relationship). Another response is to present cases that point in the other direction than the Caitlin case by supporting the idea that one can take things personally without having resentment. Pereboom (2014a) offers a case in which many are able to both take things personally and disavow and even eliminate reactive attitudes: that of parents of teenagers who, in a normal developmental stage, show disrespect. Parents can feel quite seriously hurt feelings and take things personally without feeling resentment or other reactive attitudes. This case is powerful, but it has limitations. We can ask—​as Pereboom does—​whether this special ability to experience feelings other than resentment and indignation is unique to the parental relationship. It also may be special to the relationship between parents and not fully mature adult children, in particular, insofar as there seems to be growing recognition of at least partial excuse of the behavior of adolescents.5 There may even be some reason to view these actions as not wrong, or not as violations of obligations, for the very reason that teenage behavior is seen as a normal and healthy developmental stage. Thus, though the case provides some evidence that we are psychologically capable of setting aside resentment in personal relationships, how far it generalizes is also an open question.6 Per Milam offers a variety of additional cases, 5 

See Brink (2004), for example. It is an interesting empirical question just how the process works for many parents of teenagers who behave in disrespectful ways. If (as seems sometimes to be the case) the recommended thought processes include thinking of the behavior as normal and healthy and ultimately in the service of establishing the adolescent’s own identity, and so on, the fact that it is disrespectful becomes less salient. These sorts of 6 

Relationships and Responsibility    547 as well, including ones of friendship, to elicit the intuition that one might excuse one’s friends for their transgressions (such as dropping plans at the last minute) and so cease to feel resentment, but remain vulnerable to hurt feelings related to them (2016). These intuition pumps might gain some of their power by the fact that we take our friends to be responsible in general and we are vulnerable to resentment, even while we excuse for specific transgressions on specific grounds (such as a recognized social anxiety disorder). Still, these cases and others like them provide a serious counterweight to ones like the Caitlin case.7 There are also additional limitations to the Caitlin case. First, as mentioned, even Shabo does not claim that eliminating or disavowing resentment in a personal relationship is impossible: only that it is difficult or impossible for most of us. But there is another limitation, as well. The very case that Shabo uses to make the psychological generalization plausible has a particular feature that is not present in a number of cases of hurt feelings, a feature that may be quite relevant to determining how far our reaction to it should generalize. The case is one in which Caitlin is wronged by her husband, and one in which we naturally think resentment is warranted. It is one in which we ourselves may be indignant on her behalf—​and this has nothing to do with our susceptibility to hurt feelings in the case; we don’t even know Caitlin’s husband. But, more importantly, there are many cases of hurt feelings that do not involve wrongs at all. Take a case in which one’s friend postpones a dinner in order to help another friend. A third party might feel sorry for one, but not indignant, and one might be hurt, even as one recognizes that one’s friend did nothing wrong and fails to feel resentment. Or take cases of jealousy, in which the object of a person’s love does nothing wrong. Or cases in which a child does not want to spend time with another child. These seem to be common sorts of cases in which one can have genuinely hurt feelings and take things personally, and yet not have the reactive attitudes that presuppose responsible agency and blameworthiness. If this is right, then we have additional reason to reject the idea that valuable human relationships are only possible if we have the capacity to feel resentment or other responsibility-​presupposing attitudes. Neither resentment, nor the capacity for it, is what is really doing the work here.

thought processes would seem to have some effect not only in reducing resentment, but also in taking the particular behavior personally (since the behavior is, in a sense, “not about you” the parent). Still, this is just one route to the reduction of resentment, and there may be others that would leave one’s “taking it personally” fully intact. Further, even if parents of teens with reduced resentment also take the relevant behavior “less personally,” it is a plausible hypothesis that this need entails nothing about a lower quality of the personal relationship. To the contrary, it is often recommended to parents that they take this approach. See, for example, Alfie Kohn (2005). 7 

Milam also offers an intriguing if only “suggestive” argument against the other main assumption in Shabo’s version of the Reactive Attitudes Connection, namely, that we must take things personally in order to have valuable personal relationships. He considers a daughter who ceases to take personally her mother’s manipulative behavior, but who he claims could plausibly be thought to have a valuable personal relationship with her and grieve when she dies (2016: 112).

548   Dana Kay Nelkin There are various ways to respond to these objections on behalf of the Reactive Attitudes Connection, and I focus on one recent and novel approach here. It is to suggest that hurt feelings are themselves a kind of reactive attitude and one that tracks responsible agency and blameworthiness of the relevant sort. Thus, hurt feelings can themselves do the work at issue in the Reactive Attitudes Connection. David Shoemaker (2019) has recently argued for the claim that hurt Feelings are a plausible candidate for being a reactive attitude, while pointing out the often-​overlooked fact that Strawson includes hurt feelings in his list of reactive attitudes before going on to focus most of his attention on resentment. Thus, it could be said even of the cases in which people take things personally despite lacking resentment that they still include reactive attitudes that are appropriate when their objects are responsible agents. Once we expand beyond resentment to include hurt feelings in the class of reactive attitudes, it becomes more difficult to find cases in which one can “take things personally” without having a susceptibility to at least some reactive attitudes. This is an interesting suggestion, and hurt feelings are indeed an overlooked phenomenon in recent philosophical discussions. Yet, there are good reasons behind the trend of dropping “hurt feelings” from the list of central reactive attitudes and focusing more exclusively on resentment, indignation, and guilt. One reason is that, as Shoemaker notes, accepting that hurt feelings themselves are a kind of reactive attitude prevents us from accepting an account of the conditions of responsibility and blameworthiness that are independent of our responses. In other words, then it does not seem possible to offer any set of necessary and sufficient conditions independent of the reactive attitudes in virtue of which it is the case that one is responsible or blameworthy. This is because hurt feelings seem to be appropriate in cases that don’t conform to any of the current contenders of independent conditions for responsibility and blameworthiness, as the cases earlier illustrate. For example, hurt feelings can be appropriate even if their object has no control over their actions or attitudes, even if their object displays no ill will, and so on. Trying to find a single set of conditions that makes resentment, indignation, and guilt appropriate is one thing; finding a set of conditions that makes all of those and hurt feelings appropriate would be quite another and clearly even more difficult. Thus, the hope of any unified account of responsible and blameworthy actions in response-​ independent terms will be undermined by the inclusion of hurt feelings. While some, including Shoemaker (2017) are willing to pay it, it is a high cost. Paying it would require giving up what seems the most natural view, namely, that there is such a unified set of conditions, independent of our responses, that we can point to when justifying our treating people who meet them in certain ways that are sometimes quite burdensome.8 On the alternative response-​dependent view, we would have to say that we are justified because it is somehow appropriate—​given how humans are “built”—​to respond to them in these ways. And this seems at the least 8  See Shoemaker (2017) for a defense of a response-​dependence view, as well as acknowledgment that our natural assumption is that blameworthiness is a response-​independent property that our responses track.

Relationships and Responsibility    549 to omit the intuitively plausible idea that whether responses are appropriate depends on some independent facts about the agent’s behavior in virtue of which they are blameworthy. Further, it seems that hurt feelings simply do not fit with other aspects of our practices of holding people responsible in the way that resentment does, such as the ideas of desert of negative responses or holding people to account. This observation provides another reason to think that hurt feelings are not properly thought of as tracking responsible and blameworthy agency. Again, if so, then the cases earlier in which people are vulnerable to hurt feelings without resentment succeed in showing that we can take things personally without having attitudes that presuppose responsible and blameworthy agency. In response, Shoemaker argues that when we cause hurt feelings, it is often appropriate to apologize and experience guilt feelings, and these are indeed aspects of our responsibility practices. This is an important observation. If the patterns of responses to recognizing hurt feelings are very similar to those that recognize resentment, then that would bolster the case for treating hurt feelings as reactive attitudes that also track responsible agency. But insofar as apology is an appropriate response to merely having inadvertently caused hurt feelings, it is not clear that we are employing the same notion of “apology” as the one we invoke when we think it apt to apologize for moral wrongs. And it might be that even if there is a univocal sense of apology for which apology is appropriate in both cases, we cannot read from the appropriateness of apology that the apologizer is responsible and blameworthy in a sense that makes holding them to account or responding with deserved negative attitudes appropriate. For since there is often a risk that we in fact did do something wrong in causing the hurt feelings (or wrongly omitted to prevent it from happening), there could be something appropriate about assuming that we are responsible, even if we are not.9 Something similar can be said of guilt feelings. There might be something understandable and perhaps even all things considered appropriate about feeling guilt for hurting others’ feelings given that quite often we can and ought to avoid doing so, even if those feelings are in some way strictly inapt in a case where one did not violate any obligation in the case. Thus, while the move to an expansive category of the reactive attitudes that would help to vindicate the Reactive Attitudes Connection is intriguing and certainly does important work in drawing our attention to an under-​ theorized class of emotions, it requires paying a high cost and it appears that its apparent payoffs can be explained in alternative ways. Let us turn then to a very different line of reasoning that responsible agency—​or at least an important constitutive aspect of it—​is essential for valuable human relationships such as friendship.

9  See Mason (in preparation) for the view that apology is a univocal notion of “taking blame,” but which refers to a phenomenon that has two types, one that responds to culpable wrongdoing and one that does not. See also Wolf (2001) for a related argument in a different context that it can be a virtue to take responsibility for more than one is actually responsible for.

550   Dana Kay Nelkin

2.2  The Free Choice Connection The reasoning of the Free Choice Connection is simple, and asserts that genuine friendships are ones that are freely chosen, thereby requiring free choice of the kind required for morally responsible agency. As Kane expresses the idea: There is a kind of love we desire from others —​parents, children (when they are old enough), spouses, lovers and friends—​whose significance is diminished . . . by the thought that they are determined to love us entirely by instinct or circumstances beyond their control or not entirely up to them . . . To be loved by others in this desired sense requires that the ultimate source of others’ love lies in their own wills. (1996: 88).10

This has some appeal. After all, it seems that a relationship that one finds oneself stuck with, or that one sees that one’s partner has been stuck with is less than ideal. But perhaps this is a matter of how the idea is expressed. As against the Free Choice Connection, Derk Pereboom argues that at least in many cases, we prefer that love and friendship not be a matter of deliberate decision, and that in cases in which it seems that causal determination of the will to love and friendship would be undermining of its value, what is really problematic is only certain kinds of causal determination that bypass our genuine recognition of the value in others. Thus, skepticism about free will is not incompatible with genuine friendship.11 While this is a compelling thought, there might still be a truth to be found in the spirit of the Free Choice Connection. While we want others to feel emotions and experience attitudes of love and care without choice, relationships involve more than emotions and experiences, and it might be that other aspects of relationships require a kind of control that is constitutive of responsible agency. We will see in the next section that the Special Obligations Connection provides a way of developing this thought. While feelings and attitudes might have no direct connection to free and responsible agency, a case can be made that the kinds of obligations and reasonable expectations that define special relationships over time do have such a connection.

2.3  The Special Obligations Connection The Special Obligations Argument goes as follows: (1) friendship requires special obligations or duties. (2) special obligations, being obligations, require free agency to meet them.

10  11 

See also Ekstrom (2000) and Anglin (1990) for arguments in a similar spirit. See Pereboom (2014a and forthcoming).

Relationships and Responsibility    551 Therefore, (3) friendship requires free agency. This argument captures what we can call the “Special Obligations Connection” of friendship and freedom. As we have seen, it shares something of the spirit of the Free Choice Connection, but locates the free agency at issue, not in the choice to love another or feel certain emotions, but rather in the conditions for possession of the obligations defining of friendship once begun.12 But it is also possible to see it as sharing an insight with the Reactive Attitudes Connection. For we can accept the reasoning that we do not need the reactive attitudes per se to have friendships, but at the same time think that we need something that plays a key role played by such attitudes. In particular, one role that the attitudes have seemed to play is in the appropriate making and responding to certain demands special to the relationship. It is appropriate for friends to ask certain things of friends that it wouldn’t be appropriate to ask of strangers. For example, it seems appropriate for a friend to ask for a friend’s support while in the throes of grief, or to lend an ear during a difficult time. Reactive attitudes are sometimes characterized as embodying demands.13 But even if we could imagine a friend who reacts to her friend’s voluntary absence during a difficult time without resentment or other responsibility-​ or free will-​entailing attitudes, we might still think that she would be entitled to demand assistance of her friend in a way that she wouldn’t when it comes to strangers on the street. She could point to their relationship as grounds for it being the case that her friend ought to help her in the situation. Thus, one might think that the defenders of the Reactive Attitudes Connection are onto something, but that rather than appeal to the reactive attitudes themselves to make the connection between freedom and friendship, it would be better to focus on what important role in friendship they (contingently) play, namely, that of embodying demands. If one needs a kind of free and responsible agency for such demands to be apt, then we have a similar line of reasoning, but one that appeals to the arguably more fundamental idea of demands.14 It is now a short step to the Special Obligations Argument, because it is plausible that the relevant demands are in place only if, and because, their targets have obligations to do what is demanded, or to live up to the standards to which they are rightly held. Thus, this argument can preserve insights from each of the others. To elaborate on the reasoning, note that premise (1), the idea that friendship requires special obligations, makes sense of many of our practices and assumptions. We often

12 

Further, if one takes it that the acquisition of such obligations is itself a matter of voluntary action, as some views do, then there is yet another possible place for free choice of a sort to make its case. 13  For example, Gary Watson writes, “The relevance of moral understanding to the expressive theory is this: The negative reactive attitudes express a moral demand, a demand for reasonable regard” (1987/​ 2004: 229). And see R. Jay Wallace (2009) for a similar understanding of the reactive attitudes, inspired by Strawson. Coates (2013) builds his defense of the Reactive Attitudes Connection on this idea. 14  See Nelkin (2015) for more detailed development of this line of reasoning.

552   Dana Kay Nelkin simply assume that we have special obligations to friends that we do not have to everyone, for example. We typically think that we ought to be there for our friends when they receive bad news, say, whereas we do not typically think this is the case when it comes to strangers. In fact, the premise is most often criticized by consequentialists, whose moral theory would seem to license partiality to our friends only insofar as it is a way of maximizing utility, or well-​being, or some other goods. Thus, assuming we can still have friendship if consequentialism is the true moral theory, then there is no special obligation that one has to one’s friends; there is only a general obligation to maximize good consequences, and this obligation may in some circumstances be fulfilled by treating one’s friends in special ways.15 While this is an interesting debate, for current purposes it is worth noting that the apparent existence of special obligations is one of the most challenging objections to consequentialism. Let us provisionally assume that (1) is true and turn to focus for now on premise (2) before returning briefly to premise (1). If the obligations of friendship are of the same sort or sorts as (other) moral ones, then the reasons for premise (2) in the Special Obligations Argument will just be the kinds of reasons that have been given for the thesis that moral obligations require freedom of the agents to which they apply. So let us turn to the debate over the truth of that thesis which is supported by a natural appeal to the Ought Implies Can principle, which states that if one ought to act in a certain way (or refrain from acting), then one can so act. The Ought Implies Can principle is sometimes taken to be axiomatic, though it has been argued for in a variety of ways. For example, the concept of obligation is itself a concept of action-​guidingness, and this would seem to have application only if actions could be so guided. Additionally, the principle explains our intuitions about a variety of cases in which, upon finding out that an agent cannot fulfill what we thought was an obligation, we retract our judgment that the obligation is in place and instead in many cases attribute a “second-​best” obligation.16 (For example, the plane on which you are flying to a friend’s wedding is hijacked; you are no longer obligated to be there on time. Perhaps you acquire an obligation to let your friend know why you missed her wedding when you eventually land.) There is a rich debate about the principle, but it is important to note that even if it is accepted, premise (2) still requires defense. For we must ask whether, if the Ought Implies Can principle is true, the “can” is one that requires the kind of free agency constitutive of responsible agency.17 To see how this question arises, consider the idea, familiar from 15  An exception to this commitment on the part of consequentialists could be a two-​level consequentialist account, such as that described by John Rawls (1955) or that put forward by Manuel Vargas (2013), according to which there could be special obligations. So the main objections to such special obligations come from those who take consequentialism to be true at the first order of moral theory. 16  See Nelkin (2011a: 111–​112), Howard-​Snyder (2006: 236), Brink (1994: 231), and Vranas (2007: 181–​182). 17  For a recent argument for a modified version of the principle, see Peter Graham (2011), and for an argument against it based on Frankfurt cases, which were originally intended to cast doubt on the claim that responsibility requires an ability to do otherwise, see Fischer (2003).

Relationships and Responsibility    553 the free will literature, that if determinism is true, then no one can do otherwise than they actually do. In other words, if determinism is true, then no one has the ability to do otherwise. If that is the case, and one does not do as one ought, then one thereby violates the Ought Implies Can principle. One ought to have done something that one could not have done. Thus, if the Ought Implies Can principle is true, and if determinism is true and precludes the relevant ability to do otherwise, then we have no obligations that we do not in fact meet.18 If this reasoning is sound, then it seems that the Ought Implies Can principle entails both the existence of alternate possibilities for unmet obligations and the rejection of determinism, both of which are components of some of the most influential accounts of free agency. But, as has been pointed out, even if this reasoning were sound, the acceptance of indeterminism would not by itself be sufficient for the ability to do otherwise that appears necessitated by unfulfilled obligations. As Pereboom writes, Furthermore, one might also claim that if our choices and actions are partially or truly random events, then we could never do otherwise by the sort of agency required for it to be true that we ought to do otherwise. But if it is never true that one ought to do otherwise, what would be the point of a system of moral “oughts”? (2001: 143).

Or, as Haji puts the point, . . . the sort of control agents must exercise over their actions if these are to have one or more of these normative statuses—​the sort encapsulated by the deontic principle that “ought” implies “can”—​is not the sort that they possess in “indeterministic” worlds of certain varieties, if particular versions of this deontic principle are true. Such worlds are just as threatening to moral anchors [e.g., obligation and permissibility] as are deterministic ones. (1998: 176)

In other words, adding indeterminism to the world by itself does not generate the ability to do otherwise that the Ought Implies Can principle entails for unfulfilled obligations. That ability cannot be the bare physical or metaphysical possibility that the world might turn out so that by luck we do something else. Whatever it is, the ability is one that itself requires control by the agent. It has to be in my control that I do what I ought to do if I am able to do it in the relevant sense. The kind of control that is required to do what you ought, even when you don’t, looks very much like the kind of control that is required to be blameworthy for an action. On a view I favor, the nature of the kind of freedom needed for responsibility, what one needs to act freely, is the ability to do the (or a) right thing for the right reasons. Call 18  See Haji (1998) for this reasoning. In fact, Haji goes further, arguing that one also needs alternate possibilities when one does meet one’s obligations (188). See also Pereboom (2002 and 2014a) for the articulation of similar reasoning for why determinism threatens obligations (or at least those we do not meet).

554   Dana Kay Nelkin this account the “rational abilities view.” This view entails that when one does not do the right thing, or does it for wrong reasons, one must have the ability to do otherwise if one acts freely. But one does not need the ability to do otherwise to act freely when one does the right thing for the right reasons. While I believe that there are strong independent reasons for accepting the view (including its intuitive plausibility), it is also mutually supporting of the reasoning just given for the idea that the Ought Implies Can principle implies a kind of ability to do otherwise that involves control. To see this, consider the following derivation, connecting obligation, blameworthiness, and alternate possibilities: (i) If S is blameworthy for having performed action a, then S ought not to have performed action a. (ii) If S ought not to have performed an action a, then S could have refrained from performing action a. Therefore, (iii) Principle of Alternate Possibilities-​Blame: A person is morally blameworthy for what he has done only if he could have done otherwise. (Nelkin 2011a: 100–​101)

If one can only be morally blameworthy for an action when one does something one ought not do, then the alternate possibilities requirement on blameworthy action can be seen to rest on the more fundamental thesis that when one fails to do as one ought, one could have done as one ought.19 In other words, the alternate possibilities requirement for unfulfilled obligations that is captured in the Ought Implies Can principle can be seen to ground the alternate possibilities requirement for blameworthy action. At the same time, there is no sound parallel derivation to be found for the idea that one needs alternative possibilities for praiseworthy action. And this, too, fits well with the view of the freedom required for responsibility as a substantive ability to act for good reasons. For that view implies an asymmetrical answer to the question of whether alternate possibilities are required for blameworthy and praiseworthy actions. If you have the substantive ability and exercise it, you need no alternative; if you have the

19 

One might argue that the possibility of excused wrongdoing suggests that the guiding principle here is really Blameworthiness Implies Can rather than Ought Implies Can. (Thanks to David Brink for raising and discussing this issue.) This is a large and subtle issue. I believe that excused wrongdoing is not ruled out on the view that Ought Implies Can, where “can” is a “can” of free agency. Violation of an obligation is only a necessary and not a sufficient condition for blameworthiness on the picture I am advancing. While it is true that along with such a violation comes free agency, one might be missing other elements relevant to blameworthiness. And notably, blameworthiness comes in degrees, and a number of factors might mitigate blameworthiness even to a vanishing point. One might be free to do as one ought at the same time as it is more difficult to do as one ought than is reasonable to expect, for example. Thus, there are ways to preserve the possibility of excused wrongdoing on this picture.

Relationships and Responsibility    555 substantive ability and fail to exercise it, you need an alternative.20 Thus, the idea that the Ought Implies Can principle entails that ought-​claims apply when one has precisely the freedom to do the right things for the right reasons provides mutual support to both this reading of the Ought Implies Can principle and to the conception of freedom at hand. At the same time, it is an open question whether the truth of determinism is relevant to whether one has the relevant abilities.21 I will set aside this question for now, however, as I believe that the key claim for our purposes about the implications of the Ought Implies Can principle stands without it. The key claim—​that the principle entails that one fails to fulfill an obligation that applies to one only when one could do otherwise in a sense which entails that the failure is in one’s control—​can be accepted by compatibilists and incompatibilists alike. Whether the relevant conception of control requires determinism or not, it is plausible that the Ought Implies Can principle supports the idea that a notion of control at stake in debates about freedom and responsibility is entailed by unfulfilled obligations. The rational abilities view can itself be seen to rest in part on a naturally unified account of obligations—​both fulfilled and unfulfilled—​as requiring a single kind of control captured by the substantive ability to do the (or a) right things for the right reasons. Thus, it fits neatly with, and offers mutual support for, premise (2) of the Special Obligations Argument. But even if one does not accept the rational abilities view, there is good reason to see the Ought Implies Can principle as one that implicates the control needed to do as one ought, whether one does or does not do so.22 If this is a different kind of control than that needed for freedom in the sense required for moral responsibility, then this calls out for explanation.23 Finally, it should be noted that this very line of reasoning for accepting premise (2) can offer one way of casting doubt on premise (1). If one has independent reason for skepticism about the sort of control implicated in (2), and yet endorses Ought Implies Can, then one should doubt the existence of any obligations at all. One way of taking this route, suggested by Pereboom, is to offer an alternative framework so as to try to preserve much of what we take to be captured by both moral “oughts”

20 

Note that this marks a significant disagreement with Haji, who argues that alternate possibilities are required not only for unfulfilled obligations, but also for merely permissible actions, including fulfilled obligations. 21  See Wolf (1992), Nelkin (2011)a, Vargas (2013), and Brink (2021) as examples of views that offer alternative conceptions of the relevant unexercised abilities that do not require the truth of indeterminism. 22  Indeed, one might have other reasons for accepting premise (2). Haji, for example, requires a kind of control that builds in alternate possibilities for each instance of its exercise, but also argues, partly for reasons of symmetry of deontic concepts, that such control is required for all obligations. (See Haji 2002.) 23  For examples of recent discussions that bring together questions about the control required for obligation and that required for responsible agency, see Rudy-​Hiller (2019) and Elzein and Pernu (2019), as well as earlier ones including Pereboom (2001 and 2013) and Haji (1998, 2002, and 2012).

556   Dana Kay Nelkin and the “oughts” of friendship. The framework has two parts. The first is to replace the notion of obligation with that of commitment. Friends can be committed to helping each other in circumstances of special need without it being the case that they are obligated to act in certain ways.24 Commitments of certain kinds seem distinctive to particular kinds of relationships in a way that parallels the idea that particular kinds of obligations are distinctive of particular kinds of relationships. Can this do the work that obligations seemed uniquely able to do in explaining the nature of distinctively valuable relationships and of our attitudes toward them? Perhaps not by themselves. For we might ask how to understand the basis of these commitments or how they might manifest in relationships. On one natural reading, they are a kind of promise, but that would seem to bring the idea of obligation in through the backdoor since promises seem like a paradigm of moral obligation. This is where the second part of Pereboom’s framework comes into play. It is to acknowledge that a failure to honor a commitment is a kind of wrong which a friend who has been thereby been let down can legitimately point out. And while we cannot understand “wrong” in terms of a violation of an obligation, we can instead here appeal to an “axiological ought,” rather than an “action-​guiding ought.” Doing so allows us to express with an “ought” that it is better or best to do something (such as being there for a friend) without being committed to saying that a person ought to do it in a sense that presupposes that doing it is in their control.25 A failure is then wrong in the relevant sense when it falls below a threshold of goodness such that others can legitimately protest one’s action. Whether this approach can work will depend on whether this alternative can be shown to capture enough of what we take to be essential for friendship (assuming that we keep as a fixed point the existence of valuable relationships of this kind).26 In sum, we have seen three possible ways of arguing that responsible agency—​or at least a constitutive aspect of it, namely, a kind of freedom or control—​is necessary for having certain valuable human relationships. The third combines insights from each of the others, namely, the idea that participants in close relationships can legitimately call on each other in special ways, and the idea that a kind of freedom or control is essential to relationships like friendship and others that are partly defined by the special obligations that reflect these participants’ claims. As we have also seen, serious questions can be raised about each of the three approaches.

24 

Interestingly, Arneson (2003) offers the consequentialist who denies that there are special obligations a related replacement, namely, that of “willing” that one provide special help in order to be a good friend. 25  See Pereboom (2014a and 2017), drawing on a framework for an axiological notion of wrongness from Alastair Norcross (2006). 26  See Nelkin (2015) for development of the Special Obligations Connection. See also Coates (2018) for a somewhat related line of reasoning according to which promissory obligations that are central to human relationships require a commitment to responsible agency.

Relationships and Responsibility    557

3.  Are the Conditions of Responsible Agency Essentially Relationship-​Based? Another potential connection between responsibility and relationships is found in the claim that responsible agency can only be understood in terms of particular relationships.27 T. M. Scanlon, for one, offers an account of blame and blameworthiness, central aspects of responsible agency, that make essential reference to relationships. In his view, “to claim that a person is blameworthy for an action is to claim that the action shows something about the agent’s attitudes towards others that impairs the relationship others can have with him or her . . . To blame a person is to judge him or her to be blameworthy and to take your relationship with him or her to be modified in a way that this judgment of impaired relations holds to be appropriate.”28 The account is developed first by applying it to close relationships such as friendship, and then extended to the “general moral relationship” we might have with any fellow person. The appropriate form that blame takes in each case will be different, although anyone can make the same judgment of blameworthiness. So, for example, the kind of appropriate blaming response for someone who has been betrayed by a friend would be a different kind of response from the kind that would be appropriate for a disinterested third party because they have different relationships with the offender. The account has the advantage of accommodating the plausible ideas that blame often takes different forms depending on the blamer’s relationship with the blamed, and that what is appropriate blame from a friend is often of a different kind from the blame appropriate from a stranger. And this is an insight to which we will return in the next section. Yet it does not seem essential to achieve this result either that blame be defined in terms of the blameworthy or that the conditions for blameworthiness include essential reference to relationships. In particular, it is not clear why blame must imply a judgment about an impairment in one’s relationship, or why blameworthiness entails some sort of impairment. For example, consider cases in which one blames one’s child; intuitively, this doesn’t always require a judgment of impairment in the relationship.29 And, moving in the other direction, one might question whether such judgments of impairment must in turn imply judgments of blameworthiness. Consider, for example, a case in which your friend lets you down—​and in a way that specifically violates the standards of friendship. She fails to keep a promise to you in order to help someone else. It seems possible that in such a case you are disappointed and might even re-​evaluate your relationship, but you don’t blame her at all, or even judge her to be blameworthy in an

27 

This section draws in part on Nelkin (2011b). Scanlon (2008: 128–​129). 29  See Wolf (2011) for further discussion. 28 

558   Dana Kay Nelkin intuitive sense.30 Thus, while there is surely something correct about blame having an interpersonal aspect, and how one should blame depends in part on one’s relationship to the offender, we would need more reason than has been given to think that blameworthiness itself is a matter of relationship impairment.31 Another way in which responsibility has been taken to make essential reference to relationships involves thinking of relationships in more general terms. For example, Gary Watson has begun by noting that responsibility in an important sense is always ultimately responsibility “to” someone or some others.32 Thus, argues Watson, a necessary condition for responsibility is that we have obligations to others.33 If this were correct, it would have serious implications for the limits of what we can be responsible for. As Watson himself argues, it would limit responsibility to moral responsibility, since, in other areas, while it might be that we ought to do or think in certain ways, these “oughts” are not obligations to others. Thus, while it might be that we ought not to hold inconsistent beliefs, say, and we ought to try our best in our endeavors, these are not at least first and foremost obligations to others. But it would also seem to limit what we can be responsible for to a subset of moral “oughts.” General moral obligations of beneficence, say, or a moral obligation to forgive someone who has no right to it would be precluded on the grounds that we do not owe these things to any particular others. To successfully defend the conclusion that the conditions of responsibility in the accountability sense are restricted to those in which one has obligations to others, one would have to accept that these apparently moral obligations are not the sorts of things for which one can appropriately be held responsible, and this seems a tall order. Even if we need not make essential reference to relationships in the conditions for moral responsibility, there is still an important set of observations that are key to the arguments we have seen earlier. It is that our practices of holding one another responsible are interpersonal in a variety of ways, and what sorts of blame or praise are appropriate depends on our relationships with the targets of our responses. This takes us to what is known as “the ethics of blame.”

30  It is important to note that Scanlon allows for “drifting apart” cases that do not involve blame, but these are explicitly described as cases in which standards of the relationship are not violated. 31  For interesting further arguments that Scanlon’s relationship account of blame does not plausibly extend to the case of strangers, see George Sher (2013) and Eric Brown (2017). 32  Watson (1996) distinguishes between responsibility in the “accountability” sense and responsibility in the “attributability” sense. Responsibility in the accountability sense is the one that is a matter of holding people responsible and that is associated with the reactive attitudes and desert, according to Watson, whereas responsibility in the attributability sense is a matter of attributing ends to people, and in virtue of their ends, judging them to be virtuous or vicious. It is responsibility in the former sense with which Watson is concerned here. 33  See also Darwall (2006).

Relationships and Responsibility    559

4.  Relationships and the Ethics of Blame Before turning to the debate about the ethics of blame, it is worth noting that there is an equal reason to consider the ethics of praise, but to this point, most discussions have focused on the ethics of blame.34 I will set aside this interesting issue here for reasons of space. For the same reason, I will also set aside a lively and important debate about the ethics of forgiveness, which is closely related insofar as forgiveness has been thought to be one way of ceasing to blame (or at least to mitigate it).35 To begin our exploration of the ethics of blame, recall a set of observations that are central to motivating Scanlon’s “relationship” account of blame, for example. It is plausible that the kind of blame that is morally permissible and appropriate on the part of a spouse in response to the other spouse’s wrongdoing is, at least often, of a different kind than that which would be morally permissible or appropriate from a stranger. Or, similarly, parents seem to be in a position to blame their children in ways that others are not entitled to do. As mentioned earlier, these are observations that can be accepted independently of the rest of Scanlon’s account, and indeed are accepted by theorists who otherwise differ widely about the conditions for moral responsibility and blameworthiness. In fact, a burgeoning literature on the “ethics of blame” takes these as important cases that illustrate that whether (and how) it is all things considered permissible or appropriate to blame hinges on much more than simply whether a person is in fact blameworthy. Whether someone is in fact blameworthy is but one consideration; call this consideration the condition of “fittingness.” Blame “fits” when a person is in fact blameworthy for an action. But what relationship the blamer stands in to the target of the blame also seems to determine whether and how blame is appropriate. The examples before of spouses or parents and children are obviously ones of special relationships, and in each of these cases it has been argued that being in a special relationship gives one a kind of standing to blame and to hold responsible that others lack.36 For example, some have argued that there is a “business” condition for entitlement to blame, according to which a friend can blame a person for telling a secret but a stranger on the bus 34  For a recent exception, see Emily Bingeman (in preparation), which identifies at least some important reasons for thinking that we can act wrongly by praising in certain circumstances just as we can by blaming in others. 35  See Milam (2021) for a very helpful overview, and see Chaplin (2019) for one recent discussion that parallels some of the issues discussed in this section. 36  Interestingly, standing is sometimes understood as a requirement for engaging in an activity at all; at other times it is understood as a requirement for engaging in the activity in appropriate ways. The legal term of “standing” refers to the former: One cannot bring a lawsuit without having the appropriate status. When questions of standing are raised in discussions of forgiveness, this seems to be the notion of standing at stake. But in discussions of blame, standing is most often used to refer to status needed to engage appropriately in the activity. It is taken for granted that someone can blame without having the relevant standing; what such a person lacks is the ability to blame appropriately. For further discussion of the relevant concept of standing, see Adam Piovarchy (forthcoming).

560   Dana Kay Nelkin who overhears such an offense being committed is not so permitted.37 As Linda Radzik describes the “commonsense moral view”: Sometimes the entire moral community has the standing to sanction a wrongdoer. At other times only a subset of the wrongdoer’s fellows (e.g., only her family, or only her direct victims) may hold her accountable. In still other cases, the agent is accountable only to her own conscience. . . . A failure to mind one’s own business is something for which one can be held accountable. (2011, 575).

Radzik here thinks of blame, or least expressed blame, as involving a kind of sanction. And she argues that one’s relationship, such as whether one is a family member or a friend, to the wrongdoer can affect the moral propriety of expressing blame.38 Other examples of standing include relationships thought of more broadly—​the relationship of a co-​conspirator who appears to lack standing to blame because she is complicit in the very act for which her partner in crime is blameworthy (this points to a “complicity” condition); or, thinking even more broadly, the relationship of a hypocrite who appears to lack standing to blame because she stands in the relationship of having committed the same sort of offense (pointing to a “hypocrisy” condition”).39 In all of these cases, the relationship of the blamer to the blamee in some way affects the status of blame as permissible or appropriate, even if we stipulate in advance that the blamee is truly blameworthy and even deserving of blame. And though less has been written on the way that fine-​grained distinctions in relationships affect standing, it also seems intuitive that there is a difference in the degree and kind of blame that it is appropriate for closer and more casual friends to engage in, for example. The differences in relationships of these kinds would seem also to affect what is all things considered appropriate when it comes to blame in any given instance. At this point, it is crucial to note the distinction between overt and private blame in connection with these questions of standing. It is much easier to see that how one expresses blame is a matter of ethical consideration than how unexpressed blame is. The expression of blame is a kind of action, something we do, and so something that seems subject to moral obligations and norms. We ought not to express blame in gratuitously cruel ways, for example, just as we ought not to do other gratuitously cruel things. If, however, there is such a thing as unexpressed blame, which consists of purely mental states or dispositions, then it is more controversial that this sort of blame is the kind of thing that we ought to experience in certain ways rather than others. For, at least on some views, as we have seen, ought implies can, and obligations require a kind of control that we do not have over our mental states and dispositions. But I think that we can 37  For use of the label “business condition” and critical discussion, see for example, Macalester Bell (2013). 38  See also Angela Smith (2007) for influential examples in which it seems that people in different relationships are entitled to blame in different ways, and Coates and Tognazzini (2013) for further discussion. 39  For just some examples of discussions of this condition, see Wallace (2010) and Todd (2019).

Relationships and Responsibility    561 still see unexpressed blame as subject to moral assessment because we can sometimes have at least indirect control over our attitudes, and even where we cannot, we can assess our attitudes as good or bad to have (even if we cannot appropriately be held to account for them). So sometimes we can speak of whether blame is permissible, obligatory, or forbidden, and sometimes in more general terms about whether it is appropriate or inappropriate.40 Fittingness and standing are not the only considerations that have been thought to bear on the all things considered permissibility or appropriateness of blame. Other ethical considerations are also relevant, such as that a person has suffered enough, or been blamed enough, say. So on one common framework there are three kinds of considerations that are related to the all things considered question: fittingness, standing, and other moral considerations. Within this framework, it seems that relationships are typically brought to bear when it comes to standing. But this is not the only possible taxonomy of considerations that together determine an all things considered judgment about what sort of blame is appropriate in a given case. Consider that ethical considerations take different forms. For example, some are related to rights, some to the promotion of welfare, and others to fair or equal distributions. If we begin with these large ethical categories, then it is possible to see considerations that appeal to relationships as falling into any one of these. For example, some considerations that fall under the business condition might fall under potential rights violations—​especially when it comes to expressing one’s blame. It might be that one has a general right not to be harshly blamed in an overt way in general for harms that fall mostly on oneself, say, but that this right is waived in close relationships. Relationships like marriage, which are based in part on promises, clearly generate new and distinctive rights and duties on the part of spouses. So it is easy to see how being in such a relationship might change the normative landscape when it comes to correlative rights and duties regarding blame. Couples might agree explicitly or implicitly to express their blame when they feel it. Some have argued that a failure to blame can imply a violation of respect, and the duty to express respect might be particularly strong in the case of close relationships.41 Alternatively, people in close relationships might have special duties of welfare to each other, and this might in some instances call for less or more overt blame, depending on the situation. In a complementary way, it has been argued that third parties might have a variety of different kinds of duties to allow people within a relationship to deal with their partners’ wrongdoing in a kind of private sphere (within limits), as these close relationships are central to human flourishing and

40 

A full exploration of the intersection of the ethics of blame and particular relationships would require sorting through different notions of blame, and indexing moral principles to them. For example, if skepticism about desert is true, then arguably it is wrong to blame in a sense that presupposes desert in all circumstances. But it could still be permissible, and perhaps even obligatory in some circumstances, to blame in ways that do not presuppose desert. (See Pereboom 2013, 2014a, and 2017) for discussion of these different sorts of blame.) 41  See John Robison (2019) and Kathleen Connelly (in preparation), for example.

562   Dana Kay Nelkin can often require such a sphere for their integrity.42 Such duties plausibly include duties of welfare. In sum, questions about whether relationships put one in a certain position to blame might fall naturally into more than one larger ethical category. Whatever categorization scheme we use, however, the important lesson is that we should aim to identify the ethical considerations that explain the fact that the relationship of blamer to blamee can change the all things considered permissibility to blame in a particular way. As we saw, there are a variety of options for what might justify something like the business condition, and there is a lively debate surrounding whether any ultimately succeeds.43 For example, Smith (2007) suggests that considerations of privacy might underlie such a condition, while Bell (2013) argues that the typical cases at hand, such as a stranger overhearing one party telling another about having committed some wrong, are precisely cases in which a right to privacy has been waived by having been shared in public. But it might be that there are other notions of privacy in play, such as the notion of a personal sphere free of outside interference, such as that put forward by Radzik. Nevertheless, as Radzik notes, it is difficult to draw a principled line between situations in which third parties ought to keep their blame to themselves and when they ought not. And the debate is even more unsettled when it comes to private blame, where it is more difficult to explain what is morally problematic about the blame of a stranger, as long as they keep their blame to themselves. We can further pose more fine-​grained questions, as well. Do certain kinds of friendship entitle friends to blame in ways that are different from others? Does it matter what implicit expectations the participants have of each other? And so on. My suspicion is that relationships will enter into all things considered permissibility to blame judgments in many and subtle ways, and thus it will be more accurate to replace a monolithic “business condition” with a set of conditions that capture various ways in which relationships can make morally appropriate different kinds of blame. A full accounting of the ways in which relationships affect all things considered judgments about the permissibility or appropriateness of blame would require an entry of its own. But this is clearly a very rich area for further exploration, one that requires getting clear on what exactly blame is and how relationships intersect with general ethical principles. While we saw in the last section that serious questions can be raised about whether relationships must enter into the conditions of blameworthiness, it is much harder to doubt that the ethics of at least some kind of blame must depend on the

42  Radzik (2011): 593. Note that Radzik points out that such a private sphere will have limits and if the wrongdoing is serious enough, outside intervention in the form of blame (and presumably also in the form of protective action) could be called for. She notes that it is a subtle and difficult matter to determine what these limits are. 43  For recent discussion, see MacalesterBell (2013) who rejects the Business Condition and is skeptical of the notion of “standing,” but nevertheless recognizes what she calls the “positionality” of blame, and McKiernan (2016) who defends the Business Condition, largely on the epistemic grounds that people in close relationships are in a better position to know relevant details about certain kinds of wrongdoing.

Relationships and Responsibility    563 nature of the relationship between the blamer and blamee. That there are subtle ways in which it can do so points to a rich research program already in progress.

5.  Conclusion Strawson’s exhortation to keep in mind personal relationships when exploring debates about freedom and all that is at stake, including moral responsibility, has certainly been heeded. Each of the three debates discussed here has in some way taken up the thought that responsible agency and surrounding practices of holding responsible, blaming, and praising are inextricably linked. Important work continues in each one, and new work goes beyond what even Strawson may have envisioned.44

References Anglin, W. S. (1990). Free Will and the Christian Faith. Oxford: Oxford University Press. Arneson, Richard (2003). “Consequentialism vs. Special-​Ties Partiality,” Monist 86: 382–​401. Bell, Macalester (2013). “The Standing to Blame: A Critique,” in Blame: Its Nature and Norms, edited by D. Justin Coates and Neal Tognazzini. Oxford: Oxford University Press, 263-281. Bingeman, Emily (In preparation). “In Praise of Blame.” Brink, David O. (1994). “Moral Conflict and Its Structure,” Philosophical Review 103: 215–​247. Brink, David (2004). “Immaturity, Normative Competence, and Juvenile Transfer: How (Not) to Punish Minors for Major Crimes,” Texas Law Review 82: 1555–​8155. Brown, Eric (2017). “Blame, Strangers, and the Moral Relationship,” Analysis 77: 10–​20. Chaplin, Rosalind (2019). “Taking It Personally: Third-​Party Forgiveness, Close Relationships, and the Standing to Forgive,” Oxford Studies in Normative Ethics 9: 73–​94. Coates, D. Justin (2013). “In Defense of Love Internalism,” Journal of Ethics 17: 233–​255. Coates, D. Justin (2018). “Hard Incompatibilism and the Participant Attitude,” Canadian Journal of Philosophy 49: 208–​229. Coates, D. Justin and Tognazzini, Neal, eds. (2013). “The Contours of Blame,” in Blame: Its Nature and Norms, edited by D. J. Coates and N. A. Tognazzini. Oxford: Oxford University Press, 3-26. Connelly, Kathleen Ambrose (In preparation). “Blame and Respect.” Darwall, Stephen (2006). The Second-​ Person Standpoint. Cambridge, MA: Harvard University Press. Ekstrom, Laura (2000). Free Will: A Philosophical Study. Boulder, CO: Westview Press.

44  See, for example, new work on the social dimensions of responsible agency and subtle discussion about whether various social factors, such as oppression in the family and in larger social groups, undermine responsible agency. For one example, see the essays in Hutchison, Mackenzie, and Oshana (2018).

564   Dana Kay Nelkin Elzein, Nadine and Pernu, Tuomas K. (2019). “To Be Able To or To Be Able Not To? That Is the Question. A Problem for the Transcendental Argument for Free Will,” European Journal of Philosophy 15: 13–​31. Fischer, John Martin (2002/​2003). “Frankfurt-​Style Compatibilism,” in Contours of Agency: Essays on Themes from Harry Frankfurt, edited by Sarah Buss and Lee Overton. Cambridge, MA: MIT Press, 1-26. Haji, Ishtiyaque (1998). “Moral Anchors and Control,” Canadian Journal of Philosophy 29: 175–​204. Haji, Ishtiyaque (2002). Deontic Morality and Control. Cambridge: Cambridge University Press. Haji, Ishtiyaque (2012). Reason’s Debt to Freedom: Normative Appraisals, Reasons, and Free Will. Oxford and New York: Oxford University Press. Howard-​Snyder, Frances (2006). “‘Cannot’ Implies ‘Ought Not,’ ” Philosophical Studies 130: 233–​246. Hutchison, Katrina, Mackenzie, Catriona, and Oshana, Marina (2018). Social Dimensions of Moral Responsibility. Oxford: Oxford University Press. Kane, Robert (1996). The Significance of Free Will. Oxford: Oxford University Press. Kohn, Alfie (2005). Unconditional Parenting: Moving from Rewards and Punishments to Love and Reason. New York: Atria Books. Mason, Elinor (In preparation). “Taking the Blame.” Milam, Per-​Erik (2016). “Reactive Attitudes and Personal Relationships,” Canadian Journal of Philosophy 46: 102–​122. Milam, Per-​Erik (2022). “Forgiveness,” in The Oxford Handbook of Moral Responsibility, edited by Dana K. Nelkin and Derk Pereboom. Oxford: Oxford University Press. Nelkin, Dana Kay (2011a). Making Sense of Freedom and Responsibility. Oxford: Oxford University Press. Nelkin, Dana Kay (2011b). “T. M. Scanlon, Moral Dimensions: Permissibility, Meaning, and Blame”, Philosophical Review: 603–​607. Nelkin, Dana Kay (2014a). “Free Will Skepticism and Obligation Skepticism: Comments on Derk Pereboom’s Free Will, Agency, and Meaning in Life,” Science, Religion, and Culture 1(1). Nelkin, Dana Kay (2014b). “Moral Responsibility, the Reactive Attitudes, and the Significance of Free Will,” in Libertarian Free Will: Contemporary Debates, edited by David Palmer. New York: Oxford University Press, 142–​160. Nelkin, Dana Kay (2015). “Friendship, Freedom, and Special Obligations,” Agency, Freedom, and Moral Responsibility, edited by Andrei Buckareff, Carlos Moya, and Sergi Rosell. Houndsmill: Palgrave MacMillan, 226–​250. Norcross, Alastair (2006). “Reasons Without Demands: Rethinking Rightness,” in Blackwell Contemporary Debates in Moral Theory, edited by Jamie Dreier. Malden, MA: Blackwell, 161–​179. Pereboom, Derk (2001). Living Without Free Will. Cambridge: Cambridge University Press. Pereboom, Derk (2013). “Free Will Skepticism, Blame, and Obligation,” in Blame: Its Nature and Norms, edited by D. Justin Coates and Neal Tognazzini. Oxford: Oxford University Press, 189–​206. Pereboom, Derk (2014a). Free Will, Agency, and Meaning in Life. Oxford: Oxford University Press. Pereboom, Derk (2014b). “The Dialectic of Selfhood and the Significance of Free Will,” in Libertarian Free Will: Contemporary Debates, edited by David Palmer. New York: Oxford University Press, 161–​175.

Relationships and Responsibility    565 Pereboom, Derk (2017). “Responsibility, Regret, and Protest,” Oxford Studies in Agency and Responsibility 4: 121–​140. Pereboom, Derk (2018). “Love and Freedom,” in The Oxford Handbook of the Philosophy of Love, edited by Christopher Grau and Aaron Smuts. New York: Oxford University Press. Pereboom, Derk (2021) Wrongdoing and the Moral Emotions. Oxford: Oxford University Press. Piovarchy, Adam (2020). “Hypocrisy, Standing to Blame, and Second Personal Authority,” Pacific Philosophical Quarterly 101 (4): 603–​627. Radzik, Linda (2011). “On Minding Your Own Business: Differentiating Accountability Relations Within the Moral Community,” Social Theory and Practice 37: 574–​598. Rawls, John (1955). “Two Concepts of Rules,” The Philosophical Review 64: 3–​32. Robison, John W. (2019). “When and Why Is It Wrong to Excuse an Attitude?,” Philosophical Studies 176: 2391–​2409. Rudy-​Hiller, Fernando (2020). “Reasonable Expectations, Moral Responsibility, and Empirical Data,” Philosophical Studies 177: 2945–​2968. Scanlon, T. M. (2008). Moral Dimensions: Permissibility, Meaning, and Blame. Cambridge, MA: Belknap Press of Harvard University Press. Shabo, Seth (2012). “Where Love and Resentment Meet: Strawson’s Interpersonal Defense of Compatibilism,” The Philosophical Review 121: 1 95–​124. Sher, George (2013). “Wrongdoing and Relationships: The Problem of the Stranger,” in Blame: Its Nature and Norms, edited by D. Justin Coates and Neal Tognazzini, Oxford: Oxford University Press, 49-65. Shoemaker, David (2017). “Response-​Dependent Responsibility; or, A Funny Thing Happened on the Way to Blame,” The Philosophical Review 126: 481–​527. Shoemaker, David (2019). “Hurt Feelings,” Journal of Philosophy 156: 125–​148. Smith, Angela. (2007). “On Being Responsible and Holding Responsible,” Journal of Ethics 2: 465–​484. Sommers, Tamler (2007). “The Objective Attitude,” The Philosophical Quarterly 57: 321–​341. Strawson, Peter (1962). “Freedom and Resentment,” in Free Will, 2d edition. Oxford: Oxford University Press, 72–​93. Todd, Patrick (2019). “A Unified Account of the Moral Standing to Blame,” Noûs 53: 347–​374. Vargas, Manuel (2013). Building Better Beings. Oxford: Oxford University Press. Vranas, Peter (2007). “I Ought, Therefore I Can,” Philosophical Studies 136: 167–​216. Wallace, Jay (1994). Responsibility and the Moral Sentiments. Cambridge, MA: Harvard University Press). Wallace, R. Jay (2009). The Publicity of Reasons. Philosophical Perspectives 23: 471–​497. Wallace, Jay (2010). “Hypocrisy, Moral Address, and the Equal Standing of Persons,” Philosophy and Public Affairs 38: 307–​341. Watson, Gary (1987). “Responsibility and the Limits of Evil: Variations on a Strawsonian Theme,” in Responsibility, Character, and the Emotions: New Essays in Moral Psychology, edited by Ferdinand Schoeman. Cambridge: Cambridge University Press, 256–​286; reprinted in Gary Watson (2004). Agency and Answerability. Oxford: Oxford University Press, 219–​259. Wolf, Susan (1981). “The Importance of Free Will,” Mind 90: 386–​405. Wolf, Susan (2001). “The Moral of Moral Luck,” Philosophic Exchange 31: 4–​19. Wolf, Susan (2011). “Blame, Italian Style,” in Reasons and Recognition. Essays on the Philosophy of T. M. Scanlon, edited by R. J. Wallace, R. Kumar, and S. Freeman. New York: Oxford University Press, 332–​348.

chapter 27

Re sp onsibilit y, Pe rs ona l Rel ationships , a nd t h e Si gnificanc e of t h e Reactive At t i t u de s Seth Shabo

1.  Introduction P. F. Strawson’s “Freedom and Resentment,” which was originally published in 1962, initiated a profound shift in the literature on free will and moral responsibility. According to Strawson, previous theorists had over-​intellectualized the nature of moral responsibility. His remedy was to illuminate the emotional underpinnings of our moral responsibility practices. To this end, he drew attention to a familiar range of “reactive attitudes and feelings,” of which the most important for his purposes are resentment, moral indignation, and guilt. What sets these three apart, in Strawson’s view, is that they are all reactions to behavior that exhibits an objectionable attitude or “quality of will,” such as animosity, contempt, or neglect. Such behavior offends against our demand for due regard in how people treat one another, a demand that is rooted in our natural human concern with, and sensitivity to, the interpersonal attitudes of our fellow humans. Strawson believed, in effect, that these three reactions form the emotional core of our practices of moral blame and punishment, and that once we recognize this emotional core, the questions that had divided previous theorists—​questions about what justifies these practices, and about whether the truth of determinism would threaten their justification—​seem idle (55, 60). For, he held, our proneness to these reactions is inseparable from our commitment to ordinary interpersonal relationships, which, as “part of the general framework of human life” and “a condition of our humanity,” does not require an “external” justification (55; and Strawson 1985: 33).

Personal Relationships and the Reactive Attitudes    567 Most contemporary theorists believe that the reactive attitudes are a crucial piece of the puzzle about determinism and moral responsibility, and that much hinges on whether, contra Strawson, these attitudes presuppose the falsity of determinism. But while Strawson succeeded in making these attitudes central to the current debate, his most ambitious claims about them have not been widely accepted.1 This chapter takes up one such claim. According to Strawson, relinquishing the reactive attitudes—​and, by extension, our practices of moral blame and punishment—​is not a real possibility for us, even supposing that determinism’s truth provides theoretical grounds for doing so (Strawson 1993: 53–​55). Strawson’s basis for this claim, which in previous work I dubbed “the Inseparability Thesis,” is the alleged inseparability of our susceptibility to the reactive attitudes from our commitment to ordinary interpersonal personal relationships.2 For convenience, I will reserve the term “reactive attitude” for resentment, moral indignation, and guilt. As commentators have observed, Strawson sometimes uses this term to denote a larger set of attitudes and feelings, which includes resentment, gratitude, anger, forgiveness, hurt feelings, and reciprocal love between adults (48, 52), but not moral indignation and guilt. I will use Strawson’s term “participant attitudes” for this larger set. So understood, the Inseparability Thesis implies that our commitment to personal relationships precludes relinquishing our susceptibility to resentment, moral indignation, and guilt. The Inseparability Thesis rests, in part, on two highly plausible claims. One is that ordinary interpersonal relationships depend on our concern with co-​participants’ attitudes toward us, including the esteem in which with they hold us, as exhibited in the regard with which they treat us (48–​49). The second is that resentment, which is prompted by perceived violations of our demand for due regard, is one manifestation of this concern. While highly plausible, these claims fall short of establishing the Inseparability Thesis. For even if they are true, it is a further question whether our commitment to personal relationships precludes relinquishing the reactive attitudes. Why believe this? Perhaps we could prevent ourselves from experiencing these attitudes without compromising the underlying concern with how others treat us? Moreover, even if we could not relinquish these attitudes, given our commitment to personal relationships, what bearing does this have on contemporary discussions of determinism and moral responsibility? The answers to these questions are not obvious. As Jonathan Bennett remarks,

1  One such claim is that moral responsibility is a response-​dependent property, which should be understood fundamentally in terms of when the reactive attitudes are appropriate. For a defense of this view, see Shoemaker 2017. Another claim is that the reactive attitudes do not permit or require a general justification; see Magill 2008 for a defense. 2  In Shabo 2012a, I formulated this thesis in terms of what I called our “ordinary susceptibility” to resentment. However, I will suppress this qualification here, since it isn’t as important for the purposes of this chapter.

568   Seth Shabo Strawson emphasizes—​more than I would want to—​that we could not possibly relinquish all reactive feelings. Still, ought we to try? Ought we to strive in that direction? We do have a live question about what course we should steer.3

By itself, the conclusion that we could not eliminate instances of the reactive attitudes isn’t very interesting. For it might still turn out that these attitudes are morally and rationally untenable, and that we could—​and should—​work to curtail them, so that they no longer have a prominent place in human affairs. For Strawson’s position to be worth defending, I submit, we must have reason to believe that even trying to marginalize these attitudes would somehow be misguided, given our commitment to personal relationships. By “the Strawsonian View,” I shall understand the twofold claim that, first, marginalizing these attitudes is not a credible aim of social policy; and, second, that we can see the truth of this first claim by attending to relevant features of our social and psychological reality. While Strawson doesn’t endorse precisely this view, it is a natural extension of his claim that we couldn’t eliminate these attitudes, and the considerations he adduces in support of this claim—​including his grounds for the Inseparability Thesis—​also count in favor of the Strawsonian View. In what follows, I distinguish two possible defenses of the Strawsonian View, both of which are suggested in “Freedom and Resentment,” but which have not been expressly contrasted in the literature. The first holds that we could not overcome our susceptibility to resentment without relaxing or suspending the normative expectations that partly constitute personal relationships. The second holds that we could not overcome this susceptibility without extricating ourselves from a broader network of interpersonal attitudes that reinforce resentment. I will call these arguments “the Constitutive Expectations Argument” and “the Psychological Nexus Argument,” respectively. After some preliminary remarks about the reactive attitudes in the next section, I summarize the two arguments in section 3. In section 4, I set out the main points of contention between defenders and opponents of the Strawsonian View. I then examine the two arguments in sections 5 and 6. As will be argued, the Psychological Nexus Argument is the more promising of the two.

2.  Preliminary Remarks on the Reactive Attitudes Not all forms of resentment, indignation, and guilt are relevant to the Strawsonian View. A journalist may resent a professional rival whose work has gained prominence

3  Bennett 2008: 59. Bennett is using “reactive feelings” here to include not only resentment, moral indignation, and guilt, but also the favorable attitudes, such as gratitude, that are expressed in praising behavior. However, Bennett’s question remains equally relevant if we focus specifically on whether we should try to eliminate this trio. Cf. Wallace 1994: 38–​40.

Personal Relationships and the Reactive Attitudes    569 it doesn’t merit. A con artist may feel indignant when his intended mark scoffs at his claims, which, in his mind, should seem unimpeachable. An employee may feel guilty when her position is spared after many of her friends are laid off. What matters for the Strawsonian View is that an easily recognizable variety of each attitude is a response to behavior that displays a lack of due regard. I will have this variety in mind in what follows. Of the three, Strawson saw resentment as the most natural starting point for his purposes. This is because resentment is a way of taking another’s attitudes toward us personally, and because recalling episodes of resentment is an effective reminder of our personal stake in others’ attitudes and intentions toward us (Strawson, “Freedom and Resentment”: 48–​50). This personal stake is what explains, for example, the possibility of conflicts with a distinctly personal dimension, as opposed to ones that are based solely on divergent interests. Just as it isn’t credible that we could relinquish this personal stake in most of our interpersonal dealings, the idea is, so it isn’t credible that we could immunize ourselves against resentment without broadly relinquishing this personal stake. From this starting point, it can be argued that resentment undergirds moral indignation and guilt, in that we could not immunize ourselves against these other attitudes, which concern violations of others’ demands for due regard, while remaining prone to resent violations of our own demand, since doing so would require us to suppress the recognition that other people have the same personal stake that we do in how they are treated (57–​58). Commentators have asked what unites these three attitudes, or how they “hang together” as a class (see, for example, Wallace 1996 and 2008, and Russell 2013). What does resentment, which we experience principally on our own behalf, have in common with moral indignation, which Strawson describes as its “vicarious” or “generalized” counterpart (56)? What do both share with guilt? A closer look at these questions is beyond the scope of this chapter. For my purposes, a minimal characterization will suffice. Whatever else, these three attitudes are all reactions to perceived violations of our demand for due regard, and all are inappropriate when their target hasn’t actually violated this demand, or, alternatively, when their target isn’t subject to this demand because (or to the extent to which) she isn’t a normatively competent agent (50–​52). Defenders of the Constitutive Expectations Argument and the Normative Expectations Argument believe that our commitment to personal relationships presents a powerful bulwark against efforts to marginalize the reactive attitudes. Mature friendship and reciprocal love are prime examples of such relationships. I will take opponents of the Strawsonian View to deny that marginalizing the reactive attitudes would come at the expense of such relationships as we normally understand them. Following Strawson, I will focus on resentment when setting out the two arguments below, on the assumption that these arguments could be extended to moral indignation and guilt if they succeed there.4 4 

I suggest how such an extension might go in Shabo 2012a.

570   Seth Shabo

3.  Distinguishing the Two Arguments To begin, both the Constitutive Expectation Argument and the Psychological Nexus Argument differ from a third possible defense of the Strawsonian View that can be found in the literature. According to this third defense, we cannot eliminate (or seriously curtail) the reactive attitudes because our natural susceptibility to them is simply too great. By analogy, we might think of our natural susceptibility to fear, annoyance, amusement, or boredom. We can confidently say that our psychological makeup precludes our regularly forestalling these responses while the kinds of situations that normally elicit them remain common. And, the argument continues, the same is true with the reactive attitudes; as a species, our reactive “fuse” is too short for such an effort to gain traction. One major problem with this defense is that it doesn’t show why trying to marginalize the reactive attitudes would be misguided, as the Strawsonian View requires. As Paul Russell (1992) observes, even if we are inescapably prone to these attitudes, perhaps we could nevertheless become good at preventing this proneness from being manifested. If, for example, we became convinced that the reactive attitudes are always morally objectionable, perhaps we could reliably forestall or quash their occurrences.5 Without actually undertaking this effort, there is no way to know whether this argument’s defender has underestimated our prospects. A second major problem is that this argument doesn’t address the possibility that we could strive to marginalize the reactive attitudes in part by disavowing instances that we fail to suppress. For both of these reasons, this straightforwardly naturalistic defense of the Strawsonian View falls short. Unlike this defense, the Constitutive Expectations Argument and the Psychological Nexus Argument both seek to ground the Strawsonian View, not in our sheer susceptibility to the reactive attitudes, but in the inseparability of these attitudes from involvement in personal relationships. As the name suggests, the Constitutive Expectations Argument emphasizes the constitutive role of normative expectations in personal relationships. It can be summarized this way: Personal relationships are partly constituted by normative expectations concerning how coparticipants treat one another, expectations that stem from our basic demand for due regard; and holding someone to such an expectation is being disposed to resent violations (Strawson, “Freedom and Resentment”: 63; see also 54). Since our

5  It should be noted that Russell’s discussion of whether we could prevent ourselves from experiencing the reactive attitudes is in the service of his discussion of whether these attitudes could be universally unjustified. Russell’s point, in effect, is that even if our susceptibility to these attitudes is unavoidable and does not require justification, Strawson’s opponent can still maintain that each token of these attitudes is avoidable and potentially unjustified.

Personal Relationships and the Reactive Attitudes    571 continued involvement in such relationships ensures that this disposition will persist, we cannot seriously attempt to marginalize resentment.

While this argument is often cited in connection with Strawson’s essay, it hasn’t yet received an explicit formulation. I present a formulation in section 4. The second and less obvious strand in Strawson’s thinking focuses on how other interpersonal responses reinforce resentment. In previous work, I developed this strand, which I called Strawson’s “Argument from Personal Relationships,” by arguing that we could not credibly strive to marginalize the reactive attitudes without trying to curtail the broader range of participant attitudes on which personal relationships depend.6 The Psychological Nexus Argument further develops this core idea,7 which can be summarized like this: Personal relationships depend on our distinctively personal concern with how coparticipants treat one another. Resentment is one manifestation of this concern, and it is powerfully reinforced by others (Strawson, “Freedom and Resentment”: 48–​49), which makes resentment hard to relinquish or sincerely disavow. Given this reinforcing relationship, marginalizing resentment would require extricating ourselves from this broader nexus of attitudes and feelings, which would in turn require a more far-​reaching emotional divestment from other people than we can seriously entertain.

This concern with others’ attitudes toward us is distinctively personal in that it reflects the non-​instrumental importance we attach to these attitudes, including the esteem (or lack thereof) in which others hold us, especially in the context of personal relationships. I set out this argument in section 5. As these summaries suggest, the two arguments are complementary. A defender of the Constitutive Expectations Argument might accept, for example, that we hold co-​ participants in personal relationships to the normative expectations we do at least in part because we care in a distinctively personal way about how they treat us, and that some of the various manifestations of this concern are mutually reinforcing, as the Psychological Nexus Argument implies. And a defender of the Psychological Nexus Argument will likely believe, in keeping with the Constitutive Expectations Argument, that such normative expectations are partly constitutive of personal relationships. As we shall see, however, the two arguments point to different answers to the question, “What would stop us from marginalizing the reactive attitudes if we seemed to have

6  The Psychological Nexus Argument is a further development of what I called “the Inseparability Argument” in Shabo 2012a. 7  To anticipate, the Psychological Nexus Argument elaborates more on why we (allegedly) couldn’t adopt a policy of sincerely disavowing the resentment we are unable to suppress.

572   Seth Shabo good reason to do so?” I shall argue that only the Psychological Nexus Argument has the resources for a satisfying answer to this question.8 Before proceeding, a clarification is needed. My sketch of the Constitutive Expectations Argument featured the familiar notion of holding people to normative expectations concerning how they treat us. When we speak of holding someone to such a demand or expectation, we usually have one of two things in mind. First, we may be referring to an actual response to a particular violation of the expectation, as when one person takes another to task on some occasion. Second, we may be describing a standing disposition to respond in such a manner in the event that the expectation is violated. I shall refer to these two stances as “holding someone to an expectation in the specific sense” and “holding someone to an expectation in the general sense,” respectively. In light of this distinction, we can identify two further commonalities between the two arguments before seeing where they part ways. First, defenders of both arguments believe that someone who resents a particular violation of her normative expectations, including her basic demand for due regard, thereby holds the other party to the violated expectation in the specific sense. Second, both believe that someone who is suitably disposed to resent violations of her normative expectations thereby holds the other party to those expectations in the general sense. Where the two arguments part ways is in how they link holding someone to an expectation in the general sense with being disposed to resent violations of that expectation. A defender of the Constitutive Expectations Argument believes that adopting this stance—​ or, at least, adopting it in the manner required for personal relationships—​entails having this disposition. By contrast, a defender of the Psychological Nexus Argument isn’t committed to this entailment. Perhaps she will agree that someone who lacks this disposition doesn’t count as holding the other to a normative expectation (or as doing so in the manner required for personal relationships), or perhaps she will instead see this disposition as merely characteristic of how we actually hold people to normative expectations. The important point is that, in contrast to the Constitutive Expectations Argument, nothing in the Psychological Nexus Argument turns on this entailment. Following Gary Watson (2014), I will call our basic demand for due regard in how people treat us “the Basic Demand,” and our basic concern with others’ attitudes toward us “the Basic Concern.” From the summaries mentioned, we can see that the Basic Demand and the Basic Concern feature in both arguments. As will be seen more fully, however, the emphasis is different. Whereas the Constitutive Expectations Argument gives pride of place to the Basic Demand in its answer to the question of why we couldn’t marginalize the reactive attitudes, the Basic Concern (or an aspect of it) is the load-​bearing component in the Psychological Nexus Argument. In the next section, I describe in greater detail what a policy of marginalizing the reactive attitudes would look like. I then examine the two arguments’ respective resources for showing that we could not seriously pursue such a policy. 8 

For an argument that incorporates elements of both, see Coates 2013.

Personal Relationships and the Reactive Attitudes    573

4.  The Containment Policy If the Strawsonian View is correct, marginalizing the reactive attitudes appears feasible only when we abstract away from our emotional investment in other people and their attitudes toward us. This view’s main philosophical opponents don’t deny that this emotional investment constrains what is possible for us. In their view, we could greatly reduce the role of these attitudes in human affairs without affecting the core elements of this investment, the ones that allow for relationships of mature friendship and reciprocal love (see especially Pereboom 2001: ch. 7; and 2014: ch. 7; Sommers 2007; Goldman 2014; and Milam 2016. See also G. Strawson 1986). Such an undertaking is incumbent on us, they believe, because the reactive attitudes falsely presuppose that people deserve to be the targets of these responses. We don’t genuinely deserve such responses if determinism is true, they hold, and the prospects for such desert are hardly better if determinism is false.9 I have characterized the Strawsonian View in terms of “marginalizing” the reactive attitudes. How far in this direction must we be able to go for this view’s opponents to be vindicated? In previous work, I proposed an answer that should be acceptable to both sides. Opponents of the Strawsonian View, I suggested, are committed to seeing the following as a live possibility for us: The Containment Policy . . . [W]‌e should strive as far as possible to take the reactive attitudes out of circulation. In the first place, this will involve trying to forestall these attitudes in the situations that normally elicit them. When a wrong or a slight occurs, we should remind ourselves that the agent isn’t genuinely blameworthy for it, and that it would be unfair to harbor resentment or indignation towards her, let alone to express these attitudes in a blaming way. When this reminder isn’t enough, we should continue to focus on the moral inappropriateness of our reaction so as to put it behind us as quickly as possible. When these reactions persist anyway, we should consciously disavow them . . . Finally, there may be times when people cannot be expected to sincerely disavow such feelings. In these cases . . . [when they express their resentment in blaming behavior] we won’t accord their reactions any social legitimacy, treating them instead in the variety of ways we treat other inappropriate but understandable emotional responses. (Shabo 2012a: 108)

A moral community like ours will have succeeded in marginalizing the reactive attitudes, as I put it, when its members have by and large adopted the Containment Policy. And its members will have by and large adopted this policy, in turn, when it’s generally expected that they will forestall or quickly relinquish feelings of resentment

9 

The kind of desert at issue is often described as “basic desert,” since it does not derive from contractualist or consequentialist considerations.

574   Seth Shabo and moral indignation whenever possible, sincerely disavow such feelings when they cannot be suppressed, and deny social legitimacy to blaming behavior. Underlying the disagreement about whether we could credibly pursue the Containment Policy is the shared assumption that we could not seriously strive to curtail or eliminate the typical causes of resentment—​slights, wrongs, and other manifestations of objectionable qualities of will. At issue is whether we could credibly pursue the Containment Policy without significantly curtailing such behavior. Advocates of the Containment Policy believe that we could do so, while defenders of the Strawsonian View hold that marginalizing resentment without curtailing its typical causes is hardly more feasible than trying to curtail these causes themselves.

4.1.  Disavowing the Reactive Attitudes While I continue to believe that the Containment Policy provides the right framework for assessing the Strawsonian View, I have since come to think that much hinges on how we construe the notion of disavowing the relevant attitudes, and on what is involved in doing so sincerely. Responding to my earlier argument, Pereboom contends that disavowing resentment is indeed a viable option for us, one that “can be achieved without compromising the personal nature of the care one has for the quality of the other’s will” (Pereboom 2014: 185). If the requirements for sincerely disavowing these attitudes are relatively modest, Pereboom and other advocates of the Containment Policy can plausibly deny that there is any insuperable obstacle to adopting a policy of sincerely disavowing recalcitrant instances. If, alternatively, the requirements are more demanding, this position may be harder to defend. To repeat, I will focus on resentment here. Talk of disavowing one’s resentment is fairly common in the literature, yet the question of what is involved in doing so has been neglected. When someone comes to disavow her lingering resentment, how does her outlook toward this attitude thereby change? For present purposes, what matters is not precisely how ordinary speakers use the expression “disavowing one’s resentment.” The important question is whether we can identify a stance that broadly aligns with ordinary usage, where regularly adopting this stance could anchor the successful adoption of the Containment Policy, so that resentment ceased to have any sanctioned or indispensable role in how we respond to wrongdoing. Whatever else, disavowing one’s resentment in the relevant sense will involve undertaking a commitment not to exhibit this resentment in blaming behavior. By itself, however, such a commitment isn’t enough, for we can easily imagine people who, while positively seething with resentment they regard as eminently justified, have overriding pragmatic reasons to conceal it. To support the Containment Policy, such a commitment must be undertaken as a reflection of one’s attitude toward one’s resentment. What sort of attitude will suffice?

Personal Relationships and the Reactive Attitudes    575 So as to keep the discussion manageable, let us consider two proposals: Hard Disavowal To disavow one’s resentment in the relevant sense, it is necessary and sufficient to convincingly deny that one’s resentment is indicative of one’s real or considered feelings toward the object of one’s resentment. Soft Disavowal To disavow one’s resentment in the relevant sense, it is necessary and sufficient either (i) to convincingly deny that one’s resentment is indicative of one’s real or considered feelings toward the object of one’s resentment, or, failing that, (ii) to maintain that one is genuinely ambivalent or unresolved about whether these feelings are so indicative.

Both proposals refer to one’s “real or considered feelings.” Underlying both is the thought that we can take a reflective standpoint toward our own emotional responses, a standpoint from which we may either “own” or identify with, or else “disown” or dissociate from, these responses (cf. Frankfurt 1971). Someone who disowns his emotional response doesn’t deny that it’s “really his” in the sense that he is the subject of the relevant mental states. Rather, he sees himself as subject to them “in spite of himself.” Typically, what prompts someone to take this reflective standpoint is a question about whether her resentment is an appropriate or reasonable response in the circumstances. If the subject concludes that her resentment is somehow inappropriate or misplaced—​ because she is overreacting, unfairly taking out her frustrations on others, taking things too personally, letting the pressures of the moment get to her, and so on—​she may tell herself that her resentment isn’t indicative of her true feelings, even if she can’t simply wish it away. If, conversely, she concludes on further reflection that she isn’t overreacting (say, because the other’s behavior isn’t an isolated irritant, but is part of a pattern of controlling behavior), she may well be inclined to accept her feelings of resentment as indicative of her true feelings on the matter. What distinguishes Hard Disavowal from Soft Disavowal is that the former requires a more definitive dissociation from one’s resentment: It requires the subject to deny tout court that his resentment is indicative of his true or considered feelings. In doing so, he will be affirming that, having taken stock of his true or considered feelings—​of where he stands with regard to his emotional states—​his lingering resentment isn’t part of them. By contrast, a defender of Soft Disavowal maintains that such a definitive dissociation, while sufficient, isn’t necessary; it’s enough if the subject can convincingly affirm to himself that his attitude toward his resentment is complicated in the following way. While he can’t simply deny that it’s indicative of his true feelings, it would also be inaccurate to say simply that it is indicative. On the question of where he stands, he is unresolved or divided (“of two minds”), with an inclination to accept these feelings as really his and a conflicting inclination to deny this. For convenience, I will say that someone who disavows his resentment in this “softer” sense tells himself that he is Unresolved. If Soft Disavowal is right, marginalizing the reactive attitudes is a credible policy objective provided that we can regularly and sincerely

576   Seth Shabo affirm that we are Unresolved with regard to persistent resentment. In that case, advocates of the Containment Policy need only maintain that we can indeed generally reasonably expect such “soft” disavowals of one another. Before proceeding, one further commonality between Hard Disavowal and Soft Disavowal bears emphasizing. On both, disavowing one’s resentment is distinct from judging that resentment is inappropriate or misplaced (which is normally what prompts a disavowal), as well as from refraining from endorsing one’s resentment on the basis of this judgment. For reasons we will touch on in section 4.2, a policy of not endorsing one’s resentment on this basis is not enough for the purposes of the Containment Policy—​not if we cannot be expected to sincerely disavow the resentment from which we withhold our endorsement. While I’m inclined to accept Hard Disavowal, I won’t defend it here. Instead, I’ll grant Soft Disavowal for the sake of argument, so that sincerely disavowing resentment in the weaker sense is enough to comply with the Containment Policy. I’ll argue that adopting a policy of sincerely disavowing our resentment, even in the weaker sense, isn’t a credible objective. What is to stop us from adopting such a policy? I submit that even “soft” disavowals will often ring false, just as “hard” disavowals will; for, as I will argue, soft disavowals may ring false even when we are inclined to believe that our feelings of resentment are misplaced or inappropriate and wish that we could sincerely disavow them and move on. Moreover, discovering that a disavowal rings false may put us in better touch with our emotional reality. Seeing our resentment as indicative of our true feelings after all, the feelings we must simply and unreservedly “own” on pain of emotional dishonesty, we will have discovered that how we really feel doesn’t conform to how we believe we are supposed to feel. To illustrate, consider someone who has rationalized an illustrious coworker’s rudeness toward him in the interests of preserving social harmony. He might believe that it was incumbent on him to avoid getting in the Illustrious One’s way, and that the latter’s caustic rebuke was only to be expected in light of his failure to do so. Seeing his lingering feelings of resentment as misplaced, he might initially disavow them in the stronger sense, denying tout court that they are indicative of his true or considered feelings, only to discover that the disavowal rings false. When the disavowal strikes him as forced and hollow, he will have realized that his resentment cannot be neatly separated out from how he really feels. If he next tells himself that he is Unresolved, it won’t be surprising if this “softer” disavowal rings false as well. For it may be evident to him that he isn’t being candid with himself, and that he must simply accept that these feelings are indicative of his true feelings and “go from there,” even if he doesn’t approve of this state of affairs and believes that being Unresolved would be morally preferable. Now suppose that our agent tells himself that he is Unresolved, not because he has rationalized the coworker’s behavior, but because he wishes to comply with the Containment Policy. We can still imagine this softer disavowal ringing false. That is, we can imagine that, despite believing that he should at least be Unresolved, maintaining that he is Unresolved prompts the realization that he isn’t being candid with himself, and that emotional honesty requires unreservedly accepting that his resentment is indicative of his true feelings. I will now consider how we might account for this assessment.

Personal Relationships and the Reactive Attitudes    577

4.2.  Putatively Appropriate versus Putatively Inappropriate Resentment To start, it will help to contrast putatively appropriate with putatively inappropriate resentment. By “putatively inappropriate resentment,” I mean resentment that strikes us as inappropriate or misplaced by our everyday lights, before we introduce philosophical concerns about determinism and desert. When we believe, for example, that a subject of resentment is being petty, unreasonable, bad-​tempered, self-​righteous, self-​ indulgent, self-​centered, immature, or hypocritical, or that his resentment is ginned up, self-​serving, envy-​driven, agenda-​driven, or due to a need for emotional drama, these considerations tend to count against the appropriateness of his resentment, quite apart from philosophical arguments that resentment is never an appropriate response to wrongdoing. By “putatively appropriate resentment,” I mean resentment that strikes us as emotionally apt or fitting in pre-​philosophical terms. Putatively appropriate resentment normally lacks the inappropriate-​making features described earlier. In addition, the clearest instances of such resentment tend to have the following four features. First, they seem emotionally fitting (or “make good sense”) to us (and to disinterested others) in light of an accurate understanding of the relevant facts. Second, their perceived fittingness survives mature critical reflection. Third, it survives such reflection at least in part because no other response seems to “do justice” to how we feel about how we have been treated, in that resentment alone seems to register a forceful enough rejection of the offending behavior. Fourth, even in retrospect, when it’s time to move on and relinquish resentment, we don’t revise our assessment that it was an emotionally fitting response at the time. The line separating putatively appropriate from putatively inappropriate resentment isn’t sharp, not least because putative appropriateness and inappropriateness come in degrees, because an instance of resentment might be putatively appropriate in some respects and not in others, and because there may be room for reasonable disagreement about how to weigh these different respects in particular cases. What matters for my purposes is only that we can identify clear examples of putatively appropriate resentment. Let us focus on cases where it seems to the subject that no other response does justice to how he feels about how he has been treated, because only resentment seems to register a forceful enough rejection of that treatment. I submit that such cases, if not uncommon, will present an obstacle to the Containment Policy. This is because people who on mature critical reflection regard their putatively appropriate resentment in this way will tend to have a strong sense that their resentment is an emotionally fitting response. And, the thought continues, when someone has this strong sense about her putatively appropriate resentment for this reason, it will be hard for others to hold that she can reasonably be expected to sincerely disavow it. Finally, others who understand her situation in this light will find it hard not to accord some degree of social legitimacy to her expression of resentment if she does express it. At the least, I’m inclined to think,

578   Seth Shabo they will find it hard not to respond differently than they would if they saw her resentment as putatively inappropriate.10 Why might someone believe that only resentment does justice to how she feels about how she has been treated, or that resentment alone registers a suitably forceful rejection of that treatment? I suspect that there is no single answer to this question, but the following two explanations may be illustrative. First, someone may believe that a less emotionally forceful rejection of the offending treatment, such as firm disapproval without resentment, risks downplaying the offense or minimizing the personal significance she attaches to it. Second, failing such an emotionally forceful rejection, someone may be unable to convince himself that he hasn’t partly acquiesced to the perspective of the blamed party, a perspective from which there is no compelling reason not to treat him (or others in similar circumstances) in this manner. More figuratively, we might say that only such a forceful rejection draws a sufficiently bright line between his perspective and the other party’s, so that he can be confident that he hasn’t partially internalized the view of his moral and social status that the blamed party’s treatment of him seems to reflect.11 If this thought is plausible, it won’t be a stretch to think that there may be cases where only expressing one’s putatively appropriate resentment seems like a forceful enough rejection, and that others, who see one’s expression in this light, will be moved to show that they see it as morally unlike expressions of putatively inappropriate resentment. Summing up, I have outlined a two-​tiered obstacle to pursuing the Containment Policy. First, there will be cases where people cannot reasonably be expected to sincerely disavow their resentment, since no other response does justice to how they (reasonably) feel about how they have been treated, and since even “soft” disavowals will ring false in these cases. If it were realistic to think that people would reliably deny social legitimacy to expressions of resentment in such cases, as the Containment Policy requires, this policy might still be viable. But, turning to the second tier, it seems unrealistic to expect people not to accord some measure of social legitimacy to many such expressions, if only by acknowledging that there is a morally significant difference between such expressions and expressions of putatively inappropriate resentment. We may now add that this expectation seems 10  Fleshing this thought would require distinguishing putatively appropriate from putatively inappropriate expressions of putatively appropriate resentment. I do not wish to deny that some expressions of resentment (even putatively appropriate resentment) might be so inappropriate in the circumstances that we would treat them no differently from expressions of putatively inappropriate resentment, e.g., as when the display of resentment is so “over the top” that it overshadows the offense that prompted it. 11  Starting with Watson 1987, there has been much interest in the idea that the reactive attitudes are “incipiently forms of moral address,” in that they naturally prompt complaints or protests against objectionable treatment. While I’m sympathetic to this idea, I believe it’s worth stressing that resentment and its expressions have a self-​directed role as well, in that they may provide the only convincing evidence to the subject that he hasn’t acquiesced to the offending treatment. In a related vein, see Boxill 1976: 68. See also Talbert 2012 and Smith 2013 for discussions of moral responsibility, reactive attitudes, and protest. For a conversational theory of moral responsibility, see McKenna 2012. For a discussion of the idea that someone’s conduct can convey a threat to one’s moral standing, see Hieronymi 2001.

Personal Relationships and the Reactive Attitudes    579 unrealistic even if we suppose that, in deference to philosophical objections to resentment, people often refrain from endorsing the resentment they cannot sincerely disavow. If they recognize that they (and others in similar circumstances) must unreservedly own this resentment on pain of emotional dishonesty, merely withholding their endorsement is apt to seem like a mere intellectual exercise. The crux of this response to advocates of the Containment Policy is the claim that even “soft” disavowals will often ring false in light of how people (reasonably) feel about how they have been treated. If this claim is correct, and if it’s grounded in a suitably stable feature of our psychology, identifying this feature will be a major step toward vindicating the Strawsonian View. In the next section, after providing a formulation of the Constitutive Expectations Argument, I raise doubts about whether this argument can plausibly identify such a feature. In section 6, I argue that the Psychological Nexus Argument fares better.

5.  The Constitutive Expectations Argument Both the Constitutive Expectations Argument and the Psychological Nexus Argument build on Strawson’s insight that resentment is a response to violations of the Basic Demand. Being naturally attuned to others’ attitudes toward us, we are apt to experience resentment when we encounter hostility, callousness, rudeness, disloyalty, duplicity, chauvinism, ingratitude, and so on. In the summary in section 2, we saw three of the Constitutive Expectations Argument’s key premises. First, our commitment to personal relationships is nonnegotiable. Second, such relationships are partly constituted by holding co-​participants to the Basic Demand in the general sense. Third, holding them to this demand is being suitably disposed to resent violations. Of the three, the third premise is the most contentious. For even if being disposed to resent violations is one way of holding someone to the Basic Demand, there appear to be other ways. After all, we can hold people to this demand in the specific sense by voicing criticism of a violation, by refusing to return to “business as usual” until the offense has been addressed, and by expressing such nonreactive attitudes as shock, disappointment, and sadness (Pereboom 2001: ch. 7, and 2014: ch. 7). Since someone who is firmly disposed to respond to violations in these ways plausibly counts as holding someone to this demand in the general sense, regardless of whether she is disposed to resent violations, this premise seems too strong. It might be replied that these other responses, while sufficient to hold people to the Basic Demand in the specific sense on occasion, are not sufficiently forceful by themselves to sustain the stance of holding someone to this demand in the general sense, since, the idea is, we don’t count as holding people to a demand in the general sense unless we are prepared to enforce that demand by resenting violations. Put another way,

580   Seth Shabo part of what gives some normative expectations the force of demands is that violating them carries a credible risk of incurring resentment. If this is right, holding people to the Basic Demand does require a disposition to resent violations after all. Against this reply, it seems plausible to suppose that there are more and less forceful ways of holding someone to the Basic Demand in the general sense, and that someone who regularly manifests the alternative disposition described earlier still qualifies as holding someone to this demand, even if she never resents violations. So far as I can tell, the only way around this difficulty is for the argument’s defender to say that a suitable disposition to resent violations is essential, not to holding someone to the Basic Demand per se, but to doing so in the manner required for personal relationships. I will incorporate this qualification in the third premise in what follows. Besides supporting this premise, the defender’s other major task is to close the gap between the three premises stated before and the conclusion that we cannot reliably suppress (or sincerely disavow) instances of resentment in response to violations of the Basic Demand. After all, even if our commitment to personal relationships ensures that we remain disposed to resent violations, as these premises jointly imply, it doesn’t follow that we couldn’t reliably suppress manifestations of this disposition, let alone sincerely disavow them. To close this gap, at least one further premise is needed, one that effectively forecloses this possibility. Presumably, this premise will characterize the relevant disposition with a view to showing why we could not credibly strive to suppress or disavow its manifestations. However, this step won’t be plain sailing. For however precisely the argument’s defender characterizes this disposition, it will be open to the opponent to question why we couldn’t make a credible effort to suppress its manifestations. Suppose, for example, that the defender contends, in the fourth premise, that someone is suitably disposed to resent violations only if she is highly susceptible not only to resenting violations, but to experiencing resentment in a way that makes disavowals ring false, at least when her resentment is putatively appropriate. Even if the defender can illuminate this way of experiencing resentment, the opponent will be able to counter that, for all the defender has shown, we could nonetheless credibly strive to suppress its manifestations, including by working to gradually change how we experience resentment. Thus a gap between the argument’s premises and its conclusion will remain. A stronger version of this premise would require not only that we have this disposition but that we regularly manifest it. According to this version, someone has the required disposition only if she experiences putatively appropriate resentment in a way that normally—​perhaps barring the presence of special conditions that we don’t control—​ precludes sincerely disavowing it. This stronger premise—​together with the uncontroversial further claim that pursuing the Containment Policy would require that we try to shed this susceptibility and/​or prevent its manifestations—​bridges the gap with the argument’s conclusion. However, this stronger premise is even harder to support. For the defender must now say what it is about personal relationships, and about holding people to the Basic Demand in the manner allegedly required for them, that ensures that we will continue to experience resentment in the relevant way. Pending a compelling

Personal Relationships and the Reactive Attitudes    581 answer, the opponent will regard this premise as nothing more than an empirical hypothesis, and hence as insufficient to show that striving to adopt the Containment Policy would be misguided. Perhaps there is room to support this premise by appealing to the kind of mutual emotional investment of co-​participants in personal relationships. That is, perhaps the argument’s defender could provide specific grounds for linking the nature of this emotional investment with holding people to the Basic Demand in a way that frequently precludes suppressing or sincerely disavowing resentment. But it isn’t obvious how such an argument might go. Moreover, if the nature of this emotional investment ends up doing the real work, a defender of the Strawsonian View may be able to answer the question—​“Why couldn’t we pursue a policy of sincerely disavowing our putatively appropriate resentment?”—​directly in terms of the features of this emotional investment, without reference to the requirements for adequately holding co-​participants to the Basic Demand. In that case, the defender will have moved away from the Constitutive Expectations Argument’s distinctive emphasis on this demand. Now it should be noted that there is room for yet another defense of the Strawson View, one that begins not from resentment and personal relationships, but from moral indignation and social relations writ large. Unless we regularly respond to violations of the Basic Demand with moral indignation, the idea is, we could not adequately enforce the norms that allow us to live together as members of a moral community.12 A closer look at this alternative would take us far afield. I suspect that this argument might well serve to further underline the obstacles to successfully pursuing the Containment Policy, but it’s less clear how it might establish that striving to do so would be misguided, as the Strawsonian View requires. Summing up, the prospects for the Constitutive Expectations Argument depend on whether its defender can successfully complete two tasks. First, she must find the right way (in the argument’s third premise) to link holding someone to the Basic Demand in the general sense with the disposition to resent violations. Second, she must effectively bridge the remaining gap with the argument’s conclusion via a fourth premise to the effect that we couldn’t credibly pursue a policy of suppressing and disavowing putatively appropriate resentment. If the fourth premise requires only the presence of a disposition, it’s hard to see how it could close this gap. If, alternatively, it requires regular manifestations of this disposition when violations occur, it may (together with an uncontroversial fifth premise) close this gap, but supporting it won’t be easy. Because it isn’t clear how this challenge might be met, I’m skeptical about whether the Constitutive Expectations Argument is a promising strategy for defending the Strawsonian View. More schematically, the Constitutive Expectations Argument can be set out this way: 12  Although Coates (2013) draws on the requirements for personal relationships, his argument features the idea that we could not preserve the normative expectations unless we are firmly disposed to resent violations. I am indebted to an anonymous referee for prompting me to consider a defense of the Strawsonian View that starts from moral indignation instead of resentment.

582   Seth Shabo (CE1) The Containment Policy is not a real possibility for us if pursuing it is precluded by our commitment to personal relationships. (CE2) Personal relationships are partly constituted by the manner in which co-​ participants hold one another to the Basic Demand in the general sense. (CE3) We cannot hold a co-​participant to the Basic Demand in the required manner unless we are suitably disposed to resent violations. (CE4) We are not suitably disposed to resent violations unless we have a marked susceptibility to experience putatively appropriate resentment in a way that normally precludes sincerely disavowing it, where this susceptibility is regularly manifested when violations of the Basic Demand occur. (CE5) Pursuing the Containment Policy would require that we try to shed this susceptibility and/​or regularly prevent its manifestations. Therefore, (CE6) The Containment Policy is not a real possibility for us. The argument’s opponents won’t take issue with CE1, and CE2 isn’t specific enough to be controversial in its own right. As we have seen, CE3 is contentious. However, even if such a disposition is required for holding someone to the Basic Demand in the general sense, it will be hard to show that we lack a suitable disposition to resent violations unless we meet the conditions described in CE4. Why think that someone who regularly suppresses or disavows his putatively appropriate resentment toward another party would no longer be holding that party to the Basic Demand in the manner required for personal relationships? Pending support for this premise, the Constitutive Expectations Argument does not provide a compelling answer to the question, “What would prevent us from pursuing the Containment Policy if we seemed to have good reason to do so?” I will now argue that the Psychological Nexus Argument’s defender has more resources for answering it.

6.  The Psychological Nexus Argument Like the Constitutive Expectations Argument, the Psychological Nexus Argument builds on Strawson’s observation that resentment and other reactive attitudes are responses to violations of the Basic Demand, which are rooted in the Basic Concern. But whereas the Basic Demand is central to the Constitutive Expectations Argument, the Basic Concern looms larger in the Psychological Nexus Argument. As we saw in section 3, the Psychological Nexus Argument locates the major obstacle to adopting the Containment Policy in the relationship between our susceptibility to resentment, which is one manifestation of this concern, and other participant attitudes that reinforce resentment. The defender’s main task is to illuminate this reinforcing relationship in a way that shows the Containment Policy to be misguided.

Personal Relationships and the Reactive Attitudes    583 Two participant attitudes are especially well suited to this reinforcing role: feeling aggrieved and feeling hurt (Shabo 2012a and 2012b). Both involve a sense of being impinged on by someone else’s conduct. However, the characteristic affect is different in each case. Qua aggrieved, one tends to feel the “pinch” of an undue burden or imposition. It is often this “pinch”—​or, as we might say, something’s “rankling” or “rubbing us the wrong way” on reflection—​that alerts us to the undue burden or imposition. Qua hurt, one tends to feel the “sting” of the hurtful behavior. When the undue burden is hurtful in nature, it’s not uncommon to alternate between these distinct reactions. And, the argument’s defender maintains, when the behavior that impinges on us in these ways also elicits resentment, as it sometimes does, these other responses make it harder to dissociate ourselves from that resentment in the manner required by the Containment Policy. This is not least because these other responses, which are part of how we feel about how we have been treated, can make it the case that no other response does justice to how we feel. The strategy, then, will be to use these other personally engaged responses to give content to the notion of how someone feels about how he has been treated, so as to explain both why disavowing one’s putatively appropriate resentment will often ring false, and why people cannot reasonably be expected to quash or sincerely disavow their resentment in such cases. As we shall see more fully below, what suits these other responses to this reinforcing role is that, like resentment, they are manifestations of our disposition to take maltreatment personally. Before turning to these two responses, let us begin with a schematic statement of the argument. (PN1) The Containment Policy is not a real possibility for us if pursuing it is precluded by our commitment to personal relationships. (PN2) Personal relationships are partly constituted by the manner in which co-​participants care about the esteem in which they hold each other, as manifested in the regard they show one another. (PN3) We do not care about a co-​participant’s attitudes toward us in the required way unless we are suitably disposed to take personally treatment that reflects a lack of due regard. (PN4) We are not suitably disposed to take such treatment personally if we have broadly immunized ourselves against feeling hurt and feeling aggrieved, and we cannot reliably (sincerely) disavow resentment unless we have broadly immunized ourselves against these other responses. (PN5) Pursuing the Containment Policy would require reliably (sincerely) disavowing the resentment we cannot suppress. Therefore, (PN6) The Containment Policy is not a real possibility for us.

584   Seth Shabo The crux of this argument is that we could neither plan to sincerely disavow instances of resentment without a reliable way of suppressing other manifestations of our disposition to take maltreatment personally, including feeling hurt and feeling aggrieved, nor reliably suppress these other reactions without scaling back our concern with the esteem in which our co-​participants hold us, as manifested in their treatment of us. Since our commitment to personal relationships precludes immunizing ourselves against these other reactions by scaling back this concern, the Containment Policy isn’t a real possibility for us. Challenges to the Psychological Nexus Argument will likely focus on PN4. Taking a page from the critic’s response to the Constitutive Expectations Argument, why couldn’t we strive to suppress manifestations of our disposition to take maltreatment personally in these ways without losing the relevant disposition? The answer from the Psychological Nexus Argument’s defender is that the only reliable strategy for broadly immunizing ourselves against these responses is to be significantly less emotionally attuned than we normally are to others’ attitudes toward us. For when we are so attuned, our control over whether we take displays of disregard personally in these ways is at best indirect and very limited. And since we cannot control how attuned we are to others’ attitudes without limiting our emotional investment in how they see us, there is an irresolvable practical tension between fostering personal relationships and making it a priority to avoid taking maltreatment personally in these ways. Finally, without broadly immunizing ourselves against these reactions, we won’t reliably be in a position to sincerely disavow the resentment we are unable to suppress. Let us now consider these two reactions and how they reinforce resentment.

6.1.  Resentment and Feeling Aggrieved Someone who feels aggrieved by another’s conduct or its consequences feels “put out” or “put upon,” impinged on, pained, annoyed, offended, or frustrated, and her mind often returns to the cause of these feelings. Something about her situation “pinches,” and this affect focuses her mind on the feature of the situation that she takes to be onerous or off-​putting. Feeling aggrieved is a common precursor to resentment, in that the former often enough develops into resentment as the offense sinks in or as extenuating circumstances are ruled out. But it’s important to see that “resenting” and “feeling (or being) aggrieved” are not synonymous. After all, resentment is at least plausibly thought to presuppose desert, whereas feeling aggrieved involves no such presupposition. Moreover, it isn’t uncommon to feel aggrieved without going on to experience resentment, and some considerations that count against the appropriateness of resentment don’t count against the appropriateness of feeling aggrieved, as when someone doesn’t receive the consideration for which she might reasonably have hoped, but without being denied the consideration she is due. Distinguishing these responses is important for the Psychological Nexus Argument’s claim that feeling aggrieved tends to reinforce feelings of resentment when both are present, making the latter difficult to relinquish or disavow. Examining the characteristic focus of each will shed light on how this reinforcement works.

Personal Relationships and the Reactive Attitudes    585 What feeling aggrieved makes salient to the subject is an undue burden or imposition that results from another person’s conduct. And while the subject wouldn’t feel aggrieved (as opposed to merely frustrated or annoyed) if not for the awareness of the other person as the source of the burden, it’s the imposition the person has caused, rather than the person himself, that is the immediate focus of this response. With feelings of resentment, by contrast, the focus is on the person who has caused (or whose objectionable attitude constitutes) the imposition, as the imposition’s source. This focus on the other person helps to explain why feelings of resentment sometimes develop into personal grudges, while feeling aggrieved in the absence of resentment doesn’t. When resentment occurs as well, one mechanism by which feeling aggrieved reinforces feelings of resentment is straightforward. Starting with a focus on the undue imposition, the aggrieved party’s thoughts naturally turn to the person who caused it, and—​barring extenuating circumstances—​seeing the person in this light tends to reactivate her existing feelings of resentment. By assumption, the imposition that leads the subject to feel aggrieved in such cases is a violation of her demand for due consideration. It will often be difficult in such cases for the aggrieved party to avoid thinking about the imposition, to prevent her thoughts from turning to the source of the imposition, or to keep her feelings of resentment from being rekindled whenever this mental transition occurs. This frequent reactivation is a natural obstacle to quickly relinquishing resentment, even for someone who wishes to do so (say, because he believes that the situation is consuming too much mental energy). Along with making feelings of resentment hard to relinquish, feeling aggrieved tends to make such feelings hard to disavow. This is not least because someone who feels both aggrieved and resentful will often feel “hard done by” as well. Someone who feels hard done by is keenly aware of how she has been impinged on by another’s lack of due regard. Like feeling aggrieved and unlike resentment, feeling hard done by doesn’t plausibly presuppose that the other party deserves this response, since the main focus is on the treatment and its consequences and not on the person as the source of that treatment. For this reason, advocates of the Containment Policy cannot plausibly say that feeling hard done by is inherently morally or rationally problematic. Yet when someone who feels aggrieved and hard done by experiences putatively appropriate resentment, these other reactions can help to explain why even a “soft” disavowal rings false. For disavowals will typically ring false when someone believes that only resentment does justice to how she feels about how she has been treated, and feeling aggrieved and hard done by can greatly contribute to this belief.

6.2.  Resentment and Hurt Feelings Turning now to the second reinforcing response, let us start by noting some similarities and differences among resentment, feeling aggrieved, and hurt feelings. Like resentment and feeling aggrieved, hurt feelings occur only in response to the actions and attitudes of other people. If not for our distinctively human concern with how others treat us, we would

586   Seth Shabo have trouble making sense of these reactions. However, like resentment and unlike feeling aggrieved, hurt feelings occur only when we take those actions and attitudes personally, experiencing the behavior and attitudes that elicit them as directed toward us in particular, whether in the form of hostility or disregard. By contrast, while we often do take personally the conduct that causes us to feel aggrieved, and while feeling aggrieved is sometimes a way of taking such conduct personally, we can feel aggrieved by conduct we don’t take personally (as when, for example, we are deeply inconvenienced by a clumsily crafted policy). Next, unlike both resentment and feeling aggrieved, hurt feelings don’t depend on the perception that there is something amiss or objectionable in how we have been treated. For example, a sharply worded rebuke from an admired superior may hurt its recipient’s feelings, even if the recipient believes that the rebuke was justified in the circumstances. The target of a stingingly snide remark from a coworker may realize that he is just receiving a taste of his own medicine. A parent’s feelings may be hurt when her toddler exclaims that he doesn’t like her,13 despite not really holding the toddler to normative expectations. Or someone may be hurt upon realizing that a close friend has gradually drifted away, without believing that he can legitimately expect otherwise. Of course, sometimes—​and not coincidentally—​the causes of hurt feelings are also perceived violations of one’s demand for due regard, and sometimes what is painful is precisely the perception that such consideration has been withheld. Be this as it may, hurt feelings are not in the first place responses to perceived violations of the Basic Demand. Rather, they express our vulnerability to emotional harm, a vulnerability that partly undergirds the Basic Concern. What about the relationship between hurt feelings and resentment? Earlier I described feeling aggrieved as a common precursor to resentment, since someone who feels aggrieved is already in the vicinity of resentment, even if she doesn’t go the rest of the way. By contrast, hurt feelings as such are not a step in this direction. Only when the cause of the hurt feelings is also perceived to be a violation of one’s demand for due regard is resentment likely to ensue (though even then it may not). According to the Psychological Nexus Argument, our commitment to personal relationships prevents us from making a serious effort to marginalize resentment and other reactive attitudes, given that other responses, including feeling aggrieved and feeling hurt, tend to reinforce feelings of resentment when resentment is present. How does this reinforcement work with hurt feelings? To begin, it’s important to forestall some possible misunderstandings. The argument’s defender doesn’t claim that hurt feelings often or usually give rise to resentment, or that hurt feelings are usually among the causes of resentment when resentment does occur.14 There are many cases where one of these responses occurs without the other, and where the other would seem out of place. Moreover, even when both are present, they may stem from a common cause rather than the one’s being among the other’s causes. 13 

Milam (2016: 109) considers our reactions to small children, whom we are not apt to resent, while Pereboom (2014: ch. 7) focuses on our reactions to teenagers, where we often expect ourselves to disavow our resentment. 14  In his illuminating response to Shabo 2012a, I believe that Milam (Shabo 2012a) takes me to be making such a causal claim about hurt feelings and resentment, which isn’t what I had intended.

Personal Relationships and the Reactive Attitudes    587 What the argument’s defender does claim is twofold. First, it’s not unusual for the two responses to be present together, especially when the hurtful behavior offends against the subject’s demand for due regard. Second, when both are present, hurt feelings tend to reinforce feelings of resentment, making the latter harder to relinquish or disavow than they would otherwise be. For my purposes, then, the relevant cases are ones where both responses are present, and where the behavior that causes the hurt feelings violates the offended party’s demand for due consideration. Consider an example: Towards the start of her junior year of college, Nicole receives some jarring news from home. Her grandmother, who had generously offered to fund her college education, has announced that she is abruptly withdrawing her support. Since Nicole and her parents cannot afford the costs, she won’t be returning in the spring. She grasps that the announcement has nothing to do with her: it’s the latest escalation in a years-​long feud between her grandmother and her parents. Yet she is genuinely shocked that her grandmother, however controlling, would use her education as a pawn in the family power struggle.

We can further suppose that while Nicole realizes that she is not the real target of her grandmother’s ire, she nevertheless feels that she has been the target of unprovoked aggression, and that she is deeply hurt by her grandmother’s willingness to upend her life to send a message to her parents. Suppose that Nicole feels resentment as well, and that her sense of being a target of unprovoked aggression contributes to her feeling hurt, aggrieved, and resentful. According to the Psychological Nexus Argument’s defender, when hurt feelings and resentment are both present, the former tends to reinforce the latter in two ways. First, hurt feelings can make resentment hard to relinquish by keeping the subject’s mind focused on the offending behavior. Like the “pinch” of feeling aggrieved, the “sting” of hurt feelings makes it hard to think about the offending behavior in a dispassionate manner, let alone to stop thinking about it altogether. But while both reactions involve a sense of being impinged on, there is an important difference. As we saw earlier, qua aggrieved, one’s sense of being impinged on need not reflect any personal concern with, or vulnerability to, another’s attitudes toward one. By contrast, qua hurt, one’s sense of being impinged on is a direct expression of this concern and vulnerability. In Nicole’s case, part of the emotional harm she has experienced involves having been the target of hostility from someone in whom she was emotionally invested; and whenever she thinks of this harm, we can plausibly suppose, her thoughts turn to that person, which powerfully reactivates her feelings of resentment. In these circumstances, we can plausibly imagine, she won’t be relinquishing her resentment anytime soon. Second, hurt feelings can make resentment even harder to sincerely disavow than it would otherwise be. To see this, consider another agent’s response to a similar situation. Like Nicole, Inured Nicole feels aggrieved, and the “pinch” of the undue burden enlivens her feelings of resentment whenever she thinks about her grandmother’s conduct. Unlike Nicole, however, Inured Nicole isn’t caught entirely off guard; having recognized her grandmother’s instinctive need to control others, she has learned to keep

588   Seth Shabo a safe distance. For this reason, while she is deeply upset by her grandmother’s decision, she isn’t hurt by it; she doesn’t feel the “sting” that Nicole does. Now suppose that Nicole and Inured Nicole both believe that resentment is morally and rationally untenable, and that they should disavow the resentment they can’t help feeling. Doing so won’t be easy for either of them, given how aggrieved they both feel. However, the obstacle for Nicole will be much greater. If she tells herself that she is Unresolved (or “of two minds” regarding whether her resentment is indicative of her true feelings), we shouldn’t be surprised if the disavowal rings false. For she experiences the objectionable treatment as an emotional harm, to which her personal investment in her grandmother has left her vulnerable. And, we may suppose, this way of experiencing the offending behavior prevents her from dissociating herself from her current mindset, in which it seems clear that only resentment does justice to how she feels about how she has been treated. Assuming that Nicole’s resentment is putatively reasonable, it will be hard to maintain that she can reasonably be expected to sincerely disavow it. Advocates of the Containment Policy will likely counter that feeling hurt, shocked, and disappointed, as “nonreactive analogs” of resentment, could help to supplant the resentment we sometimes can’t help feeling in such cases (Pereboom 2001, ch. 7 and 2014, ch. 7). Nicole’s case points to a different dynamic: far from providing her with an alternative to resenting her grandmother, her sense of hurt and shock reinforces her resentment, making it harder to relinquish or sincerely disavow (cf. Shabo 2012a: 115). The point is not that we can generalize from Nicole’s case to all or most cases in which hurtful behavior elicits putatively appropriate resentment. There will be cases where the hurtful behavior doesn’t involve such a serious breach of the Basic Demand, where the breach is taken less personally, and where the recipient’s emotional stake in the situation affords her a different perspective (as with Inured Nicole). The point is, rather, that Nicole’s case helps to illustrate how hurt feelings, like feeling aggrieved, can reinforce feelings of resentment, making them harder to relinquish or sincerely disavow. By reference to these reinforcement mechanisms, we can see that the Psychological Nexus Argument has resources to answer the question, “What would prevent us from pursuing the Containment Policy if we seemed to have good reason to do so?” If we accept that such disavowals will often ring false because no other response does justice to how we (reasonably) feel about how we have been treated, which includes feeling hurt and/​or aggrieved, the Containment Policy will begin to appear out of touch with our emotional reality. At least, it will appear out of touch on the assumption that others will be hard pressed to deny some degree of social legitimacy to expressions of such resentment. To be sure, the Psychological Nexus Argument’s defender still has work to do. Yet I hope to have shown that this argument affords a promising response to those who see marginalizing the reactive attitudes as a credible aim of social policy.15 15  I presented an earlier version of the chapter at the Eastern Division meeting of the American Philosophical Association in 2018. I want to thank Charles Goodman for his commentary and audience members for helpful discussion. I also want to thank the editors of this volume and an anonymous referee for providing valuable feedback on earlier drafts.

Personal Relationships and the Reactive Attitudes    589

References Boxill, B. (1976). “Self-​Respect and Protest,” Philosophy and Public Affairs 6: 58–​69. Coates, D. J. (2013). “In Defense of Love Internalism,” Journal of Ethics 17:233–​255. Frankfurt, H. 1971. “Freedom of the Will and the Concept of a Person,” Journal of Philosophy 68: 5–​20. Goldman, D. (2014). “Modification of the Reactive Attitudes,” Pacific Philosophical Quarterly 95: 1–​22. Hieronymi, P. (2001). “Articulating an Uncompromising Forgiveness,” Philosophy and Phenomenological Research 62: 529–​555. Magill, K. (2008). “Blaming, Understanding, and Justification: A Defense of Strawson’s Naturalism about Moral Responsibility,” in M. McKenna and P. Russell (eds.), Free Will and Reactive Attitudes. Burlington, VT: Ashgate, 187–​200. Milam, P. (2016). “Reactive Attitudes and Personal Relationships,” Canadian Journal of Philosophy 46: 102–​122. McKenna, M. (2012). Conversation and Responsibility. New York: Oxford University Press. Pereboom, D. (2001). Living without Free Will. Cambridge: Cambridge University Press. Pereboom, D. (2014). Free Will, Agency, and Meaning in Life. Oxford: Oxford University Press. Russell, P. (1992). “Strawson’s Way of Naturalizing Responsibility,” Ethics 102: 287–​302. Russell, P. (2013). “Responsibility, Naturalism, and ‘the Morality System,’” in D. Shoemaker (ed.), Oxford Studies in Agency and Responsibility. New York: Oxford University Press, 184–​204. Shabo, S. (2012a). “Where Love and Resentment Meet: Strawson’s Intrapersonal Defense of Compatibilism,” Philosophical Review 121: 95–124. Shabo, S. (2012b). “Incompatibilism and Personal Relationships: Another Look at Strawson’s Objective Attitude,” Australasian Journal of Philosophy 90: 131–47. Shoemaker, D. (2017). “Response-​Dependent Responsibility; or, a Funny Thing Happened on the Way to Blame,” Philosophical Review 126: 481–​527. Smith, A. (2013). “Moral Blame and Moral Protest,” in D. J. Coates and N. Tognazzini (eds.), Blame: Its Nature and Norms. New York: Oxford University Press, 27–​48. Sommers, T. (2007). “The Objective Attitude,” Philosophical Quarterly 57:321–​341. Strawson, G. (1986). Freedom and Belief. Oxford: Oxford University Press. Strawson, P. F. (1985). Sceptism and Naturalism. London: Methuen. Strawson, P. F. (1993). “Freedom and Resentment,” in J. M. Fischer and M. Ravizza (eds.), Perspectives on Moral Responsibility. Ithaca: Cornell University Press, 45–​66. Originally published in 1962. Talbert, M. (2012). “Moral Blame, Moral Competence, and Protest,” Journal of Ethics 16: 89–​109. Wallace, R. J. (1994). Responsibility and the Moral Sentiments. Cambridge, MA: Harvard University Press. Wallace, R. J. (2008). “Emotions, Expectations and Responsibility,” in M. McKenna and P. Russell (eds.), Free Will and Reactive Attitudes. Burlington, VT: Ashgate, 157–​185. Watson, G. (1987). “Responsibility and the Limits of Evil: Variations on a Strawsonian Theme,” in Ferdinand Schoeman (ed.), Responsibility, Character, and the Emotions. Cambridge: Cambridge University Press, 256–​286. Watson, G. (2014). “Peter Strawson on Responsibility and Sociality,” in D. Shoemaker and N. Tognazzini (eds.), Oxford Studies on Agency and Responsibility, Volume 2 Oxford: Oxford University Press, 15–​32.

CHAPTER 28

Forgive ne s s Per-​E rik Milam

Forgiveness is a new dimension of moral responsibility theorizing. Unlike freedom and determinism, a philosopher might reasonably fail to consider how her theory of responsibility bears on her theory of forgiveness (and vice versa). The responsibility and forgiveness literatures have only recently begun to influence one another in a significant way—​a surprisingly late convergence given the influence on the responsibility literature of P. F. Strawson’s “Freedom and Resentment” (1962) and the influence on the forgiveness literature of Jeffrie Murphy’s “Forgiveness and Resentment” (1982), itself inspired by Strawson. My aim in this chapter is to explore the connection between forgiveness and responsibility. One could describe the landscape of the forgiveness debate in any number of ways: by subject (e.g. victim-​, self-​, third-​party, and institutional forgiveness), by discipline (e.g., philosophy, theology, social psychology, and psychiatry), by comparison with cognate phenomena (e.g., excuse, justification, mercy, and letting go), by its manifestation (e.g., private or communicated), or in some other way.1 My approach is to introduce a common model of forgiveness and some recent challenges to it, each of which highlights potential connections between forgiveness and responsibility. This chapter has four parts. Section 1 presents some examples of forgiveness and draws attention both to their common features and to the diversity of our experiences of forgiving. Section 2 introduces the Standard View of forgiveness and explains its popularity and plausibility. According to this view, forgiveness requires that the forgiver overcome her blame toward the offender in response to the offender’s subsequent attitudes and actions. Section 3 considers objections to the Standard View and shows how they support a competing conception of forgiveness. Section 4 explores the implications of different conceptions of forgiveness, in particular how each view understands the limits of forgiveness. 1 

See Hughes and Warmke (2017) for one such introduction.

Forgiveness   591

1.  Examples An adequate account of forgiveness will capture the phenomenon as we experience it. This should include not only paradigmatic cases, but also less familiar examples that are nonetheless recognizable as forgiveness. While disagreement about some cases is probably inevitable, we must accept that theorizing about the nature and ethics of forgiveness—​understanding what it is, what it does, and when to do it—​ requires working from examples. With that in mind, we can begin by considering three cases. Joke. One night at a party, Jesse overhears an acquaintance, Roy, tell what sounds like a racist joke about Latinos and is offended and angry with him. Roy notices Jesse’s reaction to his joke and recognizes that he’s acted badly. He is ashamed and, later in the night, he finds Jesse and apologizes to him for telling the hurtful joke. Jesse is relieved that Roy realized that the joke was inappropriate and happy that he cared enough to talk to him about it. He forgives him and, despite some awkwardness, they continue to get along well in the future. Test. Andrea and Kelly are taking their graduate school entrance exams. During the test Andrea notices that Kelly is using her smartphone to cheat. Andrea is upset about this—​the unfair advantage, the dishonesty, the implicit arrogance—​especially when she learns that, as a result of their scores, Kelly will be preferred for admission to a prestigious graduate program to which both have applied. Andrea does not confront Kelly, but complains to mutual friends about the incident. When they tell Kelly and she reflects on her behavior, Kelly, to her credit, realizes its significance and tries to make things right as best she can. Andrea is still annoyed, but she acknowledges Kelly’s remorse, and after a few weeks she decides to forgive. While they never talk about the test, both recognize that their conflict has been resolved. Infidelity. David cheats on his partner, Donna. At first he thinks it’s no big deal, but he begins to feel more and more ashamed of his behavior and guilty about betraying her trust. David recognizes what his remorse is telling him and he recommits himself, in his own mind, to being a faithful partner. Shortly thereafter he admits to Donna what he did, apologizes to her, and assures her that it won’t happen again, explaining how guilty and ashamed he feels and how much he values their relationship. Donna is understandably upset and, at first, does not know what to do. Eventually though she comes to believe that David is sincere in his remorse and apology, that he is trustworthy, and that they can still have a healthy and fulfilling relationship together. Donna overcomes her blame and tells David that she forgives him. These cases illustrate the structure of our forgiveness practice. There are a number of similarities, but also important differences between them. Some factors seem essential to forgiving (e.g., a perceived offense), while others are common but unnecessary (e.g., an apology). Some features are absent from these cases: unrepentant wrongdoers, inappropriate forgiving, non-​victim forgiveness—​though they could be written in.

592   Per-Erik Milam Nonetheless, while different accounts inevitably disagree about which features are essential and which are not, few would reject these cases as examples of forgiveness. In addition to capturing our diverse experiences of forgiving, a theory of forgiveness aims to give a coherent and psychologically plausible account that explains its role in our shared moral lives and is compatible with a plausible ethics of forgiving. Along with its intuitive fit, we assess the adequacy of any given theory according to these criteria.

2.  The Standard View It is natural to think of forgiveness as a change of attitude. This thought is supported by paradigm cases of forgiving and by a simple but compelling line of reasoning. It is not enough simply to speak the words, “I forgive you.” Such a statement—​or an equivalent gesture, performance, or other behavior—​is consistent with an entirely unforgiving attitude. One might misunderstand one’s own attitude toward the offender or say the words in order to manipulate others’ perceptions. Even if such mistakes and subterfuge are rare, the reality of such scenarios suggests that forgiveness requires an internal change. Call this the Standard View and its central claim—​that forgiveness requires that the forgiver overcome her blame toward the offender in response to the offender’s subsequent attitudes and actions—​the Change Condition. This is the position developed by Butler (1749/​1900), resurrected by Strawson (1962), refined by Murphy (1982), and defended in a variety of forms in much of the philosophical literature on forgiveness (Kolnai 1974; Hampton 1988; Calhoun 1992; Hieronymi 2001; Griswold 2007; Allais 2008; Garrard and McNaughton 2010; Pettigrove 2012).2 The ongoing debate about the nature of forgiveness can reasonably be characterized as a series of refinements, revisions, and rejections of this view. The Standard View is compelling, in part, because it describes a recognizable practice. It paints a familiar picture of what forgiveness is and how it works, the structure of which is distilled from examples like those presented earlier. But it is more than a familiar picture. The dominance of this position in the philosophical literature rests on the fact that the Change Condition hangs together with—​supports and is supported by—​ other claims that tie forgiveness into our broader responsibility practice. Together, they provide a coherent, unified, and psychologically plausible picture of the phenomenon. We can start by identifying the object of forgiveness. The forgiver forgives the offender for her offense—​that is, her blameworthy behavior or the character it reveals. This is the only possible object. In order to forgive, the victim must view the offender’s behavior or character as blameworthy. Call this the Blameworthiness Condition. This condition is motivated by the absurdity of denying its component claims. 2  Most scholars now agree that Butler’s view does not require that the forgiver cease blaming entirely (Newberry 2001; Garcia 2011). However, he can reasonably be considered the source of this view, even if his intellectual descendants misunderstood his position.

Forgiveness   593 First, there must be something to forgive. The forgiver must view the relevant misconduct as an offense (Downie 1965; Kolnai 1973; Horsbrugh 1974; Murphy 1982; Hieronymi 2001; Bash 2007; Zaibert 2009).3 One cannot forgive another for unobjectionable behavior; nor can one forgive behavior viewed as harmful but permissible. Suppose that Andrea is applying to the same university as Brandon and that, as a result, he is less likely to be accepted. Brandon is unhappy that Andrea is hurting his chances of admission, but he also believes that her decision is reasonable and permissible. As such, forgiveness is not an option for Brandon. Andrea is not a proper candidate for forgiveness any more than a justified lawbreaker is a proper candidate for a pardon (rather than an acquittal). The forgiver must also continue to view the behavior as an offense (Hieronymi 2001). If Andrea comes to believe that Kelly was permitted to use her phone during their test because she has a learning disability and requires extra help, then she cannot forgive her. Even if she ceases blaming her, what she does is justify her conduct, not forgive (Murphy 1988: 20). Moreover, forgiveness depends on the perceptions and values of the would-​be forgiver rather than the reality of the situation (Adams 1991). One cannot forgive a person one believes to be innocent, even if they are guilty (Novitz 1998); nor can one forgive an action one doesn’t take to be wrong. For example, Valerie cannot forgive Donna for having premarital sex if she believes that doing so is entirely appropriate, but Donna can forgive herself if she believes that she ought to have abstained. To see why this is so, consider the alternative. If forgiveness required actual wrongdoing, then agents with false moral beliefs would be unable to forgive perceived transgressions—​for example, Jesse’s grandmother could not have forgiven him for marrying someone from another religion. But such agents are doing something so nearly identical to what we call forgiveness that it seems implausible to exclude it from the category. Now, this requirement can be understood in different ways, so that forgiveness is a wider or narrower phenomenon. As with blame, the object of forgiveness is usually taken to be wrongdoing, whether actions or omissions, but one can arguably also forgive others’ ‘wrongbeing’—​for example, vices like inconsiderateness (Bell 2008; Pettigrove 2012). Others might defend an even wider category, arguing that one can forgive morally permissible but socially proscribed actions of various sorts—​for example, suberogatory actions (Driver 1992), violations of non-​obligatory basic decency (Calhoun 2004), or morally permissible moral mistakes (Harman 2016). For example, Brandon may forgive Brenda for refusing to donate a kidney to him even if he believes she wasn’t obligated to do it. Similarly, if telling offensive jokes is tasteless but permissible, Jesse might nonetheless forgive Roy for doing so. A still more inclusive conception might even allow for nonmoral forgiveness as the complement to nonmoral blame (Björnsson 2017, 143). If we can blame a reasoner for their sloppy inferences, perhaps we can also forgive such failures. 3  Espen Gamlund (2011) appears to deny this and argues that one can forgive both justified and excused conduct. However, even in his cases, one cannot forgive behavior one believes to be justified, only behavior for which one reasonably rejects a justification that is, in fact, adequate (2011: 112–​115).

594   Per-Erik Milam Second, the forgiver must view the offender as having been morally responsible for the offense (Kolnai 1973; Murphy 1982; Calhoun 1992; Hieronymi 2001; Murphy 2003; Griswold 2007; Allais 2008; Warmke 2015 and 2016a).4 One cannot forgive an offense that one believes resulted from non-​culpable ignorance, inability, or lack of opportunity. For example, Valerie should visit her friends, but she is not a candidate for forgiveness if she didn’t know they were in town, was too sick to travel, or couldn’t afford the trip. And, as with offense, forgiveness requires perceived rather than actual responsibility. If Brandon and Brenda believe that Valerie chooses to go skiing each year rather than visit them, then they can forgive her despite the fact that she wasn’t actually able to come. Their decision to forgive is based on a misunderstanding, but it’s forgiveness nonetheless. Finally, as with wrongdoing, we sometimes realize that a purported instance of culpable wrongdoing was not actually the offender’s fault. Such realizations usually prompt a change of attitude toward the offender—​for example, we stop blaming them—​but relinquishing blame for this reason is to excuse, not forgive (Murphy 1988: 20). There is nearly unanimous agreement that, in order to forgive, one must continue to view the offender as having been responsible for their offense. However, if responsibility skepticism is true and no one is responsible for any of their behavior (Pereboom 2001; Rosen 2004; Levy 2011), then this condition would seem to make (non-​deluded) forgiveness impossible. One can respond to this possibility in two ways. The first is to accept the conclusion as a radical implication of a radical theory. The second is to argue that forgiveness doesn’t require the kind of responsibility that is threatened by the skeptical arguments (Pereboom 2001). The Blameworthiness Condition is plausible, but not sufficient. According to the Standard View, forgiving is not just forming an attitude about the offense or offender; it is a change of attitude. Thus, an adequate account of forgiveness must identify what it is a change from. If forgiveness is a positive change of attitude, and we normally have positive attitudes toward others, then it must be preceded by a negative change of attitude about the same object, namely, the offense. The Standard View holds that the relevant negative attitude is blame and that the relevant change is overcoming blame. This is what we see when we reflect on our experiences and attempt to distill an account of forgiveness from real and imagined cases, like those discussed so far. In order to forgive, a victim must first have blamed the offender for her offense. Call this the Blame Condition. It is not enough to view the offender as having culpably done wrong. One must hold the wrongdoing against the offender (Nelkin 2017). For example, suppose Brenda is a vegan and believes that it is blameworthy to eat eggs and dairy products, but that it is easy and nice to eat cheese while vacationing in Paris. If, during her trip, she indulges in custards, croissants, and cheese, never caring that she’s acting contrary to her moral commitment, it seems impossible for her to forgive herself. Only if, back in London, 4  This condition is widely accepted in the empirical literature on forgiveness, too (Karremans et al. 2003), though empirical studies often fail to operationalize the distinction successfully (Struthers et al. 2008; Fehr et al. 2010; Pronk et al. 2010).

Forgiveness   595 she reflects on these luxurious days and regrets her decisions does it seem possible for her to forgive herself. The same can be said of Andrea’s response to Kelly’s cheating and Jesse’s reaction to Roy’s joke. The best explanation of our judgments in such cases is that blame—​understood here as holding the offense against the offender—​is a precondition on forgiving. The Blame Condition also allows for different formulations. Different accounts have identified the necessary response to offense as resentment (Murphy 1982; Hieronymi 2001; Walker 2006; Griswold 2007), as any retributive attitude (Kolnai 1973), or as any negative reactive attitude or hard feeling generally (Richards 1988; Hampton 1988; Murphy 1998; Allais 2008; Pettigrove 2012; Bell 2013; Blustein 2014). However, while they require different initial responses, all seem to require blame of some sort. Moreover, in addition to disagreeing about the form blame must take—​for example, resentment or any hard feeling—​proponents of the Standard View can also disagree about what blame is, such that forgiveness might be understood as overcoming a feeling (Murphy 1982), judging the situation differently (Hieronymi 2001), changing how one view’s the offender (Hampton 1988; Allais 2008), or simply ceasing to hold the offense against the offender (Warmke 2011). For example, a more restrictive view might hold that forgiveness is a personal practice, rather than one concerned with broader impersonal morality, and that one can only forgive actions that one takes personally. Thus, even if Brandon blames the politicians who propose to slash the foreign aid budget, he cannot forgive them unless he takes their indifference personally. However, a less restrictive view might deny this and suggest instead that it is no less plausible that one can forgive than that one can blame in such a case. However, while these disagreements about how precisely to formulate the Blame Condition suggest a wide range of conceptions, all seem to accept that in order to forgive one must hold the offense against the offender. Whether we think of this as blaming or as a particular kind of blaming is a matter of refining the view. The Standard View, as I have described it, gives a compelling account of the nature and function of forgiveness—​what it is and what it does. The picture of forgiveness as overcoming blame at apparent culpable wrongdoing is conceptually coherent, psychologically plausible, and true to our experience of what forgiving and being forgiven is like. It also explains common sense distinctions that we deploy in our moral dealings with one another. For example, it distinguishes cognate phenomena like excuse and justification that resemble forgiveness in their nature, function, or consequences by explaining that one can cease to blame an offender for different reasons. One may judge an action not to be wrongful or an offense not to be culpable.5 The consequences may be

5 

The Standard View is also consistent with distinguishing forgiving from letting go, whereby one overcomes blame but for reasons that seem inconsistent with forgiving—​e.g., in order to achieve cardiovascular benefits. Many philosophers recognize the distinction between forgiving and other ways of overcoming blame toward a culpable wrongdoer. See Milam (2018a) for an account that distinguishes the forgiving and letting go in terms of one’s reasons for overcoming blame.

596   Per-Erik Milam the same in each case (overcoming blame) and the different practices may have the same function (resolving moral conflicts), but they are nonetheless distinct. It also allows us to explain the role of forgiveness in our broader responsibility practice. Forgiveness is understood by reference to blame. Blame and forgiveness are conceived of as complementary practices, with forgiveness as a way of overcoming blame. It is part of the repertoire of practices—​ including blame, apology, and reconciliation—​by which we work through the moral conflicts that inevitably arise between individuals. We can conceive of our responsibility practice in different ways, but a useful model is Michael McKenna’s account of holding responsible as a conversation (2012: 89):

•​

Moral contribution: X commits an offense O against Y (e.g., Roy tells a racist joke that hurts and offends Jesse) • ​ Moral address: Y responds to X about O (e.g., Jesse blames Roy for telling the joke) •​ Moral account: X gives an account of the offense (e.g., Roy apologizes to Jesse) •​ Moral response: Y responds to X’s account (e.g., Jesse forgives Roy) The practice may go smoothly or not. In a best-​case scenario, it proceeds as above. However, sometimes Jesse’s blame may make Roy defensive or Jesse may reject Roy’s apology as insincere. In such cases reconciliation may be partial or not occur at all. Whichever account of our responsibility practice we prefer, the conversation model shows how the Standard View understands the function of forgiveness as one element of a larger practice. On this view, a forgiver is engaging in one part of a practice of holding responsible that includes (or is partially constituted by) blame, apology, forgiveness, and other responses to the attitudes and actions of seemingly responsible agents. Finally, the Standard View is consistent with a range of claims about the ethics of forgiveness. Typically, it does not settle moral claims in virtue of the nature or function of forgiveness. It leaves open whether and when forgiveness is good, right, or virtuous, whether the would-​be forgiver’s standing can render forgiveness inappropriate (Pettigrove 2009), and whether it is by its nature elective (Allais 2013; Milam 2018b). This wide compatibility—​leaving the ethics of forgiveness to normative ethics—​is another virtue of the Standard View.

3.  Revision and Rejection I wouldn’t call this the Standard View if there weren’t opponents. In this section, I introduce some objections and explain their motivation. These objections suggest that the Standard View is too exclusive and target either the Blame Condition or the Change Condition.

Forgiveness   597

3.1  Forgiveness Without Blame One might reject the Standard View as an overly narrow account of the nature of forgiveness. Contrary to the Blame Condition, one could argue that, while the forgiver must have a positive change of attitude, they need never blame in the first place, so the required change may be something other than overcoming blame. David might admit to Donna that he was unfaithful, apologize, and show genuine remorse. In response Donna might immediately forgive him without having ever blamed him.6 The same is true if one can forgive preemptively. For example, suppose Susan has injured herself while skiing and that she and Brandon need help if she is going to get down the mountain. Brandon can still ski, but Susan doesn’t want him to leave her alone. Nonetheless, she might say to him, “I forgive you if you decide to leave in order to find help.” If we accept this as genuine forgiveness, Susan preemptively forgives Brandon and does so without ever blaming him (Cornell 2017). In both cases, the forgiver has a change of heart but does not cease to blame. Instead, David’s apology and Susan’s acceptance of Brandon’s likely decision inhibit Donna and Susan’s respective dispositions to blame.7 The defender of the Standard View must either accept that Donna and Susan cannot forgive in this scenario or argue that, in those cases where they seem to forgive, they actually do blame in some sense.

3.2  Forgiveness Without Ceasing to Blame Alternatively, one might contend that the Change Condition excludes recognizable cases of forgiveness. In particular, one might think that, if forgiveness is voluntary and elective, then one can forgive even if one hasn’t yet overcome one’s blame. Whether one rejects the Blame Condition or the Change Condition, these objections must be accompanied by an alternative account of what changes when we forgive. In this section, I consider four suggestions, each of which views the connection between forgiveness and responsibility in a different way. First, one could accept that blame is a precondition on forgiving, but deny that the forgiver must overcome blame. This might seem like a minor refinement of the Standard View and in some versions of this view it is (Newberry 2001). It does not depart much 6 

Thanks to Karl Martin Adam for this suggestion. This argument applies equally well to affective, cognitive, and conative accounts of the nature of blame (Coates and Tognazzini 2012). Whether blame entails an emotional response, a change of judgment, or a change in how one behaves or is disposed to behave toward the offender, the claim is that one need not blame in order to forgive. If this objection is correct, Donna and Susan need not blame in any of these senses. However, if blame merely entails judging that the offender is ”to blame” for the offense—​i.e., that they are blameworthy (Vargas 2013)—​then blame is necessary for forgiveness because to cease to judge blameworthy would be to justify or excuse the offender (Hieronymi 2001; Murphy 2003). Forgiving cannot be a change in this attitude, i.e., the judgement of blameworthiness (at the time of action). 7 

598   Per-Erik Milam from the spirit of the view to require that one significantly, but not entirely, overcome one’s blame, and the two views do seem to shade into one another. Nonetheless, some versions of this position differ significantly from the Standard View. For example, Ernesto Garcia interprets Butler as arguing that to forgive is, fundamentally, to resent to the appropriate degree given the fact that the offender remains a moral agent who deserves the same basic regard and goodwill as any other human being (2011; 9).8 On this view, one forgives by coming to view the offender as morally redeemable despite their offense. One must cease to blame only insofar as blaming is incompatible with recognizing the offender’s capacity to reform. This is a much thinner and more inclusive notion of forgiveness than the Standard View. It also allows us to make sense of forgiveness as, by its nature, a virtuous mean between excess and deficiency in a way that proponents of the Standard View have struggled to make plausible.9 Alternatively, one might argue that forgiving requires only that one cease to endorse one’s blaming emotions (Schönherr 2018). Second, one could reject the Change Condition based on how one understands the nature of blameworthiness. It’s often thought that an offender remains blameworthy forever, even if blaming them ceases to be justified. But if one can cease to be blameworthy for a past culpable wrongdoing (Khoury and Matheson: 2018), then we can understand forgiveness as ceasing to view an offender as presently blameworthy, even if one continues to view them as having been blameworthy at the time of the offense, as seems necessary. For example, one could argue that: in order for an offender to be blameworthy for their action, it must be the result of a bad (or substandard) quality of will; a repentant offender does not have a bad quality of will; therefore, the offender is no longer blameworthy, though it may still be true that she was blameworthy.10 We often conceive of forgiving as separating the offender from their offense (Murphy 1988; Hieronymi 2001; Allais 2008). According to this view, separating the two supports ceasing to view the offender as blameworthy, rather than merely ceasing to blame. Third, one could argue that it is sufficient for forgiveness that the victim changes her mind about personally punishing the offender for their offense, even if she still blames them for it. On this view, in order to forgive, the victim must come to believe that the world would be a worse place if she tried to do something to offset the offense (e.g., take retribution) and, on these grounds, must deliberately refuse to do so (Zaibert 2009: 387). The change required for forgiveness, then, is a change in belief and the decision to act (or refrain) on the basis of that belief.11

8 

Jean Hampton’s discussion of moral hatred and biblical forgiveness gives a similar view (1988: 37). See, among others, Richards (1988), McGary (1989), and Griswold (2007). 10  Technically, on this view, the change required for forgiveness is a change in one’s perception of the offender’s present blameworthiness. However, insofar as blaming requires judging blameworthy, one must, as a matter of fact, also cease to blame. Thus, this view can capture the intuition that the forgiver must overcome their blame, but also allow that this is not the relevant change. 11  For responses to this view, see Warmke (2011) and Russell (2016). 9 

Forgiveness   599 Finally, one could argue that it is sufficient for forgiving that one choose to release an offender from the obligation generated by their offense, whether one continues blaming them or not (Nelkin 2013). On this view, blameworthy behavior generates an obligation that the offender has to the victim and to forgive is to release the offender from this obligation. For example, one may be obliged to repent, show remorse, or make amends in some way. Forgiveness also changes the normative significance of the offense from one which warrants various negative responses to a state where at least some of these responses are no longer justified (Warmke 2016a). Among other things, this view captures examples of forgiveness that resemble debt forgiveness or waiving a promise and that do not involve overcoming blame (Twambley 1976; Adams 1991; Nelkin 2013; Warmke 2016b; Cornell 2017; Bennett 2018) and it extends both to easy forgiveness of minor offenses (Adams 1991) as well as to the immediate or unconditional forgiveness of horrific wrongs often reported in the media (Walters 2006; Berman 2015).12 This last option may be the most promising and radical departure from the Standard View.13 Call it the Norm Changing View. It aims to explain the role and importance of forgiveness in our shared moral lives (Warmke 2016a; 687–​688). However, in order to constitute a legitimate alternative to the Standard View, more must be said about what exactly changes when one forgives. In the remainder of this section, I consider some possible explanations. Altering the norms between victim and offender is not itself sufficient for forgiveness. After all, these norms can be altered by the offender or by relevant changes in the world. For example, if Dylan crashes his father’s car, but suffers serious injury such that he is no longer a responsible agent, then he will not have obligations from which his father can release him. It seems, then, that the change of norms must be brought about by a change in the forgiver. However, it is not enough to intend to release the offender from their obligation because one might still fail to carry through or abandon the intention. One option is that the forgiver’s decision to forgive makes the difference (Zaibert 2009: 387). However, this would seem to be in tension with two desirable features of the Norm Changing View, namely, its ability to explain the change of status between forgiver and offender and the

12  It is unclear whether, on this view, forgiveness requires that the offender actually be blameworthy or whether it is enough that the forgiver perceive her as being so. Perceived wrongdoing is sufficient on the Standard View, but if forgiving releases the offender from an acquired obligation, that would seem to imply that she is actually guilty of culpable wrongdoing. 13  A particularly radical version of this view might argue that forgiveness is neither a change of attitude nor even a response to the offender or their offense, but rather a persisting attitude expressed in the requisite speech act or other performance. On this view, being wronged and forgiving function like being promised and waiving that promise. One can only wave a promise if one has been promised something, but it can be waived at will and for any reason (or no reason). Thus, for example, one might have a persisting desire not to be owed anything beyond what is owed to everyone and, for this reason, immediately and unconditionally waive all promises and forgive all wrongs as soon as one acquires the power to do so—​or, if possible, do so preemptively.

600   Per-Erik Milam similarity between interpersonal forgiveness and exercises of official authority like debt forgiveness or pardon. Altering the norms between forgiver and offender requires communication of the relevant decision, so a personal decision to forgive is not sufficient. A person is not free to leave prison as soon as the governor decides to pardon them, nor is a person released from their obligation to pay back a loan as soon as the creditor decides to forgive it. Communication of the decision is sometimes required to alter permissions and requirements.14 This suggests, then, that communicating the decision, rather than the decision itself, is what alters the relevant norms, whether via a speech act (“I forgive you”) or some equivalent gesture or behavior. There is disagreement about the nature of the communicative dimension of forgiveness. It has been characterized variously as a behabitive (Haber 1991), a commissive (Pettigrove 2012), and a declarative (Warmke 2016a).15 This is the change relevant to forgiveness. However, even communication might be insufficient if forgiveness also requires uptake by the offender in order to alter norms. It is not clear that a forgiver releases an offender from his obligation if the offender fails to recognize or accept this decision. In this way, forgiveness may operate like consent or refusal. Donna fails to refuse her stylist’s offer to shampoo her hair if the stylist interprets Donna’s “That’s okay” as “Okay, go ahead” rather than “Thank you, but no.” Similarly, Dylan fails to consent to a medical procedure if he is not recognized as consenting because he fails to meet the legal requirements for doing so.16 The same may be true of forgiving if, for example, an offender fails to recognize or refuses to accept forgiveness. The alternatives to the Standard View canvassed in this section, especially the Norm Changing View, offer a significantly different conception of the nature, function, and ethics of forgiveness. They dispute the preconditions for forgiveness (e.g., whether blame is necessary), what must remain fixed when one forgives (e.g., whether one must continue to view the person as responsible), and what must change when one forgives (e.g., overcoming blame, forswearing punishment, altering norms between victim and offender). At the same time, these views attempt to capture the same experiences of forgiving and the same distinctions between forgiveness and cognate phenomena.17 For example, forgiving is different from excusing or justifying, in part, because the former requires a decision, while the latter do not. The fact that an offender was excused or justified alters what the victim can permissibly say or do to the offender, but the victim does not possess this power to alter. 14  Warmke recognizes the importance of communication, but does not argue that communication is necessary for forgiveness (2016a: 699). 15  These accounts of the nature of communicative forgiveness are themselves neutral between the Standard View and Norm Changing View. One could argue that forgiving requires a speech act, but that the speech act must communicate that one no longer blames the person (Pettigrove 2012). 16  This is different from consenting but being refused, as when a person consents to give blood, but is refused because they weigh too little. Brunning and Milam defend an uptake condition (2018: 155–​157). 17  Warmke claims only to be concerned with paradigmatic forgiveness (2016a: 691 and 698), but he presumably wants his complete view to be true of the broader practice.

Forgiveness   601 The challengers to the Standard View also give a different account of the function of forgiveness. It is still understood as a way to resolve moral conflicts, but its primary significance is as a way of influencing norms of permission and obligation rather than as a step in our practice of holding responsible. For proponents of the Norm Changing View, wrongdoing and forgiveness function like promising and waiving promises—​ forgiveness dispels obligations just as promises create them. However, for proponents of the Standard View, forgiveness complements blame. One implication of this is that the plausibility of the Standard View of forgiveness is tied to the plausibility of its complementary account of blame. Meanwhile, the plausibility of the Norm Changing View is tied to the plausibility of a normative ethics that recognizes forgiving as a normative power. This difference in function has implications for the ethics of forgiveness on these views. The Standard View leaves the permissibility and impermissibility of forgiving to normative ethics, which is itself independent of the nature and function of forgiveness. Whether and when forgiveness is permissible or impermissible will depend on one’s account of moral permissibility—​for example, one may hold that forgiving is permissible if and only if doing so will produce the greatest happiness. However, the Norm Changing View takes a theory of forgiveness to be part of a normative ethical theory. If forgiving makes demanding restitution impermissible, then any normative theory that denies this must be mistaken. This view has significant implications. For example, if forgiveness functions like other normative powers, then one may be unable to reverse the normative changes one makes by forgiving. The ethics of forgiveness will also depend, for each of these views, on the degree to which it takes forgiving to be voluntary. We have only partial control over our beliefs, attitudes, and actions, so any view that requires a change of attitude—​whether abjuring punishment, overcoming blame, or accepting an offender’s remorse—​must allow that forgiving often proceeds indirectly and that some are, at times, unable to forgive. By contrast, views according to which forgiving is a decision that one can make independently of one’s beliefs and attitudes may suggest that forgiving is often under our direct control. Given the widespread acceptance of some version of the ought-​implies-​can principle, how voluntary one takes forgiveness to be will influence one’s account of whether and when one ought or ought not to forgive. I have presented these alternatives as challenges to the Standard View, but, before moving on, we should consider whether the Norm Changing View might supplement rather than supplant the Standard View. On the one hand, the Standard View offers necessary conditions on forgiveness, while its opponents reject these conditions and argue that other changes are sufficient. And at least some versions of the Norm Changing View reject the conception of forgiveness as an emotional change (Warmke 2016a and 2016b; Bennett 2018). On the other hand, however, an account very like the Standard View, but which holds that its conditions are jointly sufficient but not necessary for forgiveness, may be more plausible insofar as it can be broadened to capture a wider variety of forgiveness practices.

602   Per-Erik Milam

3.3  Pluralism Instead of rejecting the Standard View in favor of an alternative account, or vice versa, it might seem reasonable to advocate pluralism about forgiveness—​that is, to claim that there are multiple legitimate conceptions of forgiveness, including (perhaps) the Standard View and one (or more) of its challengers. However, justified pluralism requires more than a few irreconcilable theories. For it to be plausible (at least) the following conditions must be met: (1) each conception of the phenomenon is inadequate or incomplete on its own; (2) the competing conceptions are not parts of a single adequate account; and (3) the pluralist account is more plausible than the best monist account according to at least one of the criteria by which accounts are assessed. For example, a pluralist account of forgiveness would be plausible if, (1) theory X captured one class of examples and theory Y captured another class (e.g., private and communicated forgiveness); (2) the two theories are inconsistent (e.g., one restricts forgiveness to victims, while the other does not); and (3) neither of the theories is able to account for the plausibility of the other (e.g., if preemptive forgiveness cannot be understood merely as granting permission). While many philosophers accept that we deploy multiple seemingly irreconcilable conceptions of forgiveness (Hampton 1988; Adams 1991; Calhoun 1992; Garcia 2011; Nelkin 2011; Pettigrove 2012), Christopher Bennett (2003) is among the few who have developed an explicitly pluralist account. Instead, forgiveness theorists tend to reject one or more of the above conditions for pluralism. For example, in order to explain how forgiveness is elective, Cheshire Calhoun distinguishes aspirational from minimal forgiveness, but argues that the latter is actually a form of excuse (1992: 81). Miranda Fricker (forthcoming), also in an attempt to resolve the tension between elective and non-​elective forgiveness, identifies two seemingly inconsistent conceptions but argues that one is, in fact, a derivative form of the other. Ernesto Garcia (2011) identifies two interpretations of Butler’s account of forgiveness, but argues that one is not only a better exegesis but also a more plausible account of the phenomenon. Indeed, the most commonly drawn distinction between “kinds” of forgiveness—​ that is, between internal (or private) and external (or communicated) forgiveness—​is rarely presented as a genuinely pluralist theory. Rather they are taken to be different practices, the latter of which may (but need not) accompany the former (Zaibert 2009). An exception may be Marilyn McCord Adams’s account, according to which external (“performative”) and internal (“from the heart”) forgiveness can occur independently of one another (1991: 294). The other distinction that might motivate pluralism—​and has motivated reconciling monisms like Calhoun’s and Fricker’s—​is that between elective and non-​elective forgiveness. For example, suppose that Kelly was seriously injured as a result of Steve’s recklessness and that he has avoided her since the accident. In response, Kelly might forgive Steve despite his unwillingness to take responsibility and apologize because she suspects that it is his shame and defensiveness that is keeping him from repenting and atoning; at the same time, she may be unwilling or unable to forgive

Forgiveness   603 without an apology. On Bennett’s (2003) pluralist view, Kelly offers personal forgiveness, which is elective, but not redemptive forgiveness, which is owed or not depending on the offender’s response to the victim’s blame.

4.  Implications Different accounts of forgiveness emphasize different features or versions of the practice. In this section, I highlight some of the implications of the Standard View and the Norm Changing View. In doing so, I focus on how they bear on salient experiences of forgiveness, key debates about forgiveness, and potential limits of forgiveness. Difficulty. Sometimes forgiving is difficult. Steve may want to forgive Dylan for betraying him, but be unable to do so. He may struggle with the decision itself, wondering if Dylan really deserves his forgiveness; or he may have decided to forgive but struggle to actually follow through. Both scenarios seem true to life. The Standard View recognizes both dimensions of difficulty as obstacles to forgiveness. However, if forgiveness does not require an emotional change, then the struggle to follow through is not part of forgiving. This particular difficulty may be real and significant, but it is not an obstacle to forgiving. One accomplishes forgiveness by making or communicating one’s decision and this does not require overcoming blame—​though it may be difficult to communicate one’s decision if one has not overcome one’s blame. Which view one thinks best captures the difficulty of forgiving will depend on a number of factors, only some of which I have taken up in this chapter. These include whether one sees forgiveness in terms of its normative significance or in terms of its role in our responsibility practice, and whether one conceives of forgiveness as entirely voluntary or as partially involuntary. Reversal. Sometimes a person forgives, but changes their mind and withdraws their forgiveness. Brandon might forgive Valerie for lying to him because she apologizes and appears remorseful. However, if he learns that her apology was insincere and her remorse feigned, he might withdraw his forgiveness and continue blaming her. Again, such scenarios seem true to life.18 The Standard View takes such phenomena at face value, as can some other views (Zaibert 2009). Other accounts have a harder time capturing this possibility. If forgiving alters what it is permissible to demand of or do to the offender (Nelkin 2013; Warmke 2016a and 2016b), then withdrawing forgiveness would seem to be either impossible or else to require that the norm change is also reversed. However, one might think that the significance 18 

There is also a further question about withdrawing forgiveness. Geoffrey Scarre (2016) allows that forgiveness can be withdrawn in cases like Brandon’s, but denies that one withdraws forgiveness just because one’s hard feelings return. Forgiveness can only be withdrawn in response to recognizing that one forgave on the basis of bad information.

604   Per-Erik Milam of a norm change comes from its irreversibility. Forgiving, like waiving a promise, may be significant because one cannot reclaim the normative power once it has been used. Mistakes. Sometimes a person is mistaken about whether they have forgiven. Donna may think she has forgiven David for betraying her trust, but slowly come to realize that she hasn’t. This realization may stem from the recognition that she still ruminates on David’s misbehavior or from others pointing out that she still seems detached and suspicious of him—​that is, she blames him. The Standard View gives a straightforward account of such mistakes. Whether one has actually stopped blaming the offender is not always or entirely transparent to the would-​ be forgiver, so one might be mistaken about whether one has forgiven. Moreover, given the valorization of forgiveness in many societies, we might often be motivated to believe we have forgiven when we actually haven’t. The Norm Changing View also has room for mistaken forgiveness, but the phenomenon it describes is different. If forgiveness is a decision or communicative act, then mistakes like Donna’s are less common and may be impossible. However, one can be mistaken about whether one has forgiven insofar as one can be mistaken about the fact of the wrongdoing and, thus, about whether one has actually altered the relevant norms. For example, Andrea might forgive Brandon for allowing an anti-​Semitic public figure to speak on their campus, only to realize that Brandon wasn’t the one who made the decision. Thus, while she intended to release him from the obligation created by his offense, she didn’t have the power to do so because he hadn’t committed the offense, and she is thus mistaken about having forgiven. Non-​Victim Forgiveness.  Our everyday understanding of forgiveness seems to allow that individuals other than the victim can forgive. We speak of individuals forgiving themselves, third parties forgiving those who have harmed their friends, and groups seeking and offering forgiveness. These are familiar phenomena and are recognized as cases of forgiveness.19 A theory must either account for these practices as forgiveness or explain why they are not. The Standard View takes the first approach. Anyone can, in principle, overcome their blame and do so for the right kinds of reasons. If Steve can forgive Valerie for lying to him, why can’t Valerie forgive herself, or Brandon, as Steve’s friend, forgive her for the same reasons? However, if forgiving changes the normative situation between victim and offender, and only a victim can do this, then non-​victim forgiveness is impossible. While Valerie and Brandon might do something that looks like forgiveness, only Steve can make the relevant change. Thus, whether one finds one or the other position more plausible will depend on whether they can provide satisfying answers to various questions. For example, if non-​victim forgiveness is impossible, then what exactly are agents like Valerie and Brandon doing when they try to forgive? And what exactly do we object to when we view self-​forgiveness and third-​party forgiveness as inappropriate or wrongful?

19 

On self-​forgiveness, see Holmgren (1998), Dillon (2001), and Milam (2017). On third-​party forgiveness, see Pettigrove (2012), Walker (2013), and MacLachlan (2017). On political forgiveness, see Amstutz (2004) and MacLachlan (2012).

Forgiveness   605

5.  Conclusion The aim of this chapter has been to provide an introduction to the philosophical debate about the nature, function, and ethics of forgiveness. In particular, I have tried to place that debate in the context of a broader discussion of the nature of moral responsibility and to highlight the connections between these two contexts. I began by presenting the Standard View of Forgiveness, according to which to forgive is to overcome one’s blame toward a culpable wrongdoer. I then introduced a number of objections to this view, each of which disputed the nature of the change required for forgiveness. I focused, in particular, on the Norm Changing View, according to which forgiving changes what is permissible and impermissible for the victim and offender to do and to forgive is to bring about that change. I closed by considering the implications of these two views for how we answer some open questions about forgiveness. Inevitably, I have been forced to neglect a number of important topics and questions, including the following: elective vs. non-​ elective forgiveness (Allais 2013; Milam 2018b); standing to forgive (Pettigrove 2009 and 2012); reasons to forgive (Milam 2018a, Schönherr 2018); divine forgiveness (Warmke 2018a and 2018b); the history of forgiveness (Konstan 2010); and how forgiveness works in non-​ideal circumstances (Walker 2006; MacLachlan 2009; Brunning and Milam 2018). However, I hope that the way in which I have introduced the debate will give readers some insight into how to investigate and answer these questions.

References Adams, Marilyn McCord. 1991. “Forgiveness: A Christian Model.” Faith and Philosophy 8.3, 277–​304. Allais, Lucy. 2008. “Wiping the Slate Clean: The Heart of Forgiveness.” Philosophy and Public Affairs 36.1, 33–​68. Allais, Lucy. 2013. “Elective Forgiveness.” International Journal of Philosophical Studies 21.5, 637–​653. Amstutz, Mark. 2005. The Healing of Nations: The Promise and Limits of Political Forgiveness. Lanham, MD: Rowman and Littlefield. Bash, Anthony. 2007. Forgiveness and Christian Ethics. Cambridge: Cambridge University Press. Bell, Macalaster. 2008. “Forgiving Someone for Who They Are (and Not Just What They’ve Done).” Philosophy and Phenomenological Research 77.3, 625–​658. Bell, Macalaster. 2013. Hard Feelings: The Moral Psychology of Contempt. Oxford: Oxford University Press. Bennett, Christopher. 2003. “Personal and Redemptive Forgiveness.” European Journal of Philosophy 11.2, 127–​144. Bennett, Christopher. 2018. “The Alteration Thesis: Forgiveness as a Normative Power.” Philosophy and Public Affairs 46. 2, 207–​233. Berman, Mark. 19 June 2015. “‘I Forgive You’: Relative of Charleston Church Shooting Victims Address Dylann Roof.” The Washington Post. Accessed 3 December 2017.

606   Per-Erik Milam Björnsson, Gunnar. 2017. “Explaining Away Epistemic Skepticism about Culpability,” in Oxford Studies in Agency and Responsibility. Ed. David Shoemaker. Oxford: Oxford University Press, 141–​164. Blustein, Jeffrey. 2014. Forgiveness and Remembrance: Remembering Wrongdoing in Personal and Private Life. Oxford: Oxford University Press. Brunning, Luke and Per-​Erik Milam. 2018. “Oppression, Forgiveness, and Ceasing to Blame.” Journal of Ethics and Social Philosophy 14.2, 143–​178. Butler, Joseph. 1900. The Works of Bishop Butler. London: Macmillan. Calhoun, Cheshire. 1992. “Changing One’s Heart.” Ethics 102.1, 76–​96. Calhoun, Cheshire. 2004. “Common Decencies,” in Setting the Moral Compass: Essays by Women Philosophers. Oxford: Oxford University Press, 128–​141. Coates, D. Justin and Neal. A. Tognazzini. 2012. “The Nature and Ethics of Blame.” PhilosophyCompass 7.3, 197–​207. Cornell, Nicolas. 2017. “The Possibility of Preemptive Forgiveness.” Philosophical Review 126.2, 241–​272. Dillon, Robin S. 2001. “Self-​Forgiveness and Self-​Respect.” Ethics 112, 53–​83. Downie, R. S. 1965. “Forgiveness.” The Philosophical Quarterly 15.59, 128–​134. Driver, Julia. 1992. “The Suberogatory.” Australasian Journal of Philosophy 70.3, 286–​295. Fehr, Ryan, Michele J. Gelfand, and Monisha Nag. 2010. “The Road to Forgiveness: A Meta-​ Analytic Synthesis of Its Situational and Dispositional Correlates.” Psychological Bulletin 136.5, 894–​914. Fricker, Miranda. 2019. “Forgiveness: An Ordered Pluralism.” Australasian Philosophical Review 3.1, 241–​260. Gamlund, Espen. 2011. “Forgiveness Without Blame,” in The Ethics of Forgiveness. Ed. Christel Fricke. New York: Routledge, 107–​129. Garcia, Ernesto V. 2011. “Bishop Butler on Forgiveness and Resentment.” Philosopher’s Imprint 11.10, 1–​19. Garrard, Eve and David McNaughton. 2003. “In Defence of Unconditional Forgiveness.” Proceedings of the Aristotelian Society 100, 39–​60. Garrard, Eve and David McNaughton. 2010. Forgiveness. Durham, NC: Acumen Publishing Ltd. Griswold, Charles. 2007. Forgiveness: A Philosophical Exploration. Cambridge: Cambridge University Press. Haber, Joram Graf. 1991. Forgiveness. Savage, MD: Rowman and Littlefield. Harman, Elizabeth. 2016. “Morally Permissible Moral Mistakes.” Ethics 126.1, 366–​393. Hieronymi, Pamela. 2001. “Articulating an Uncompromising Forgiveness.” Philosophy and Phenomenological Research 62.3, 529–​555. Holmgren, Margaret R. 1998. “Self-​Forgiveness and Responsible Moral Agency.” Journal of Value Inquiry 32, 75–​91. Horsbrugh, H. T. N. 1974. “Forgiveness.” Canadian Journal of Philosophy 4.2, 269–​282. Hughes, Paul M. and Brandon Warmke. 2017. “Forgiveness.” Stanford Encyclopaedia of Philosophy (Summer 2017 Edition). Ed. Edward N. Zalta. Karremans, Johan C. et al. 2003. “When Forgiving Enhances Psychological Well-​Being: The Role of Interpersonal Commitments.” Journal of Personality and Social Psychology 84.5, 1011–​1026.

Forgiveness   607 Khoury, Andrew C. and Benjamin Matheson. 2018. “Is Blameworthiness Forever?” Journal of the American Philosophical Association 4.2, 204–​224. Kolnai, Aurel. 1974. “Forgiveness.” Proceedings of the Aristotelian Society 74, 91–​106. Konstan, David. 2010. Before Forgiveness: The Origins of a Moral Idea. Cambridge: Cambridge University Press. Levy, Neil. 2011. Hard Luck. Oxford: Oxford University Press. MacLachlan, Alice. 2009. “Practicing Imperfect Forgiveness,” in Feminist Ethics and Social and Political Philosophy: Theorizing the Non-​Ideal. Ed. Lisa Tessman. 185–​204. MacLachlan, Alice. 2012. “The Philosophical Controversy Over Political Forgiveness” in Public Forgiveness in Post-​Conflict Contexts. Eds. Bas van Stokkom, Neelke Doorn, and Paul van Tongeren. Cambridge: Intersentia Publishing, 37–​64. MacLachlan, Alice. 2017. “In Defense of Third-​Party Forgiveness” in The Moral Psychology of Forgiveness. Ed. Kathryn Norlock. London: Rowman and Littlefield International. 135–​160. McKenna, Michael. 2012. Conversation and Responsibility. Oxford: Oxford University Pres. Milam, Per-​Erik. 2017. “How Is Self-​Forgiveness Possible?” Pacific Philosophical Quarterly 98, 49–​69. Milam, Per-​Erik. 2018. “Against Elective Forgiveness.” Ethical Theory and Moral Practice 21.3, 569–​584. Milam, Per-​Erik 2019. Reasons to forgive. Analysis 79. 2, 242–​251. Murphy, Jeffrie. 1982. “Forgiveness and Resentment.” Midwest Studies in Philosophy 7.1, 503–​516. Murphy, Jeffrie and Jean Hampton. 1988. Forgiveness and Mercy. Cambridge: Cambridge University Press. Murphy, Jeffrie. 1998. “Jean Hampton on Immorality, Self-​Hatred, and Self-​Forgiveness.” Philosophical Studies 89.2, 215–​236. Nelkin, Dana K. 2011. Making Sense of Freedom and Responsibility. Oxford University Press. Nelkin, Dana K. 2013. “Freedom and Forgiveness,” in Free Will and Moral Responsibility. Eds. Ishtiyaque Haji and Justin Caouette. Newcastle: Cambridge Scholars Press. Nelkin, Dana K. 2017. “Blame”. Newberry, Paul A. 2001. “Joseph Butler on Forgiveness: A Presupposed Theory of Emotion.” Journal of the History of Ideas 62.2, 233–​244. Novitz, David. 1998. “Forgiveness and Self-​ Respect.” Philosophy and Phenomenological Research 58.2, 299–​315. Pereboom, Derk. 2001. Living Without Free Will. Cambridge: Cambridge University Press. Pettigrove, Glen. 2009. “The Standing to Forgive.” The Monist 92.4, 583–​603. Pettigrove, Glen. 2012. Forgiveness and Love. Oxford: Oxford University Press. Pronk, Tila M. et al. 2010. “What It Takes to Forgive: When and Why Executive Functioning Facilitates Forgiveness.” Journal of Personality and Social Psychology 98.1, 119–​131. Richards, Norvin. 1988. “Forgiveness.” Ethics 99.1, 77–​97. Rosen, Gideon. 2004. “Skepticism About Moral Responsibility.” Philosophical Perspectives 18, 295–​313. Russell, Luke. 2016. “Forgiving While Punishing.” Australasian Journal of Philosophy 94.4, 704–​7 18. Scarre, Geoffrey. 2016. “On Taking Back Forgiveness.” Ethical Theory and Moral Practice 19.4, 931–​944. Schönherr, Julius. 2019. “When Forgiveness Comes Easy.” Journal of Value Inquiry 53. 4, 513–​528.

608   Per-Erik Milam Strawson, P. F. 1962. “Freedom and Resentment,” in Free Will. Ed. Gary Watson. Oxford: Oxford University Press. Struthers, C. Ward et al. 2008. “The Effects of Attributions of Intent and Apology on Forgiveness: When Saying Sorry May Not Help the Story.” Journal of Experimental Social Psychology 44, 983–​992. Twambley, P. 1976. “Forgiveness.” Analysis 36.2, 84–​90. Vargas, Manuel. 2013. Building Better Beings. Oxford: Oxford University Press. Walker, Margaret Urban. 2006. Moral Repair. Cambridge: Cambridge University Press. Walker, Margaret Urban. 2013. “Third Parties and the Social Scaffolding of Forgiveness.” Journal of Religious Ethics 41.3, 495–​512. Walters, Joanna. 2 October 2006. “‘The Happening’: 10 Years after the Amish Shooting.” The Guardian. Accessed 3 December 2017. Warmke, Brandon. 2011. “Is Forgiveness the Deliberate Refusal to Punish?” Journal of Moral Philosophy 8, 613–​620. Warmke, Brandon. 2015. “Articulate Forgiveness and Normative Constraints.” Canadian Journalof Philosophy 45.4, 490–​514. Warmke, Brandon. 2016a. “The Normative Significance of Forgiveness.” Australasian Journal of Philosophy 94.4, 687–​703. Warmke, Brandon. 2016b. “The Economic Model of Forgiveness.” Pacific Philosophical Quarterly 97.4, 570–​589. Warmke, Brandon. 2017a. “Divine Forgiveness I: Emotion and Punishment-​Forbearance Theories.” Philosophy Compass 12, 1–​9. Warmke, Brandon. 2017b. “Divine Forgiveness II: Reconciliation and Debt-​Cancellation Theories.” Philosophy Compass 12, 1–​8. Zaragoza, Kevin. 2012. “Forgiveness and Standing.” Philosophy and Phenomenological Research 84.3, 604–​621.

CHAPTER 29

Re c onciliation a nd t h e End of Resp onsi bi l i t y Linda Radzik

1.  Introduction In this chapter, I assume that at least many of the familiar practices by which we hold ourselves and one another accountable are justifiable. These practices include blaming people, resenting those who mistreat us, and feeling guilty when we mistreat others. My interest here is in another practice of responsibility: reconciliation. Reconciling involves normalizing relationships in the wake of conflict, in large part by bringing an end to the other practices by which we have been holding ourselves or others accountable. As Herbert Morris notes, “wrongdoing arises in a world in which there is a conception of righting the wrong. It arises in a world in which persons possess a conception not just of separation from others but of coming together again with them.”1 So, we might think of reconciliation as providing the happy ending to stories of moral responsibility for wrongdoing. But reconciliation is a controversial subject—​ especially in political contexts—​ precisely because it seems to suggest an “end” to responsibility. By ceasing to hold wrongdoers responsible for past transgressions, critics suggest, we risk either denying their culpability or denying the significance of their culpability. We wind up sending the message that their actions weren’t really wrong or else that wrongdoing doesn’t really matter. One might reply that reconciliation instead merely indicates that wrongdoers are no longer answerable for their past misdeeds. Yet philosophical analyses of responsibility typically suggest that it is permanent. Moral responsibility for an action is grounded in the person’s causal relationship to the action, in her mental states and

1 

Herbert Morris, “Guilt and Suffering,” in On Guilt and Innocence (Berkeley: University of California Press, 1976), 96.

610   Linda Radzik competence leading up to and at the moment of action, and perhaps in certain counterfactual conditions relating to her ability to do otherwise. The factors that are crucial to responsibility are either prior to or concurrent with the action. A wrongful action might have been excused or justified under the circumstances at the time, but no ameliorative step the person takes later will change the fact that she is blameworthy for this action. But if this is correct, then Morris is mistaken. Wrongs can’t be righted. Instead we face Josiah Royce’s “hell of the irrevocable”: “I am, and to the end of endless time shall remain, the doer of that willfully traitorous deed. Whatever other value I may get, that value I retain forever. My guilt is as enduring as time.”2 No happy endings after all. This chapter introduces current thought about reconciliation, which emerges from moral, legal, and political writings about the aftermath of wrongdoing. We will see how reconciliation, as conceived in these contexts, seems to be in tension with our other practices of responsibility and why, for that reason, those who advocate reconciliation are often accused of condoning wrongdoing. Yet, we will also consider the case for believing reconciliation to be not just permissible but also desirable in a wide range of contexts.

2.  Defining Reconciliation In its ordinary sense, “reconciliation” refers to the restoration of good relations between two or more parties following a period of conflict or alienation. The parties in question may be either individuals or groups. The more technical use of the term “reconciliation,” which appears in the scholarly literature and which is my topic here, differs from the ordinary concept in a few ways. First, while people fall out with one another for all sorts of reasons, philosophical discussions of reconciliation are motivated specifically by cases of wrongdoing or violent conflict.3 Another difference is that, in ordinary discourse, to describe two parties as reconciled is to imply that, before the period of conflict, the parties enjoyed a special, personal relationship (such as a friendship) and that this relationship was a reasonably healthy one. Both of these implications are dropped in the philosophical literature. Parties who were strangers to one another prior to the offense are frequently described as reconciling as fellow citizens or as fellow human beings.4 Parties whose relationship has always been bad are also described as reconciling.5 For an example of the latter usage, we can point to South Africa’s transition from apartheid to 2 

Josiah Royce, The Problem of Christianity (Chicago: University of Chicago Press, 1968), 161–​162. It is at least theoretically possible that none of the parties engaged in a violent conflict, such as civil war, is guilty of wrongdoing. 4  See, for example, Trudy Govier, Forgiveness and Revenge (New York: Routledge, 2002); and Darrel Moellendorf, “Reconciliation as a Political Value,” Journal of Social Philosophy 38, no. 2 (2007): 205–​221. 5  Trudy Govier, Taking Wrongs Seriously: Acknowledgment, Reconciliation and the Politics of Sustainable Peace (Amherst, NY: Humanity Books, 2006); and Colleen Murphy, A Moral Theory of Political Reconciliation (New York: Cambridge University Press, 2010). 3 

Reconciliation and the End of Responsibility    611 democracy in the 1990s, which presented itself as a politics of reconciliation despite the fact that the white population had never lived on decent terms with the Black population.6 In its philosophical usage, “reconciliation” is taken to include the establishment of good relations for the first time. Finally, the philosophical conception of reconciliation does not even require that at least two parties be involved. A wrongdoer who overcomes guilt, self-​doubt, or dysfunction is sometimes described as reconciling with herself.7 To speak of an improvement in relations is to draw attention primarily to how the parties interact—​how they behave with respect to one another (e.g., whether they cooperate, whether they speak civilly). However, these interactions are shaped by underlying beliefs, attitudes, emotions, and expectations, which the parties hold toward one another. So discussions of reconciliation—​and especially discussions of how it is to be achieved and rendered stable—​emphasize improvements in how the parties regard one another. Do they still believe the other poses a risk? Are they still resentful, afraid, or ashamed? Have respect and trust been restored, and if so, is it mutual? Reconciliation is generally considered to be different from, though related to, forgiveness. Scholars offer different accounts of forgiveness, but the most widely used one is an attitude-​centered account according to which forgiveness is a matter of overcoming or forswearing negative reactive attitudes that result from a misdeed, such as resentment, anger, or disappointment.8 Many theorists add that forgiveness also requires one to adopt a more positive attitude toward the wrongdoer, such as compassion or goodwill.9 Given an attitude-​centered account, forgiving is a step in the direction of reconciliation. However, forgiveness alone will fall short of reconciliation if the improvement in attitude has little effect on how the parties interact with one another. Reconciliation also seems to be possible without forgiveness, insofar as continuing resentment can be contained or compartmentalized, so as not to significantly impair the parties’ interactions. The distinction between an attitude-​ centered account of forgiveness and an interaction-​centered account of reconciliation fits with the intuition that it is possible to forgive the dead but not to reconcile with them. Another difference is that forgiveness is one-​directional, proceeding from the forgiver to the forgiven, while reconciliation is mutual. Even if the victim takes generous steps toward reconciliation, if she is rebuffed by the wrongdoer then reconciliation does not occur. The role of third parties differs 6 

Antjie Krog, “‘This Thing Called Reconciliation,’ Forgiveness as Part of an Interconnectedness-​ Towards-​Wholeness,” South African Journal of Philosophy 27, no. 4 (2008): 353–​366. 7  Howard Zehr, Changing Lenses: A New Focus for Crime and Justice (Scottsdale, PA: Herald Press, 1990); and Linda Radzik, Making Amends: Atonement in Morality, Law, and Politics (New York: Oxford University Press, 2009). 8  For example, see Jeffrie G. Murphy and Jean Hampton, Forgiveness and Mercy (New York: Cambridge University Press, 1988); Charles L. Griswold, Forgiveness: A Philosophical Exploration (New York: Cambridge University Press, 2007); and Margaret R. Holmgren, Forgiveness and Retribution: Responding to Wrongdoing (New York: Cambridge University Press, 2012). Alternative accounts of forgiveness are surveyed in Griswold, Forgiveness, and Glen Pettigrove, Forgiveness and Love (New York: Oxford University Press, 2012). 9  Joseph Butler, Fifteen Sermons (Charlottesville, VA: Lincoln-​Rembrandt Publishing, 1993); and Pettigrove, Forgiveness and Love.

612   Linda Radzik as well. Many theorists claim that it is a conceptual truth that only victims can forgive wrongdoers. Third parties to the wrong, no matter how upset or indignant they may be, are in no position to do anything properly described as “forgiving.”10 However, the idea of third parties reconciling with wrongdoers is taken to be conceptually unproblematic, even though it may raise other kinds of objections. Different kinds of relationships, such as being close friends, casual friends, spouses, or fellow citizens, are characterized by typical patterns of attitudes, interactions, and expectations, as well as by different norms for what counts as a good relationship of that type. What is expected of a close friend may be improper from a mere neighbor. So, discussions of reconciliation are typically relative to the relationship in question. Two people might reconcile as business partners after a transgression but lose their friendship. A couple might repair their friendship yet decide to divorce. Reconciliation is a matter of degree, since the ways in which relationships improve are a matter of degree.11 The closer the relationship comes to the paradigm of a morally acceptable relationship of that type, the greater the degree of reconciliation. For this reason, we can characterize reconciliation as the normalization of relationships, where “normal” is an evaluative rather than a statistical standard.12 When relationships are normalized through reconciliation, the degrees of mutual respect, goodwill, and trust that are appropriate to a relationship of this type (e.g., casual friendship, marriage) have been restored or established. However, especially in the political literature, theorists disagree about how much a relationship must improve before it should be described as a successful reconciliation. In the aftermath of civil war or severe human rights abuses, mere peaceful coexistence can be a major achievement, which might be worth acknowledging by applying a term with positive connotations.13 At the other end of the spectrum, theorists require something more like political solidarity or “a shared vision of a collective future” before they will describe a group as reconciled.14 Such strong versions of political reconciliation face 10  For the debate about whether third parties can forgive, see Pettigrove, Forgiveness and Love, ch. 2, and Margaret Urban Walker, “Third Parties and the Social Scaffolding of Forgiveness,” Journal of Religious Ethics 41, no. 3 (2013): 495–​512. The popular view that only victims can forgive makes for an awkward combination with the equally common view that people can forgive themselves for wronging others, as noted in Per-​Erik Milam, “How Is Self-​Forgiveness Possible?” Pacific Philosophical Quarterly 98, no. 1 (2017): 49–​69. 11  David A. Crocker, “Reckoning with Past Wrongs: A Normative Framework,” Ethics and International Affairs 13, no. 1 (1999): 43–​64; and Griswold, Forgiveness. 12  I am assuming here that the relationship is not inherently morally problematic, such as mugger/​mugged or con-​man/​patsy. Reconciliation efforts in the aftermath of a crime, such as those recommended by restorative justice models, typically pursue the reconciliation of the parties as fellow citizens or fellow human beings. See, for example, John Braithwaite, “Repentance Rituals and Restorative Justice,” Journal of Political Philosophy 8, no. 1 (2000): 115–​131. 13  Rajeev Bhargava, “The Difficulty of Reconciliation,” Philosophy and Social Criticism 38, no. 4–​5 (2012): 369–​377. 14  Timothy Longman, Phuong Pham, and Harvey M. Weinstein, “Connecting Justice to Human Experience: Attitudes Toward Accountability and Reconciliation in Rwanda,” in My Neighbor, My Enemy: Justice and Community in the Aftermath of Mass Atrocity, eds. Eric Stover and Harvey M. Weinstein (New York, Cambridge University Press, 2004), 207. See also Desmond Tutu, No Future

Reconciliation and the End of Responsibility    613 criticism as illiberal (insofar as they leave little room for reasonable disagreement over values)15 or as antithetical to the interests of minority groups (who may prefer greater sovereignty or even secession rather than political unity with the majority).16 Where the minority groups have been the victims of severe oppression or atrocities at the hands of the majority, as they have been in white settler states such as the United States, Canada, and Australia, the critique of strong accounts of reconciliation is particularly powerful.17 More moderate accounts of political reconciliation focus on values such as democratic reciprocity or mutual respect for agency.18

3.  Reconciling as Ceasing to Hold Responsible This brief introduction to the philosophical conception of reconciliation helps us see how reconciling involves a relenting in the means by which parties hold themselves and others responsible for wrongful or violent behavior.19 We can see this by noting the changes in relationships that count as reconciling. The negative beliefs, attitudes, and behaviors that are rendered justifiable (or perhaps excusable) by wrongdoing are ones that are dropped, to one degree or another, in reconciling.20 Victims and third parties become less inclined to believe that the wrongdoer is dangerous or morally corrupt. The parties feel less resentful, indignant, ashamed or guilty. They are increasingly willing to Without Forgiveness (New York: Image/​Doubleday, 1999); and Simon Cabulea May, “Moral Compromise, Civic Friendship, and Political Reconciliation,” CRISPP: Critical Review of International Social and Political Philosophy 14, no. 5 (2011): 581–​602. 15  Amy Gutmann and Dennis Thompson, “The Moral Foundations of Truth Commissions,” in Truth v. Justice: The Morality of Truth Commissions, eds. Robert I. Rotberg and Dennis Thompson (Princeton, NJ: Princeton University Press, 2000), 22–​44. 16  Andrew Woolford, “The Limits of Justice: Certainty, Affirmative Repair, and Aboriginality,” Journal of Human Rights 3, no. 4 (2004): 429–​444. 17  Jeff Corntassel and Cindy Holder, “Who’s Sorry Now? Government Apologies, Truth Commissions, and Indigenous Self-​Determination in Australia, Canada, Guatemala, and Peru,” Human Rights Review 9, no. 4 (2008): 465–​489; and Victoria Freeman, “In Defence of Reconciliation,” Canadian Journal of Law and Jurisprudence 27, no. 1 (2014): 213–​223. 18  Gutmann and Thompson, “The Moral Foundations of Truth Commissions”; and C. Murphy, A Moral Theory of Political Reconciliation. 19  An assumption of this discussion is that holding S responsible for A is distinct from merely believing that S is responsible for A. Intuitively, the following statement makes sense: “I believe that he is responsible, but I’m not going to hold him responsible.” Such a speaker might consider the wrong in question none of her business, or believe that she would be a hypocrite if she held the wrongdoer responsible. 20  See T. M. Scanlon, Moral Dimensions: Permissibility, Meaning, Blame (Cambridge, MA: Belknap/​ Harvard University Press, 2008); Murphy and Hampton, Forgiveness and Mercy; Govier, Taking Wrongs Seriously; and Margaret Urban Walker, Moral Repair: Reconstructing Moral Relations after Wrongdoing (New York: Cambridge University Press, 2006).

614   Linda Radzik trust the other, wish the other well, and hope for good future relations. They are less punitive and better able to function in the presence of the other. They are more willing to cooperate. Where parties reach a high degree of reconciliation, their relationships are no longer tainted by the past wrong. They no longer relate to one another in terms of the roles “victim” or “wrongdoer.”21 The claim that reconciling typically includes ceasing to hold responsible receives further support when we consider the issue of standing. Arguably, only those parties who have the standing to hold the wrongdoer responsible have the option of reconciling. The standing of both victims and wrongdoers is not controversial. Victims may hold their abusers, and wrongdoers may hold themselves, accountable through their attitudes and behaviors. Third parties are more complicated. It is not always clear whether people are entitled to respond to a wrong or conflict to which they are merely bystanders.22 May they express indignation? Are they entitled to utter rebukes, demand that the victim receive an apology, or boycott a wrongdoer’s business? In some cases, third parties should help hold wrongdoers to account. Failing to do so might even amount to complicity in wrongdoing or abandoning the victim. But in other cases, third parties would be justly criticized for not minding their own business. The difference often turns on the type of wrong in question and the third party’s relationship to the main parties of the conflict. It can be difficult to judge who has standing in particular cases, but the standing to hold blameworthy actors responsible seems to correlate with the standing to engage in the kind of reconciliation that is our topic here. Suppose that Joe, after learning of some misdeed by eavesdropping on a private conversation, adopts a rude tone every time he has to interact with the estranged friend of his office mate. After the friends make up, Joe’s interactions with the wrongdoer improve. But describing them as reconciling would lend too much dignity to Joe’s meddling behavior. Were Joe to declare his willingness to put the past behind him, he would likely meet with bewilderment or resentment. The association between reconciling and ceasing to hold the wrongdoers accountable is also found in the political realm. In the last 30 years or so, reconciliation has formed a prominent part of the agenda of states attempting to transition from repressive rule or civil conflict to better forms of government.23 Such agendas often include amnesty programs.24 Criminal prosecution is the long-​established practice by which states hold 21 

Jean Hampton characterizes forgiveness this way, but it is a better description of reconciliation. Murphy and Hampton, Forgiveness and Mercy, 36–​43. 22  Linda Radzik, “On Minding Your Own Business: Differentiating Accountability Relations Within the Moral Community,” Social Theory & Practice 37, no. 4 (2011): 574–​598. 23  Martha Minow, Between Vengeance and Forgiveness: Facing History After Genocide and Mass Violence (Boston: Beacon Press, 1998); and Ruti Teitel, Transitional Justice (New York: Oxford University Press, 2002). 24  David Dyzenhaus, “Justifying the Truth and Reconciliation Commission,” Journal of Political Philosophy 8, no. 4 (2000): 470–​496; and Mark Freeman, Necessary Evils: Amnesties and the Search for Justice (New York: Cambridge University Press, 2011).

Reconciliation and the End of Responsibility    615 people responsible for significant criminal wrongdoing. Yet, in the name of reconciliation, some new governments have granted blanket amnesties for past crimes, including terrible human rights violations, such as murder and torture. In other cases, amnesty was made conditional on providing testimony to a truth commission. In many of these transitional states, the choice was described in stark terms. Citizens would have to choose between peace and justice, between reconciliation and accountability.25 Not surprisingly, reconciliation efforts have frequently been condemned as condoning wrongdoing. Amnesty programs amount to a policy of impunity for gross human rights violations. States that prioritize reconciliation over punitive justice are charged with acquiescing in evil or with forgetting the past.26 They are accused of victimizing the survivors a second time, and as forcing victims to bear the costs of future peace in addition to the harms of the past.27 While the charge of condoning wrongdoing is particularly heavy in cases of political reconciliation, given the severity of past wrongs, the same objection appears in other areas of the literature on both reconciliation and forgiveness. Recall, the attitude-​ centered account of forgiveness portrays forgiving as the victim overcoming or forswearing attitudes like resentment or anger, and (on some versions) restoring goodwill. Yet, resentment, anger, and the withdrawal of goodwill are paradigmatic ways by which a victim holds the person who mistreated him accountable. So, unsurprisingly, one of the central debates in the moral literature on interpersonal forgiveness turns on whether it is possible for a victim to overcome resentment without thereby condoning the wrong.28 In discussing the role of third parties in the aftermath of interpersonal wrongdoing, Margaret Urban Walker introduces the category of “normative abandonment.”29 Her concern is that, “When wrong is done to someone, then, the failure of others to be moved to the censure expressed by resentment [or indignation] is an abandonment of either the norm, or the responsibility of the offender, or the victim.”30 Insofar as forgiveness or reconciliation relents in holding a blameworthy party responsible, it acquiesces in the wrong.

25 

Robert I. Rotberg and Dennis Thompson, eds., Truth v. Justice: The Morality of Truth Commissions (Princeton, NJ: Princeton University Press, 2000); and Daniel Philpott, Just and Unjust Peace: An Ethic of Political Reconciliation (New York: Oxford University Press, 2012). 26  Paul M. Hughes, “Moral Atrocity and Political Reconciliation: A Preliminary Analysis,” International Journal of Applied Philosophy 15, no. 1 (2001): 123–​133; and Max Pensky, “Amnesty on Trial: Impunity, Accountability, and the Norms of International Law,” Ethics and Global Politics 1, no. 1–​2 (2008): 1–​40. 27  Dyzenhaus, “Justifying the Truth and Reconciliation Commission”; and Brandon Hamber, Transforming Societies After Political Violence: Truth, Reconciliation, and Mental Health (Dordrecht: Springer, 2009). 28  Aurel Kolnai, “Forgiveness,” Proceedings of the Aristotelian Society 74 (1974): 91–​106; Pamela Hieronymi, “Articulating an Uncompromising Forgiveness,” Philosophy and Phenomenological Research 62, no. 3 (2001): 529–​555; and Pettigrove, Forgiveness and Love. 29 Walker, Moral Repair, 20. 30  Walker, “Third Parties and the Social Scaffolding of Forgiveness,” 505.

616   Linda Radzik

4.  Reconciliation Without Condonation The chief challenge to justifying practices of reconciliation, then, lies in finding ways to separate reconciliation after wrongdoing from the condonation of wrongdoing. In a few cases, this might be impossible, and any degree of reconciliation would be wrong.31 But, in most cases, reconciliation can become permissible, at least to some degree. A first, important step lies in distancing reconciling from forgetting. Forgetting is sometimes explicitly advocated as a means to reconciliation. In 1999, the prime minister of Cambodia notoriously called on his fellow citizens “to dig a hole and bury the past,” in advocating amnesty for Khmer Rouge leaders.32 But on the whole, advocates of reconciliation, especially those writing about politics, emphasize the importance of preserving the memory of the past as part of the reconciliation process.33 This is evinced by the popularity of physical memorials34 and truth commissions among those who advocate political reconciliation. One argument for truth commissions over criminal trials is that the former are better able to establish and communicate the facts about the past wrongdoing than the latter.35 While offers of amnesty might well be seen as a high price to pay for truth telling, this must be balanced against the difficulty of securing convictions, especially against high-​level offenders in transitional political contexts.36 Official acknowledgments of responsibility for past wrongdoing are also seen as key to successful political reconciliation. In many of these cases, the state publicly denied wrongdoing in the past. Victims were “disappeared” and families were left with no information and no legal recourse.37 Public and detailed accounts of who did what to whom, and what their motives were, are argued to be necessary for survivors’ ability

31 

For the debate about whether a wrong can be unforgivable, see Vladimir Jankélévitch, Forgiveness, trans. Andrew Kelley (Chicago: University of Chicago Press, 2005); and Trudy Govier, “Forgiveness and the Unforgivable,” American Philosophical Quarterly 36 (1999): 59–​75. 32  David Chandler, “Coming to Terms With the Terror and History of Pol Pot’s Cambodia (1975–​79),” in Dilemmas of Reconciliation: Cases and Concepts, eds. C. A. L. Prager and T. Govier, (Waterloo, Ontario: Wilfrid Laurier University Press, 2003), 310. 33 Govier, Taking Wrongs Seriously; and C. Murphy, A Moral Theory of Political Reconciliation. 34  Jeffrey M. Blustein, Forgiveness and Remembrance: Remembering Wrongdoing in Personal and Public Life (New York: Oxford University Press, 2014). 35  Jennifer Llewellyn, and Robert Howse, “Institutions for Restorative Justice: The South African Truth and Reconciliation Commission,” University of Toronto Law Journal 49, no. 3 (1999): 355–​388. 36  Dyzenhaus, “Justifying the Truth and Reconciliation Commission.” 37  Jose Zalaquett, “Confronting Human Rights Violations Committed by Former Governments: Principles Applicable and Political Constraints,” in Transitional Justice, ed. Neil J, Kritz (Washington, DC: United States Institute of Peace Press, 1995), 3–​31; and Lawrence Weschler, A Miracle, A Universe: Settling Accounts with Torturers (Chicago: University of Chicago Press, 1998).

Reconciliation and the End of Responsibility    617 to gain closure and heal psychologically,38 the re-​establishment of the rule of law,39 the rebuilding of trust between the citizenry and officials,40 and the recognition that victimized groups are equally valuable members of the political community.41 In cases of interpersonal wrongdoing as well, the acknowledgment of wrongdoing may be crucial to separating reconciliation from condonation.42 This theme appears frequently in discussions of forgiveness in the moral literature. When a victim fails to feel resentment over being wronged, there is at least prima facie reason to worry whether she respects herself as a valuable person who deserves to be treated properly.43 Even advocates of a relatively demanding forgiveness ethic recommend that victims let themselves feel resentment, and acknowledge that it is fitting, before working to overcome it.44 This advice about forgiveness is also appropriate for victims’ decisions to reconcile. In fact, an honest acknowledgment of the fact of wrongdoing and the wrongdoer’s blameworthiness may be even more important for reconciliation than for forgiveness. Insofar as reconciling involves normalizing interactions with the wrongdoer, the victim leaves herself vulnerable to further mistreatment. Another way in which reconciliation can be separated from condonation is by making reconciliation conditional. One might argue that a victim’s willingness to reconcile with a wrongdoer is properly made conditional on the wrongdoer making amends.45 Apologies and promises not to repeat the wrong help to withdraw the insults and the threats that the wrongful action communicated. Voluntary offers of restitution repair the harms caused by wrongdoing, symbolize the wrongdoer’s remorse, and provide evidence of renewed respect for the victim and his rightful claims. Given that normalizing relations with the wrongdoer typically leaves the victim open to future harm, victims are entitled to require evidence of renewed trustworthiness. Depending on the nature of the wrongful action, the victim might also require some sort of acknowledgment of wrongdoing from the community, such as an expression of censure or the punishment of the offender. As Walker says, communities provide the “social scaffolding” for moral repair—​the context in which norms are established, communicated, and maintained.46 Waiting until conditions of atonement, social sanctioning or legal punishment have been met before dropping the attitudes and

38 

Daniel W. van Ness and Karen Heetderks Strong, Restoring Justice, 2nd ed. (Cincinnati, OH: Anderson Publishing, 2002). For evidence against this claim, see Hamber, Transforming Societies after Political Violence. 39  C. Murphy, A Moral Theory of Political Reconciliation. 40  James L. Gibson, Overcoming Apartheid: Can Truth Reconcile a Divided Nation? (New York: Russell Sage, 2004). 41  Llewellyn and Howse, “Institutions for Restorative Justice”; and Elizabeth Kiss, “Moral Ambition within and Beyond Political Constraints: Reflections on Restorative Justice,” in Truth v. Justice, 68-​98. 42 Walker, Moral Repair; and Radzik, Making Amends. 43  Murphy and Hampton, Forgiveness and Mercy. 44 Holmgren, Forgiveness and Retribution. 45  For accounts of making amends, see Walker, Moral Repair; and Radzik, Making Amends. 46  Walker, “Third Parties and the Social Scaffolding of Forgiveness.”

618   Linda Radzik behaviors characteristic of holding wrongdoers responsible would clearly distinguish reconciling from condoning. Notice, punishment and reconciliation are not necessarily incompatible.47 Reconciliation is indeed a relenting in holding wrongdoers accountable. But the relenting might come after the wrongdoers have already been held accountable by appropriate and proportional sanctions. While recent history provides multiple examples where reconciliation has been promoted as an alternative to punitive forms of justice,48 they are not necessarily opposed to one another. Indeed, reconciliation remains an important moral issue even where punitive justice has been secured. Will criminal offenders be fully integrated back into civil society after serving their prison terms? Does the criminal punishment of corrupt leaders provide sufficient reason for the citizenry to reconcile with the state, or for citizens to reconcile with one another? That being said, we should resist the simple equation of justice with punitive justice. Punishment has a role to play in the aftermath of wrongdoing, but so does the corrective justice of restitution and reparations.49 Some theorists in this literature suggest that the healing of relationships is itself a form of justice, called restorative justice.50 Even without adopting a restorative justice framework, however, one may reject the claim that reconciliation is necessarily in tension with justice. Although extraordinary circumstances might force people to choose between peace and justice, in most cases, justice helps secure peace. The legitimacy of making reconciliation conditional upon the satisfaction of just claims is important. But some unconditional forms of reconciliation may also be able to avoid condonation, especially in interpersonal rather than political contexts. Victims sometimes find ways to communicate their clear opposition to the wrong while still reconciling with unrepentant, unpunished wrongdoers. One way to do this is by a gracious extension of forgiveness to the recalcitrant wrongdoer. Using the language of forgiveness marks the act as wrong; yet, the compassion shown to the wrongdoer models a better way of relating to others and may even inspire repentance and reform.51 Rather

47 

Christopher Bennett, The Apology Ritual: A Philosophical Theory of Punishment (New York: Cambridge University Press, 2008); and Ernesto Verdeja, Unchopping a Tree: Reconciliation in the Aftermath of Political Violence (Philadelphia: Temple University Press, 1990). 48  These include defenses of reconciliation in criminal justice contexts. See Zehr, Changing Lenses; and Gerry Johnstone, Restorative Justice: Ideas, Values, Debates (Collompton UK: Willan Publishing, 2002). 49  Elazar Barkan, Guilt of Nations: Restitution and Negotiating Historical Injustices (New York: W. W. Norton & Co., 2000); Janna Thompson, Taking Responsibility for the Past: Reparation and Historical Injustice (Cambridge: Polity Press, 2002); and Roy L. Brooks, Atonement and Forgiveness: A New Model for Black Reparations (Berkeley: University of California Press, 2004). 50  South African Truth and Reconciliation Commission Final Report, 1998. Vol. 1, Ch. 1, para. 36; Johnstone, Restorative Justice; and Lucy Allais, “Restorative Justice, Retributive Justice, and the South African Truth and Reconciliation Commission,” Philosophy and Public Affairs 39, no. 4 (2012): 331–​363. 51  Eve Garrard and David McNaughton, “Conditional Unconditional Forgiveness,” in The Ethics of Forgiveness: A Collection of Essays, ed. Christel Fricke (New York: Routledge, 2011), 97–​106; and Holmgren, Forgiveness and Retribution.

Reconciliation and the End of Responsibility    619 than condoning wrongdoing, the victim inspires the wrongdoer to acknowledge and make amends for the transgression. As I noted earlier, third parties can also influence the dynamics of reconciliation between the main parties to the conflict. Where wrongdoers remain unrepentant, for example, third parties can help pave the way to reconciliation. Their indignation might encourage the wrongdoer to make amends, either by helping her see that her action was indeed wrong or by providing her with an incentive to repair it. By clearly marking the wrong as a wrong, third parties can also reduce the victim’s need to continue protesting the wrong and leave him more free to reconcile. Of course, third parties can also make things worse. In some cases, the failure of third parties to hold wrongdoers responsible further undermines and harms the victim. Cases of bullying provide poignant examples. Studies suggest that most bullying is performed before an audience; the humiliation of the victim is one of the characteristic harms of bullying.52 In cases like these, the failure of third parties to protest the wrong arguably makes them complicit in it, turning them from mere bystanders into wrongdoers. So, third parties influence whether and how victims and wrongdoers resolve their conflicts with one another. But they also face decisions about how to repair their own relationships with either the wrongdoer or the victim. Members of the local community who witness or come to know about the wrongdoer’s misdeed may react with indignation or mistrust and decide not interact with the wrongdoer in the future. In other cases, like those involving bullying, the wrong might also damage the relationship between the victim and community members. Shame and uncertainty make it difficult for them to interact normally. Writers in these literatures often suggest that victims have certain sorts of prerogatives in making decisions about how reconciliation will proceed, including the reconciliation between the wrongdoer and third parties. For example, victims are said to have a prerogative in forgiveness. Third parties may not declare someone forgiven on the victim’s behalf; nor may others attempt to coerce them into forgiving.53 In the case of South Africa’s Truth and Reconciliation Commission, many objected that the state was wrong to reconcile its own relationship with the wrongdoers before victims had made their decisions.54 Victims are seen as having a special degree of authority in specifying what the conditions for reconciliation should be.55 Even in criminal justice cases, where decisions about how to sanction wrongdoers have historically been left to juries and judges, some theorists argue that the victims of crime should be allowed to help

52 

P. Flaspohler, J. Elfstrom, K. Vanderzee, H. Sink, and Z. Birchmeier, ‘‘Stand by Me: The Effects of Peer and Teacher Support in Mitigating the Impact of Bullying on Quality of Life,’’ Psychology in the Schools 46, no. 7 (2009): 636–​649. 53  Trudy Govier and Wilhelm Verwoerd, “Forgiveness: The Victim’s Prerogative,” South African Journal of Philosophy 21, no. 2 (2002a): 97–​111. 54  Govier and Verwoerd, “Forgiveness,” 97–​111. 55 Radzik, Making Amends, ch. 5.

620   Linda Radzik determine sentences.56 But, I would argue that a theory of reconciliation should also recognize that victims sometimes get things wrong. Victims may set conditions on reconciliation that are disproportionate or inherently objectionable. They may suffer from vices such as hard-​heartedness or hypocrisy. When this is the case, third parties must make their own decisions about reconciling with wrongdoers, and wrongdoers must find other ways of resolving their negative self-​regarding attitudes. In this section, we have explored a variety of strategies that parties can follow to prevent reconciliation from collapsing into condonation—​to prevent a relenting in holding a wrongdoer responsible from collapsing into a denial of wrongdoing or a lack of concern about wrongdoing. One might add that time alone can make a difference, at least in some cases. The ordinary transgressions of everyday life might well have expiration dates, such that only people lacking in virtue will let them continue to disrupt normal relations. Holding wrongdoers accountable is one thing; holding a grudge is something else. Finally, in evaluating the objection that reconciliation involves condoning wrongdoing, we should keep in mind that reconciliation is a scalar notion. Relationships can be reconciled to varying degrees, and this means that a reduction in some of the attitudes and behaviors by which we hold one another responsible is compatible with the continuation of others. One might renew cooperation while continuing to resent, for example. This point is especially important in political cases, where the preservation of the memory of the past—​of what was done and who was responsible for it—​is one of the best defenses against its being repeated. Full reconciliation, wherein the roles of “victim” and “wrongdoer” are fully dropped, may not be permissible in all cases. Yet, even in these cases, given the value of restored relationships—​of cooperation, trust, and good will—​partial reconciliation can be morally commendable.

5.  Conclusion There is surely a sense of responsibility under which Royce’s claim that guilt is timeless is correct. That action A at time T was attributable to subject S is timelessly true. Assuming that the action was wrong in a way that was neither justified nor excused for S at time T, then S was guilty of this wrong; and this is also timelessly true. But this is not the only sense of responsibility that we have. One might also agree with the Strawsonian view that a person is accountable for a wrongful action only if it is legitimate to hold that person accountable.57 In thinking through the nature and ethics of reconciliation, we 56 

Nils Christie, “Conflicts as Property,” British Journal of Criminology 17, no. 1 (1977): 1–​26; and Braithwaite, “Repentance Rituals and Restorative Justice.” 57  P. F. Strawson, “Freedom and Resentment.” Proceedings of the British Academy 48 (1962): 187–​ 211. On the distinction between attributability and accountability see Gary Watson, “Two Faces of Responsibility,” Philosophical Topics 24, no. 2 (1996), 227–​248.

Reconciliation and the End of Responsibility    621 are led to ask: held accountable by whom, and by what modes of response; held accountable when, and for how long? Our practices of responsibility include practices through which we relent in holding the wrongdoer accountable. If the reader is inclined to agree with me that there is value in reconciliation—​in letting go of anger, working through distrust and alienation, and re-​establishing a morally healthy relationship—​then a few further conclusions suggest themselves. First, those wrongs for which accountability must be permanent are likely the exception rather than the rule. Some wrongs may be such that a significant degree of reconciliation with the wrongdoer would be intolerable. But, in most cases, wrongdoers apologize, they suffer punishment, victims forgive, or time passes; and it is enough.58 Continuing to hold the wrongdoer accountable would itself be objectionable. Secondly, as Angela M. Smith has noted, whether a person is accountable for a wrongful act in the Strawsonian sense of being a legitimate target of responses like resentment will depend on a number of factors that have nothing to do with that act or what the agent was like when she performed it.59 The wrongdoer’s accountability will depend, at least in part, on other people’s circumstances, including whether anyone else has the standing to hold the actor responsible. Smith presents this as an objection to Strawsonian accounts, suggesting that they have simply changed the subject away from a core sense of moral responsibility to an ethics of moral response.60 But perhaps we should consider the possibility that at least some aspects of moral responsibility are deeply social in just the way that the Strawsonian view implies. Perhaps the question of whether a particular person is accountable for a particular wrong does not have a single answer. Perhaps instead the answer must be relative to different members of the moral community and different times. At a particular time, the wrong might have been righted within some of these relationships, though not in others. Finally, we should note an advantage of the view that responsibility for wrongdoing can “end” in the sense that practices of holding wrongdoers responsible can be brought to a satisfactory resolution. It encourages us to take a more optimistic view of people who are guilty of wrongdoing: They are redeemable. They can see the error of their ways, apologize, and make amends. The very capacities for agency and moral decision-​making that make people responsible for their misdeeds in the first place also enable them to work through that responsibility. This is a happy ending, indeed, especially since we are sometimes the wrongdoers.61

58  French defends the thesis that what a wrongdoer does after the action can change whether he is properly blamable for the action. However, French’s argument, unlike the one I offer here, turns on a change of identify in the wrongdoer. Peter A. French, “Self-​Blaming, Repentance, and Atonement,” Journal of Value Inquiry, 48, no. 4 (2014): 587–​602. 59  Angela M. Smith, “On Being Responsible and Holding Responsible,” Journal of Ethics 11, no. 4 (2007): 565. 60  See also Angela M. Smith “Responsibility as Answerability,” Inquiry 58, no. 2 (2015): 99–​126. 61  The author thanks Dana Kay Nelkin, Derk Pereboom, Colleen M. Murphy, and Robert R. Shandley for helpful comments on earlier versions of this material.

622   Linda Radzik

References Allais, Lucy. 2012. “Restorative Justice, Retributive Justice, and the South African Truth and Reconciliation Commission.” Philosophy and Public Affairs 39, no. 4: 331–​363. Barkan, Elazar. 2000. Guilt of Nations: Restitution and Negotiating Historical Injustices. New York: W. W. Norton & Co. Bennett, Christopher. 2008. The Apology Ritual: A Philosophical Theory of Punishment. New York: Cambridge University Press. Bhargava, Rajeev. 2012. “The Difficulty of Reconciliation.” Philosophy and Social Criticism 38, no. 4–​5: 369–​377. Blustein, Jeffrey M. 2014. Forgiveness and Remembrance: Remembering Wrongdoing in Personal and Public Life. New York: Oxford University Press. Braithwaite, John. 2000. “Repentance Rituals and Restorative Justice.” Journal of Political Philosophy 8, no. 1: 115–​131. Brooks, Roy L. 2004. Atonement and Forgiveness: A New Model for Black Reparations. Berkeley: University of California Press. Butler, Joseph. 1993. Fifteen Sermons. Charlottesville, VA: Lincoln-​Rembrandt Publishing. Chandler, David. 2003. “Coming to Terms with the Terror and History of Pol Pot’s Cambodia (1975–​79).” In Dilemmas of Reconciliation: Cases and Concepts, edited by C. A. L. Prager and T. Govier, 307–​326. Waterloo, Ontario: Wilfrid Laurier University Press. Christie, Nils. 1977. “Conflicts as Property.” British Journal of Criminology 17, no. 1: 1–​26. Corntassel, Jeff, and Cindy Holder. 2008. “Who’s Sorry Now? Government Apologies, Truth Commissions, and Indigenous Self-​Determination in Australia, Canada, Guatemala, and Peru.” Human Rights Review 9, no. 4: 465–​489. Crocker, David A. 1999. “Reckoning with Past Wrongs: A Normative Framework.” Ethics and International Affairs 13, no. 1: 43–​64. Dyzenhaus, David. 2000. “Justifying the Truth and Reconciliation Commission.” Journal of Political Philosophy 8, no. 4: 470–​496. Flaspohler, P., J. Elfstrom, K. Vanderzee, H. Sink, and Z. Birchmeier. 2009. “Stand by Me: The Effects of Peer and Teacher Support in Mitigating the Impact of Bullying on Quality of Life.” Psychology in the Schools 46, no. 7: 636–​649. Freeman, Mark. 2011. Necessary Evils: Amnesties and the Search for Justice. New York: Cambridge University Press. Freeman, Victoria. 2014. “In Defence of Reconciliation.” Canadian Journal of Law and Jurisprudence 27, no. 1: 213–​223. French, Peter A. 2014. “Self-​Blaming, Repentance, and Atonement.” Journal of Value Inquiry 48, no. 4: 587–​602. Garrard, Eve, and David McNaughton. 2011. “Conditional Unconditional Forgiveness.” In The Ethics of Forgiveness: A Collection of Essays, edited by Christel Fricke, 97–​106. New York: Routledge. Gibson, James L. 2004. Overcoming Apartheid: Can Truth Reconcile a Divided Nation? New York: Russell Sage. Govier, Trudy. 1999. “Forgiveness and the Unforgivable.” American Philosophical Quarterly 36: 59–​75. Govier, Trudy. 2002. Forgiveness and Revenge. New York: Routledge. Govier, Trudy. 2006. Taking Wrongs Seriously: Acknowledgment, Reconciliation and the Politics of Sustainable Peace. Amherst, NY: Humanity Books.

Reconciliation and the End of Responsibility    623 Govier, Trudy, and Wilhelm Verwoerd. 2002a. “Forgiveness: The Victim’s Prerogative.” South African Journal of Philosophy 21, no. 2: 97–​111. Govier, Trudy, and Wilhelm Verwoerd. 2002b. “Trust and the Problem of National Reconciliation.” Philosophy of the Social Sciences 32, no. 2: 178–​205. Griswold, Charles L. 2007. Forgiveness: A Philosophical Exploration. New York: Cambridge University Press. Gutmann, Amy, and Dennis Thompson. 2000. “The Moral Foundations of Truth Commissions.” In Truth v. Justice: The Morality of Truth Commissions, edited by Robert I. Rotberg and Dennis Thompson, 22–​44. Princeton, NJ: Princeton University Press. Hamber, Brandon. 2009. Transforming Societies After Political Violence: Truth, Reconciliation, and Mental Health. Dordrecht: Springer. Hieronymi, Pamela. 2001. “Articulating an Uncompromising Forgiveness.” Philosophy and Phenomenological Research 62, no. 3: 529–​555. Holmgren, Margaret R. 2012. Forgiveness and Retribution: Responding to Wrongdoing. New York: Cambridge University Press. Hughes, Paul M. 2001. “Moral Atrocity and Political Reconciliation: A Preliminary Analysis.” International Journal of Applied Philosophy 15, no. 1: 123–​133. Jankélévitch, Vladimir. 2005. Forgiveness. Translated by Andrew Kelley. Chicago: University of Chicago Press. Johnstone, Gerry. 2002. Restorative Justice: Ideas, Values, Debates. Cullompton, UK: Willan Publishing. Kiss, Elizabeth. 2000. “Moral Ambition Within and Beyond Political Constraints: Reflections on Restorative Justice.” In Truth v, Justice: The Morality of Truth Commissions, edited by Robert I. Rotberg and Dennis Thompson, 68–​98. Princeton, NJ: Princeton University Press. Kolnai, Aurel. 1974. “Forgiveness.” Proceedings of the Aristotelian Society 74: 91–​106. Krog, Antjie. 2008. “‘This Thing Called Reconciliation,’ Forgiveness as Part of an Interconnectedness-​Towards-​Wholeness.” South African Journal of Philosophy 27, no. 4: 353–​366. Llewellyn, Jennifer, and Robert Howse. 1999. “Institutions for Restorative Justice: The South African Truth and Reconciliation Commission.” University of Toronto Law Journal 49, no. 3: 355–​388. Longman, Timothy, Phuong Pham, and Harvey M. Weinstein. 2004. “Connecting Justice to Human Experience: Attitudes Toward Accountability and Reconciliation in Rwanda.” In My Neighbor, My Enemy: Justice and Community in the Aftermath of Mass Atrocity, edited by Eric Stover and Harvey M. Weinstein, 206–​225. New York, Cambridge University Press. May, Simon Cabulea. 2011. “Moral Compromise, Civic Friendship, and Political Reconciliation.” CRISPP: Critical Review of International Social and Political Philosophy 14, no. 5: 581–​602. Milam, Per-​Erik. 2017. “How Is Self-​Forgiveness Possible?” Pacific Philosophical Quarterly 98, no. 1: 49–​69. Minow, Martha. 1998. Between Vengeance and Forgiveness: Facing History After Genocide and Mass Violence. Boston: Beacon Press. Moellendorf, Darrel. 2007. “Reconciliation as a Political Value.” Journal of Social Philosophy 38, no. 2: 205–​221. Morris, Herbert. 1976. “Guilt and Suffering.” In On Guilt and Innocence, 89–​110. Berkeley: University of California Press. Murphy, Colleen. 2010. A Moral Theory of Political Reconciliation. New York: Cambridge University Press.

624   Linda Radzik Murphy, Jeffrie G., and Jean Hampton. 1988. Forgiveness and Mercy. New York: Cambridge University Press. Pensky, Max. 2008. “Amnesty on Trial: Impunity, Accountability, and the Norms of International Law.” Ethics and Global Politics 1, no. 1–​2: 1–​40. Pettigrove, Glen. 2012. Forgiveness and Love. New York: Oxford University Press. Philpott, Daniel. 2012. Just and Unjust Peace: An Ethic of Political Reconciliation. New York: Oxford University Press. Radzik, Linda. 2009. Making Amends: Atonement in Morality, Law, and Politics. New York: Oxford University Press. Radzik, Linda. 2011. “On Minding Your Own Business: Differentiating Accountability Relations Within the Moral Community.” Social Theory & Practice 37, no. 4: 574–​598. Rotberg, Robert I., and Dennis Thompson, eds. 2000. Truth v. Justice: The Morality of Truth Commissions. Princeton, NJ: Princeton University Press. Scanlon, T. M. 2008. Moral Dimensions: Permissibility, Meaning, Blame. Cambridge, MA: Belknap/​Harvard University Press. Smith, Angela M. 2007. “On Being Responsible and Holding Responsible.” Journal of Ethics 11, no. 4: 465–​484. Smith, Angela M. 2015. “Responsibility as Answerability.” Inquiry 58, no. 2: 99–​126. South African Truth and Reconciliation Commission Final Report. 1998. [Available online http: //​www.justice.gov.za/​trc/​report/​index,htm] Strawson, P. F. 1962. “Freedom and Resentment.” Proceedings of the British Academy 48: 187–​211. Teitel, Ruti. 2002. Transitional Justice. New York: Oxford University Press. Thompson, Janna. 2002. Taking Responsibility for the Past: Reparation and Historical Injustice. Cambridge: Polity Press. Tutu, Desmond. 1999. No Future Without Forgiveness. New York: Image/​Doubleday. van Ness, Daniel W., and Karen Heetderks Strong. 2002. Restoring Justice, 2nd ed. Cincinnati, OH: Anderson Publishing. Verdeja, Ernesto. 2009. Unchopping a Tree: Reconciliation in the Aftermath of Political Violence. Philadelphia: Temple University Press. Walker, Margaret Urban. 2006. Moral Repair: Reconstructing Moral Relations After Wrongdoing. New York: Cambridge University Press. Walker, Margaret Urban. 2013. “Third Parties and the Social Scaffolding of Forgiveness.” Journal of Religious Ethics 41, no. 3: 495–​512. Watson, Gary. 1996. “Two Faces of Responsibility.” Philosophical Topics 24, no. 2: 227–​248. Weschler, Lawrence. 1998. A Miracle, A Universe: Settling Accounts with Torturers. Chicago: University of Chicago Press. Woolford, Andrew. 2004. “The Limits of Justice: Certainty, Affirmative Repair, and Aboriginality.” Journal of Human Rights 3, no. 4: 429–​444. Zalaquett, Jose. 1995. “Confronting Human Rights Violations Committed by Former Governments: Principles Applicable and Political Constraints.” In Transitional Justice, edited by Neil J. Kritz, 3–​31. Washington, DC: United States Institute of Peace Press. Zehr, Howard. 1990. Changing Lenses: A New Focus for Crime and Justice. Scottsdale, PA: Herald Press.

CHAPTER 30

Resp onsibil i t y a nd Religi on Daniel Speak

It would be hard to deny that contemporary philosophical theorizing about the nature and conditions of moral responsibility has been deeply influenced by religious thought. In some cases, the influence is overt and an explicit feature of the theorizing. In others, however, the influence may be harder to discern. The primary purpose of this chapter is to explore some central elements of this influence. In order to have a manageable frame of inquiry, this chapter will limit its attention to the influence of traditional theistic commitment upon the contemporary debate, where we will take traditional theism to be the view that there exists a maximally powerful, knowledgeable, and benevolent creator and sustainer of the universe. The limitations imposed here are substantial; they establish boundaries that are at once overly narrow and overly broad. The boundaries are overly narrow because some religious traditions that could plausibly be thought to have influenced contemporary philosophical reflection on moral responsibility are not theistic in the present terms. They are overly broad because there are a great many differences among alternative theistic religious traditions—​between Judaism, Christianity, and Islam, for example—​that could be relevant to theorizing about moral responsibility. Nevertheless, we can hope that exploration of this influential, though admittedly limited, terrain will be intrinsically worthwhile and an encouragement to further related work of both coarser and finer grain. One way to approach this subject matter would be by way of constructing a straightforward list of ways that theistic thought and practice have had philosophical effects on contemporary theorizing about moral responsibility. Here we would be working, as it were, from the ground up: from cases in which we can see the influence, but without any particular insight into what might be unifying or explaining this influence. My strategy, however, will be different; and, I hope, somewhat more illuminating. I begin by asking what it is, more generally, about theistic commitment that would seem to impose

626   Daniel Speak pressure on our attitudes about moral responsibility. That is, we start by considering what aspects of theism, taken more or less in the abstract, could be expected to make a difference at the level of our concerns with moral responsibility. From this point of view, I submit, there is some additional explanatory power to be had; though of course to access it we will still need to think carefully about the concrete cases and about the nature of the relevant influence. Why, then, is theistic commitment likely to impose unique demands on our theorizing about moral responsibility? I will be suggesting five general answers: (1) because theism has emphasized human likeness to God, (2) because theism proposes divine purposes for our agency, (3) because theism invites eternal expectations for our moral lives, (4) because theism finds itself with internal tensions that it is under philosophical pressure to resolve, and (5) because theism faces relevant external challenges to its philosophical coherence and plausibility to which it is at pains to respond. While I hope that moving through these points of pressure in this order has a certain natural reasonability, it is also worth noting that this ordering is also, generally speaking, from the more abstract and speculative influence of theism on theorizing about moral responsibility to more concrete and overt impact.

1.  Human Likeness to God According to theism, the universe has been created by, and is sustained by, omnicompetent agency. Divine agency is thus naturally conceived of as agency par excellence—​as agency at its conceptual best. For the theist, then, such agency is not merely an intriguing conceptual possibility to be considered in the abstract. It is, rather, the fundamental fact grounding all of reality. That is, according to theism ideal agency exists, and its effects are all around us in the created order. Just what is entailed by the theistic ideal of divine agency is complex and controversial, of course, with a great many competing theistic accounts vying for theological market share. Here the point is just that the theist will not be especially tempted by accounts of the nature of agency simpliciter that are essentially deflationary: for example, according to which agency is an accidental feature of the universe or is necessarily premised upon limitation, whether metaphysical or moral. Nor will the theist be particularly attracted to views that treat the concept of agency as one that does not carve nature at a fundamental metaphysical joint. Since theists accept that God exists and is an ideal agent, the concept of agency will enjoy a privileged status in theistic ontology. Furthermore, theists have typically accepted some version of the Imago Dei doctrine. In the Hebrew Scriptures, and particularly in the Book of Genesis, human beings are said to have been created in the “image” or “likeness” of God.1 Once again, the specifics

1 

See Genesis 1:26–​27 and 5:1–​2.

Responsibility and Religion    627 here have varied considerably over time and tradition. Some theists have insisted that the image of God reflected in the human person is substantial; indeed, so substantial that we can expect to gain insight into philosophical anthropology by, for example, seeking to understand our moral psychology on the model of the Christian trinity.2 Other theists, however, have been content to think of our likeness to the divine principally in relational or functional terms. In almost all cases, though, the Imago Dei doctrine has been taken to suggest important parallels between human and divine agency. Our actions are, in crucial respects, like God’s, and especially like God’s action in creation. For theism, then, divine agency characteristically functions as a model of human agency—​or, if not as a model, then at least as a foil for considerations of the nature and limits of its human instantiation. The consequences of this divine modeling conception of human agency for theistic accounts of moral responsibility are not, by any means, straightforward. Still, there are two points worth noting here. First, conceptions of the divine ideal, together with the emphasis on human beings as created in the likeness of God, have naturally pushed theistic thinkers toward what I will call “optimistic” views of human agency. As I am thinking of the matter, optimistic views are those characterized by higher hopes and expectations for human agential capacity and achievement. This is to say that such views are optimistic about the prospects of our meeting more rigoristic agential demands. Theistic agential optimism can be seen to take two broad forms. One form involves a tendency to be optimistic about the original natural powers and reach of human agency. Given that human agency is relevantly like God’s, we can expect, at least under normal conditions, to have naturally robust agential abilities: for example, to reason well, to perceive the moral landscape accurately, and to bring our practical reasoning to fruition in coherent action.3 A second form of optimism has to do with the capacities of human agency in the face of inevitable obstacles. Consider our agential capacities for avoiding cognitive bias in our practical reasoning, for overcoming akrasia, or for mitigating the effects on our values of a deficient upbringing. Again, just as theism envisions God as essentially unaffected by the appearance of obstacles to the achievement of divine purposes, theism will also incline those under its influence to view the human agent as able to manage volitional obstructions without substantial cost. Optimistic views will be those, then, that emphasize our having what it takes, both in terms of our basic powers and in terms of our abilities to overcome obstacles, to succeed in agential efforts of the sort clearly crucial 2 Augustine, De Trinitatae. 3 

As a counterpoint, theistic traditions have also placed considerable emphasis on human sinfulness or “fallenness.” On a number of classical theistic views of the human fall from divine grace, the original capacities for rationality and moral action emphasized in the body of the text are seriously damaged by human sin. This is to take note of the fact that there are impulses within theism to a potentially counterbalancing form of pessimism. Note, though, that this pessimism is not about our basic agential endowment but about what we are left with after the sinful damage has been done. How this affects our theorizing about moral responsibility will partly depend on whether our theorizing on this topic purports to be of the ideal or non-​ideal variety.

628   Daniel Speak for morally responsible agency. In one sense, this leads to another species of the point that theists will be naturally pulled away from skeptical or nihilistic accounts of moral responsibility, especially when the skepticism is grounded in doubts about human agential capacities. This isn’t to say that there are no pressures coming from theism that push philosophers and theologians toward moral responsibility skepticism; there are indeed such pressures, as we will see. Still, there is clearly some foundational pressure here away from such skepticism. A related second point has to do with the way that the Imago Dei impulse invites those influenced by theism to think of human agential powers in substantially individualistic terms. On the standard picture, nothing outside of God is necessary for divine rational action. God does not need external support in order to act, and nothing outside the divine will can contribute to its rationality. When God acts, it is all and only the divine volition that explains the action. Thus, when theists emphasize the human parallels with God on this score, they frequently contribute to another form of human agential optimism—​an optimism regarding our ability to act, so to speak, independently of the social structures in which we find ourselves. This is to say that our likeness to God has frequently been taken to justify confidence in a form of agential autonomy that we should expect to affect an account of moral responsibility. From this point of view, it is not surprising to find the late (or indeed, absence of) emergence, within the theistic traditions of concerns for the way that concrete social and political realities impact individual moral responsibility. Supposing ourselves to possess a God-​like autonomy, we might very well have good grounds for ignoring the potential for circumstantial effects on our agential capacities. In a similar spirit, the fairly radical Kantian thought—​that our moral responsibility depends on our ability to have chosen our characters and our actions from outside of nature and time—​would be hard to take seriously apart from a view about our likeness to God. In an even more straightforward instance of the central thought, recall Chisholm’s surprising claim in his effort to resolve the compatibility problem: “If we are responsible, and if what I have been trying to say is true, then we have a prerogative which some would attribute only to God: each of us, when we act, is a prime mover unmoved.”4 The pressure here is, of course, not directly logical, such that an emphasis on human likeness to God rules out an appeal to the sociality of our agency. Nevertheless, it no doubt partially explains how little attention the Western tradition, especially as it has been influenced by theism, has paid to the dependence of responsible human agency upon various felicitous and contingent circumstances.5

4 

“Human Freedom and the Self,” p. 12 Further implications of the Imago Dei doctrine for human moral responsibility will depend a good deal on how the theist (or the brand of theism) in question conceives of divine freedom and, if such a thing is coherent, divine moral responsibility. For consideration of these questions, see William Rowe, Can God Be Free? (2004) and William Alston, “Some Suggestions for Divine Command Theorists” (1990). 5 

Responsibility and Religion    629

2.  Divine Purposes Another core feature of traditional theism that will have implications for our theorizing about moral responsibility involves the supposed divine aims for human agency. For the theist, the universe is not a cosmic accident; and neither are its constituent parts, including human persons and their capacities. Here it might be helpful to contrast the theistic picture with a compelling naturalistic alternative. On the form of naturalism I have in mind, human beings are indeed cosmically accidental, emerging from the sturm und drang of essentially random chemical and biological evolutionary processes that make no pretense of intentionality. As a consequence, the capacities with which we find ourselves will be explained not by their requirement for some larger purpose for human existence but by their radically contingent good fortune in supervening upon whatever has contributed sufficiently to our Darwinian fitness. I trust the contrast here can be seen to be quite sharp. The theist (who need not deny the crucial role of broadly Darwinian mechanisms in speciation) will conceive of human agency as the consequence of an intentional process with a telos, while the naturalist we are imagining will not. The implications for moral responsibility of a divinely purposed conception of human agency are substantial. I will consider only two here. First, for those who take the purposive picture seriously, it will be sensible to be suspicious of views about responsible agency that treat it as something of a moral spandrel. On some contemporary accounts, the aim of our theorizing about moral responsibility is to make the best of the accidentally crooked timber with which we find ourselves. On such accounts, we shouldn’t expect our capacities to match up with our highest moral expectations or to have any larger teleological aims. Consider, on this count, the broad revisionist approach according to which what we should seek to achieve in developing a theory of moral responsibility is a tolerable vindication of the standard practices of praising and blaming—​and punishing and rewarding—​that our social milieu has already accepted.6 Or consider approaches that emphasize the social construction of moral responsibility or simplistic forward-​looking (i.e., consequentialist) justifications for our responsibility practices. These are natural (and perhaps even salutary) ways of theorizing about responsibility from the point of view of cosmic accidentality. However, those influenced by the theistic picture are unlikely to be moved by these methodological impulses since what lies beneath them appears to be a preparedness to treat our agential capacities as a lucky result of purposeless mechanisms. This first point was made without appeal to any specific account of God’s purposes for morally responsible agency. But we can see a second further implication of theistic influence once a specific claim about the divine purpose for human agency is on the table. Suppose, for example, and following many theistic traditions, that God intends human beings to be capable of a morally serious and personally intimate divine-​human 6 

See Vargas, Building Better Beings, especially c­ hapter 6.

630   Daniel Speak relationship. Clearly, intimacy is a graded concept, so it should be taken to admit of degrees; in consequence, there will be much variation among theists regarding the nature and depth of the intimacy that God might seek with creatures like us. In any case, taking the minimal model of our best human relationships, it will be natural for theists to think of this intimacy as requiring a good deal of our agential capacities. After all, anything even recognizably resembling reciprocity with God will demand that human beings are generally capable (or can become capable, consistent with constraints of personal identity) of the forms of mutual regard we might expect to find within a community of our most admired moral exemplars. Since such relationships will involve fundamental vulnerability between participants, each must be able to trust the other, each must be able to have an abiding confidence in the good will and good character of the other. This means that each one will have to be deeply morally sensitive, sharply self-​aware, and highly self-​controlled. The point is not that all human agents must already meet these high-​powered moral demands for intimate relationship but rather that they can reasonably be taken to have what it takes to meet these demands. At the very least, then, this particular purposive account of human agency is likely to lead to a relatively high view of our agential capacities, since the aim of a kind of friendship with God is a consistent animating image. As a result, we can further expect theists and those influenced by theism to have particularly elevated views of our basic capacities, with correspondingly rigoristic conceptions of moral responsibility. Put another way, with divine purposes like this one in the background of our theorizing, we can expect theistic theories generally to make it harder to avoid blame and harder to earn praise, since they suggest that human beings are perhaps better endowed with respect to agency than they can sometimes appear to be.

3.  Eternal Expectations A related conceptual pressure can be seen to be applied by the dramatic eschatological expectations of most forms of traditional theism. From the standpoint of the theistic traditions, human beings are typically taken to have an eternal future. That is, most forms of theism expect each human being to end up either in heaven or in hell, where these states are taken to be maximally honorific and maximally punitive, respectively. Thus, theism ordinarily proposes an ultimate reckoning of our moral lives that will in the end justify eternal punishments and rewards. Clearly this theistic picture of human destiny will have powerful implications for a theory of moral responsibility. The thought that our moral lives have eternal stakes is very likely to encourage, once again, a robustly optimistic view of our capacities as moral agents. Insofar as the ideas of heaven and hell—​as states of normative judgment—​are treated as morally and metaphysically coherent, a correspondingly elevated conception of the morally responsible agent will be required in order for the ultimate divine judgments to meet conditions of fairness or justice.

Responsibility and Religion    631 What might initially appear as a surprising upshot of this influence has been in the direction of moral responsibility skepticism. In Galen Strawson’s development of what he calls “the basic argument” to the conclusion that no one is truly morally responsible, the ideas of heaven and hell play a crucial role.7 The heart of the argument is that true moral responsibility requires that an agent be the fundamental source of its own capacities; it requires the agent, in short, to be a causa sui, the cause of itself. But, of course, no human being can be a causa sui. Therefore, Strawson concludes, no human being can be truly morally responsible. Why, however, does he think that true moral responsibility demands anything so exotic as self-​causation? It is because he believes that the idea of true moral responsibility is a demanding one indeed. In an effort to explain what he takes to be the core idea of this true moral responsibility, Strawson tells us that, [a]‌n old story is very helpful in clarifying this question. This is the story of heaven and hell. As I understand it, true moral responsibility is responsibility of such a kind that, if we have it, then it makes sense, at least, to suppose that it could be just to punish some of us with (eternal) torment in hell and reward others with (eternal) bliss in heaven.8

Strawson is quick to add that one need not believe that this story is true to see his point. This is because his point purports to be about what would need to make sense in order for one to be truly morally responsible. If someone is truly morally responsible, then it would have to make sense that she could be justly condemned to hell or consigned to heaven on the basis of her actions. With this demand in mind, it is certainly easier to see how Strawson can plausibly claim that true moral responsibility requires that an agent be the ultimate source of the springs of her actions. With eternal punishments and rewards on the line, it will be reasonable to expect that targeted agents have complete control over themselves, since anything less will entail an unacceptable amplification of moral luck. If the agent is not a causa sui, then infinite judgment will end up turning on factors beyond the agent’s control in a way that will be transparently unfair. This isn’t the place to evaluate Strawson’s argument.9 What I want to take notice of here is only the powerful influence of a theistic perspective on this brand of argument regarding moral responsibility. Though Strawson’s conclusion is ultimately in favor of skepticism about moral responsibility, his approach vindicates rather than undermines the point with which we began this section. The point, you will recall, was that the ideas of heaven and hell that are central to many forms of traditional theism animate a particularly exalted view of the morally responsible agent. What Strawson’s argument allows

7 

“The Impossibility of Moral Responsibility,” Philosophical Studies, vol. 75, 1994) “The Impossibility of Moral Responsibility,” 9. 9  For evaluation, however, see; Hurley, 2000, “Is Responsibility Essentially Impossible?” Philosophical Studies 99, 229–​268; and Clarke, 2005, “On an Argument for the Impossibility of Moral Responsibility.” Midwest Studies in Philosophy 29, 13–​24. 8 

632   Daniel Speak us to see is that this exaltation can drive theory in two different directions. If we are convinced that we are in fact morally responsible agents, then the exaltation will accrue to us as agents who can justly be targeted with eternal punishment and rewards. If we are convinced on independent grounds that we are not the kinds of agents who can be justly targeted with eternal punishments and rewards, then the exalted view of morally responsible agency will drive us toward skepticism regarding true human moral responsibility.10

4.  Internal Pressures Another way to see into the influence that theism is prone to have on theorizing about moral responsibility is to consider relevant internal tensions that theistic religions have sought to relieve. In point of fact, there are too many of these kinds of tensions internal to theism to treat them all here. I will be content to look in some detail at two of the more influential problems that implicate concerns about moral responsibility, both generated by standard views about God’s relationship to the created world.

4.1  Divine Foreknowledge Given that traditional theism is partly premised on the idea that God is omniscient, the problem of divine foreknowledge arises almost immediately. After all, it certainly seems natural to think that an omniscient being has exhaustive foreknowledge, where this will amount to God’s having infallible knowledge about every future event, including those events that constitute human actions. The problem, then, put tersely, is how human freedom and responsibility is supposed to be compatible with this kind of exhaustive foreknowledge. To make the problem particularly clear, we can briefly consider Nelson Pike’s influential regimentation of it.11 Bringing together various strands of thought—​from Augustine, Anselm, Paley, and especially Boethius—​Pike presents the foreknowledge problem as something like a trilemma. Suppose that someone has performed some action at a particular time; let’s say that Dana has raised her hand in support of a certain departmental policy at noon on October 15, 2018. Now also assume that God exists and has exhaustive foreknowledge.

10 

Of course, other options include (1) rejecting Strawson’s claim that heaven and hell responsibility (“true” responsibility) is what we have all along been concerned with in the moral responsibility debate; and (2) rejecting Strawson’s claim that “true” moral responsibility must, at least in principle, be such as to make sense of justified eternal punishment or reward. But, again, the larger point here is that there are theistic impulses that make opting for either of these somewhat undesirable. 11  Pike, Nelson, 1965, “Divine Omniscience and Voluntary Action,” The Philosophical Review 74(1): 27–​46.

Responsibility and Religion    633 This will mean, Pike claims, that God has known for a very long time, indeed always, that Dana would raise her hand at noon on the day in question. We can assume, for example, that God knew 10,000 years ago that Dana would raise her hand at noon on this day. Now, we can ask whether, given God’s knowledge 10,000 years ago, Dana could have refrained from raising her hand at the time she did. Pike claims that in order for it to be true that Dana could have so refrained, one of three things would have to be true. One thing that could be true is that Dana had an ability, at noon, to do something such that in doing it she would have changed the past. In particular, she could have had the power to make it that, while God knew (and therefore believed) 10,000 years ago that Dana would raise her hand at noon, the past is changed so that God did not in fact believe this 10,000 years ago. That is, perhaps Dana’s action could have changed the past. Another possibility is that Dana had an ability, at noon, to do something such that in doing it she would have caused God to be wrong 10,000 years ago in believing that she would raise her hand at noon on the future day. That is, perhaps Dana’s action could have made God wrong. A final possibility, by Pike’s lights, is that Dana had an ability, at noon, to do something that would have caused God to cease to exist 10,000 years ago. That is, perhaps Dana’s action could have made God nonexistent (and if God does not exist back 10,000 years ago, then there is no divine knowledge at that time to restrict Dana’s freedom now). Of course, Pike also contends quite reasonably that each of these three potential powers of Dana’s are impossible. No one can cause the fixed past to be different from the way it actually is. No one can cause an essentially omniscient being to have a false belief. And no one can cause a necessary being to cease to exist. However, Pike’s claim is that the only way for Dana to be able to refrain from raising her hand, given God’s foreknowledge, is for her to have one or another of these impossible powers. Therefore, he concludes, God’s foreknowledge rules out her ability to have refrained from raising her hand at noon. And, of course, the point here generalizes to all human actions that an omniscient being foreknows. Since Pike insists that having the ability to refrain from an action is constitutive of its being a voluntary action, his big conclusion is that divine foreknowledge undermines all voluntary action; if God has exhaustive foreknowledge, then no one acts freely. In addition, on the assumption (nearly ubiquitous in the Western tradition) that acting freely is a necessary condition for being morally responsible, divine omniscience appears to be in fundamental tension with the view that human beings are morally responsible agents. Again, tensions of the kind enshrined in Pike’s argument have animated theistic reflection for two millennia, even if his crystallization of the issues is decidedly contemporary. What Pike’s version of the problem brings out crisply are the decision points for theism regarding divine omniscience and human moral responsibility. Clearly we are not in position to assess the full range of theistic responses to these tensions here.12 Nor is it clear that considering the full range would help our present inquiry. However, it can

12 

But for an influential assessment (and treatment of many of the details elided here), see The Dilemma of Freedom and Foreknowledge by Linda Zagzebski (Oxford, 1991).

634   Daniel Speak be illuminating to see a couple of the decision points that transparently affect theorizing about moral responsibility. The first point of decision we can consider regards how committed the theist should be to the view that human beings are indeed morally responsible. After all, it would seem that a simple resolution of the tensions generated by the foreknowledge problem might take the form of giving up the view that human beings are in fact free and responsible. It is worth noticing that a substantial majority of the theistic responses to the dilemma have, however, aimed to vindicate the view that human freedom and responsibility are compatible with divine foreknowledge.13 Put another way, one point to take away from the foreknowledge controversy is that the theistic tradition has treated the claim that human beings are morally responsible as nearly axiomatic. Though I am making no effort to explain this nearly axiomatic commitment to human moral responsibility here, it is one more point at which we can see a brand of theistic inoculation against responsibility skepticism. A second point of decision involves the relationship between the claim that God has exhaustive foreknowledge and the claim that determinism is true. Taking notice of this point is crucial because the contemporary debate about moral responsibility has been so thoroughly structured by concerns regarding determinism. Furthermore, there are extremely suggestive parallels between the form of the foreknowledge problem—​ especially as Pike formulated it—​and the form of consequence argument for the incompatibility of determinism and moral responsibility developed by Peter van Inwagen.14 At this decision point, we can identify three broad impulses. The first is to see a deep conceptual relationship between foreknowledge and determinism while at the same time having considerable confidence in the compatibility of responsibility with divine foreknowledge. The output of this way of seeing the terrain will be an inclination toward a form of compatibilism also about determinism and responsibility. The second broad impulse also sees a deep conceptual relationship between divine foreknowledge and determinism; but added to this is a strong conviction, perhaps generated by consequence-​ style reasoning or by a manipulation argument, of the incompatibility of determinism

13  One particularly relevant contemporary divergence from this majority takes the form of what has come to be called “Open Theism.” On this view, the preferred way for the theist to respond to Pike’s argument is to give up the claim that God has exhaustive foreknowledge. Open theists do not, however, abandon the claim that God is omniscient. Rather, they seek to provide an account of omniscience that does not include exhaustive foreknowledge. See Tuggy, Dale, 2007, “Three Roads to Open Theism,” Faith and Philosophy 24(1): 28–​51. 14  In (very) brief, the consequence argument supports incompatibilism by attempting to demonstrate that, given determinism, a person could do something other than what she has done only by having one of two impossible powers: either the power to change the past or the power to change the laws of nature. See An Essay On Free Will by Peter van Inwagen (Oxford, 1983). John Martin Fischer has been especially adept at recognizing the parallels in the forms of concern about freedom and responsibility generated by both divine foreknowledge and determinism. For some details, see his The Metaphysics of Free Will (Oxford, 1994). It may also be profitably noted that the foreknowledge dilemma bears at least equally striking similarities to the classical fatalistic sea battle argument that vexed Aristotle.

Responsibility and Religion    635 and moral responsibility. Here the upshot is likely to be a higher degree of doubt about the compatibility of divine foreknowledge and responsibility. The third broad impulse is to resist claims about the depth of the conceptual relationship between foreknowledge and determinism, emphasizing that there may be very important differences, relevant to assessments of moral responsibility, regarding the ways that each of them could threaten to undermine our responsibility-​relevant freedom. To the degree that these differences can be shown to be substantial, there is room to endorse compatibilism about foreknowledge and moral responsibility while also endorsing the incompatibility of determinism and moral responsibility.

4.2  Divine Providence Traditional theism has also typically thought of God as providential. A providential God does not merely have knowledge about the goings-​on in the created order but exerts some level of control over them. In fact, the standard view has been that this control is, in some sense, quite thorough. For example, Tom Flint takes note of the expression of providence in the Westminster Confession of 1647: God, the great Creator of all things, doth uphold, direct, dispose, and govern all creatures, actions and things, from the greatest even to the least, by his most wise and holy providence, according to his infallible foreknowledge, and the free and immutable counsel of his own will, to the praise and the glory of his wisdom, power, justice, goodness, and mercy.15

The emphasis, then, for the providential view of God is upon divine governance. And, as this traditional expression highlights, this governance is taken to go beyond foreknowledge (though it depends upon this foreknowledge) and extends to all aspects of the created order, including human actions. Understood in its strongest form, what we can call, following the tradition, “meticulous” providence is the view that every event occurs precisely as God has planned; no events occurs contrary to God’s overarching and maximally good intentions. There have, of course, been weaker views of divine providence than this one. For example, there have been more general accounts of God’s providence that are premised upon the idea that God exerts a broad and somewhat unspecific control over the way things unfold in the created order but without meticulous control over each detail. In any case, what theism has bequeathed to its progeny here is an even more pointed challenge than the foreknowledge problem discussed earlier. That the providence problem is indeed more pointed can be seen by noting that foreknowledge can, at least in principle, be taken to have no causal impact on human activity, even

15 

Divine Providence: The Molinist Account, 15. Flint credits this quote to Paul Helm, The Providence of God (Downers Grove: InterVarsity Press, 1994)

636   Daniel Speak if it does (following Pike) entail the absence of alternative possibilities for action. It is at least prima facie difficult to see how divine governance over human actions can be executed without causal impact. This is to say that the divine control implied by doctrines of providence appears to be a greater threat to moral responsibility than does divine foreknowledge because control suggests even stronger parallels to manipulation than does simple knowledge. Still, we should take note of the fact that traditional theism has generally been at pains (as with the foreknowledge problem) to render a compatibility verdict with respect to providence and human responsibility. Again, that human beings are morally responsible has continued to operate as a kind of axiom, forcing theists to reconcile it with metaphysical threats to it that could somewhat easily be parried by abandoning the near axiom. Given theistic commitment both to the moral responsibility of human beings and to the highest views of divine knowledge and control, we can find in this corner of the theistic tradition some strong (though resistible) pressures toward broadly compatibilist conceptions of moral responsibility.

5.  External Challenges A final family of pressures on our theories of moral responsibility that we can consider here arise from the influential philosophical challenges that theism has had to face, particularly in the form of the problem of evil and its sister problem, the so-​called problem of divine hiddenness.

5.1  The Problem of Evil To be frank, there may be no greater pressure placed upon theistic theorizing about human agency than that imposed by the contemporary problem of evil. If this is right, it is largely because the philosophical challenge strikes so deeply at the core of theistic coherence and plausibility. Some version of the problem of evil has been on the table for theists from the beginning; but most versions of the problem throughout intellectual history have taken the form of an explanatory aporia. Like the problems of divine foreknowledge and divine providence, accommodating the existence of evil in the world largely amounted to a reconciliation project internal to theism. This is to say, the problem of evil was not generally understood to constitute a threat to the truth of the claim that God exists, though, of course, there were some pointed critics who edged up to atheistic argumentation. In the 20th century, however, the contemporary problem of evil became a transparently polemical one, constituting a clear and forceful attack on the plausibility of theism. Given the stakes in the contemporary debate, then, it is understandable that the polemical problem of evil has exerted so much pressure on recent theorizing about moral responsibility.

Responsibility and Religion    637 The contemporary polemical problem of evil gets perhaps its most pointed and influential expression in J. L. Mackie’s rightly famous 1955 paper, “Evil and Omnipotence.”16 In it, Mackie attempted to demonstrate “not that religious beliefs lack rational support, but that they are positively irrational, that the several parts of the essential theological doctrine are inconsistent with one another [ . . . ].”17 In brief, then, Mackie took his argument to show that the theist could be shown to be involved in a form of fairly straightforward logical self-​contradiction. That is, Mackie’s argument purported to establish that, by virtue of accepting that there are indeed evils in the world, logical consistency would require theists to give up one or more of their essential commitments. The argument itself is, in fact, deceptively simple. It can be understood as the claim that the collection of propositions endorsed by the theist constitutes a logically inconsistent set. To see the proferred inconsistency, here are the propositions to which Mackie took theists to be essentially committed:

1. 2. 3. 4.

God exists God is omnipotent God is omnibenevolent Evil exists

His strategy, then, is to draw out an inconsistency among these commitments. He argues that if God exists and is omnipotent, then God would be able to create a world lacking evil. At the same time, if God exists and is omnibenevolent, then God would be fundamentally motivated to create a world lacking evil. According to Mackie, this means that, given the truth of the first three propositions, the fourth proposition would have to be false. However, since the fourth proposition is clearly true (and an essential commitment of theism to boot), at least one of the first three must be false. But if any one of these propositions is false, then theism is false. Therefore, we should conclude that theism is false. And notice, again, that the conclusion is not that theism is probably false or unsupported by good evidence; it is that theism must be false by virtue of a formal inconsistency among its essential commitments. One thing to see here, then, is that the stakes for the theist in responding to this argument could not be much higher. Without some compelling reply, theism stood to be incoherent. What I want to emphasize in what follows is that the two most influential initial theistic responses to Mackie’s argument depended upon a substantial and controversial view about human freedom and moral responsibility. In the first instance, Alvin Plantinga developed his “free will defense” in response to Mackie’s argument.18 As Plantinga conceived of a “defense,” it was an effort, in short, only to show that an argument for atheism, like Mackie’s, had not made its formal case. 16 

J. L. Mackie, 1955, “Evil and Omnipotence,” Mind, 64: 200–​212. Mackie, “Evil and Omnipotence,” 200. 18  See chapter IX of his The Nature of Necessity (Oxford, 1974) and, somewhat more accessibly his God, Freedom, and Evil (Eerdmans, 1977). 17 

638   Daniel Speak The distinction to be drawn is a with a “theodicy,” which purports to show what may reasonably be considered to be God’s good reasons for allowing evil. Plantinga’s defense does not take up the project of theodicy; it makes no claims, he insists, about what God’s reasons for allowing evil should be thought to be. Rather, Plantinga attempts to demonstrate that the Mackie argument falls short, so to speak, on its own terms. Put otherwise, Plantinga takes his defense to show that the theist is under no philosophical pressure from the Mackie argument because it does not in fact demonstrate the logical inconsistency it so colorfully describes. What is crucial for our purposes are two points here, one philosophical and one sociological (or near enough). The philosophical point is that Plantinga’s defense turned crucially on at least the coherence of libertarianism regarding free will and moral responsibility.19 Given his defensive purposes, all Plantinga took himself to be in need of showing was that the theist could accept the four propositions above without being involved in logical incoherence. To do this, Plantinga took note of the fact that the propositions could all be true if God had some good reason for allowing the world’s evil. Then, all he needed was the plausible claim that it was merely metaphysically or logically possible that God have a good reason. To animate this possibility, Plantinga supposed, for the sake of the argument, that God might want human beings to enjoy a form of freedom that could ground deep moral regard and the consequent possibility of divine intimacy. He further supposed that such deep moral regard could not be enjoyed by agents who were determined by antecedent factors to perform their actions. That is, he supposed the truth of incompatibilism with respect to deep freedom and responsibility. With these suppositions in place, it wasn’t hard to see how the four theistic propositions could all be true together. After all, if God had a good reason for allowing robust free and responsible human agency, then the occurrences of evil resulting from poor exercises of this agency on the part of human beings could not be taken to be inconsistent either with God’s power or with God’s goodness. Importantly, the point was not that libertarian agency in fact justifies God in allowing the world’s evils. Instead, the point was that the logical possibility of the story told by Plantinga regarding freedom, responsibility, and evil demonstrated that Mackie was wrong that there is a formal contradiction lurking within the propositions endorsed by the theist. The sociological point is that Plantinga’s defense came to be widely understood to be decisive against Mackie’s argument and similar instantiations of the logical problem of evil. At the very least, it became part of the standard lore of 20th-​century philosophy of religion that Plantinga’s free will defense had silenced the Mackie objection.20 I take 19  In his original argument, Mackie somewhat vaguely suggests that the libertarian conception of agency is itself incoherent, though he makes no effort to establish this. He goes on instead to resist the appeal to a free will response to his argument by claiming, in essence, that, given compatibilism, God should have been able to create a world in which human beings are free and yet always choose well. In an important respect, then, Mackie fails to see that his purported aim (to demonstrate the incoherence of theism) requires the stronger view at which he has merely gestured, namely, that there is something logically problematic about the libertarian view. Plantinga’s defense exploits this point. 20  In minimal support of this sociological claim is William Rowe’s concession, setting up his own inductive version of the atheistic argument from evil: “Some philosophers have contended that the

Responsibility and Religion    639 note of this sociological point because we can be sure that this perception of the import and success of Plantinga’s argument has played an important role in continuing to attract theistically inclined philosophers to libertarianism regarding freedom and moral responsibility. Around about the same time that Plantinga was developing his free will defense, John Hick was constructing the most influential theodicy of the late 20th century. In his Evil and the God of Love, Hick attempted to provide a set of morally sufficient good reasons that God may have had for allowing the world’s evils.21 The central claim of the theodicy was that reasons sufficient to justify God’s permissions of evil could be found in the domain of what Hick came famously to describe as “soul-​making.” On the assumption that God is aiming to bring human beings through a form of moral development that could ultimately lead them into a finally satisfying intimacy with God, Hick argued that human beings would need to overcome serious obstacles and engage in robust character transformation. Here again, the controversial libertarian view about freedom and responsibility takes center stage. Intriguingly, according to Hick, this isn’t because human beings could not, in principle be determined in the soul-​making process. Instead, his concern is more axiological than metaphysical. He claims that “one who has attained goodness by meeting and eventually mastering temptations, and thus by rightly making responsible choices in concrete situations, is good in a richer and more valuable sense than would be one created ab initio in a state of either innocence or virtue.”22 Furthermore, he insists that the soul-​making process “is not taking place . . . by a natural and inevitable evolution, but through a hazardous adventure in individual freedom.”23 To be clear, Hick explicitly insists that this adventure in individual freedom is of the libertarian variety. And though he is aware of the difficulty of defending the view that human beings enjoy this brand of responsible agency, he does insist that “it is the one that seems intuitively most adequate to our ordinary experience as moral agents.”24 In the time since the seminal contributions of Plantinga and Hick, theistic philosophers have sought to respond to the problem of evil in ways that do not presuppose a libertarian conception of responsible agency. In fact, a number of philosophers have explicitly attempted to develop, or at any rate take note of the possibility of

existence of evil is logically inconsistent with the existence of the theistic God. No one, I think, has succeeded in establishing such an extravagant claim. Indeed, granted incompatibilism, there is a fairly compelling argument [Plantinga’s] for the view that the existence of evil is logically consistent with the existence of the theistic God” (Rowe, 1996, p. 10, note 1). Of course, there have been philosophers unwilling to sign on with the majority regarding the success of Plantinga’s argument, including DeRose 1991; Adams 1999; Howard-​Snyder and O’Leary-​Hawthorne 1998; Otte 2009; Pruss 2012. It may be worth noting here that in almost all of these cases, the philosophers criticizing Plantinga’s argument also go on to argue that local changes to and adaptations of Plantinga’s reasoning do succeed in showing that the logical argument from evil is not finally probative. 21 

Hick 1978. Hick 1978, 255. 23  Hick 1978, 256. 24  Hick 1978, 278. 22 

640   Daniel Speak developing, compatibilist versions of the free will defense.25 Some have even suggested that a successful defense can be launched from the platform assumption of responsibility skepticism.26 Still, the larger point is that the most influential theistic responses to the problem of evil have centralized libertarianism.27 This is a point of pressure from theism on theories of moral responsibility that it would be hard to overstate.

5.2  The Problem of Divine Hiddenness Perhaps the most recent external philosophical challenge to theism has come from the problem of divine hiddenness, due to J. L. Schellenberg.28 In his original presentation of this argument for atheism, the hiddenness of God is constituted by the fact that there are fair-​minded and morally sensitive people who do not take themselves to have enough evidence to believe that the God of traditional theism exists. As Schellenberg argues, if there were indeed a being like the God of traditional theism, there would be no such people. This is because God would be sure that anyone who is indeed not irrationally resistant would be able to be in a relevant relationship with God. And, according to Schellenberg, to be in a relevant relationship with God requires that one believe in God. Therefore, out of divine love and goodness, God would be sure to make every reasonable and nonresistant person also capable of believing that God in fact exists. That is, God would not remain hidden from such people. However, since there do seem to be reasonable and nonresistant nonbelievers, we should conclude that there probably is no maximally powerful and maximally good being. One thing to notice about this argument is that Schellenberg is careful to frame it in such a way that it doesn’t transparently require the falsity of libertarianism about moral agency. This point comes out in the way he expresses the idea of hiddenness. The simple fact that there are people who do not believe in God does not constitute divine hiddenness as he is thinking of it. After all, he allows, it could very well be that many of those who do not believe that God exists are in this state of unbelief by virtue of various bad free choices on their own part. Any such instances of unbelief will not entail hiddenness. Thus, the hiddenness argument cannot obviously be addressed simply by appeal to libertarianism, since it is only those persons who are not using their libertarian agency badly in their attitudes toward the proposition that God exists who count toward the hiddenness claim.

25 

Bishop 1993; Howesepian 2007; Turner 2013. See Derk Pereboom, “Libertarianism and Theological Determinism,” in Free Will and Theism, eds. Kevin Timpe and Daniel Speak. (Oxford, 2016). 27  In addition, libertarian moral agency remains central to most recent theistic responses to the problem; for example in Richard Swinburne’s Providence and the Problem of Evil (Oxford, 1998); in Peter van Inwagen’s The Problem of Evil (Oxford, 2006); and in Eleonore Stump’s Wandering in Darkness (Oxford, 2012). 28  Divine Hiddenness and Human Reason, Ithaca, NY: Cornell University Press, 1993. 26 

Responsibility and Religion    641 As with the other problems for theism that we have considered in this chapter, we have no space for a full consideration of the theistic responses to the hiddenness argument.29 What is worth seeing, however, is once again the way that important responses to it have turned upon views that are likely to influence theorizing about the nature of moral responsibility. For example, one important response to Schellenberg’s argument has attempted to explain the existence of reasonable unbelief by way of a concern with epistemic coercion.30 The animating thought is that what God would have to do in order to remove all reasonable unbelief might make the divine reality too palpable for our moral good. That is, in providing sufficiently strong evidence, God might thereby make some important forms of moral action impossible. What this response brings out is the possibility that our morally responsible agency can be threatened epistemically. As Michael Murray develops this idea, it is a species of a Hick-​style soul-​making concern, since his claim is that in order for God to allow us to develop in the deepest moral ways God must not provide too much evidence of divine reality. Once again, then, we can see a way in which a theistic outlook, motivated by the demand to respond to an external challenge, will have substantial implications for a view of human responsibility. This is because a response of this kind to the hiddenness argument will depend upon a view of our moral responsibility that makes it subject to the forms of epistemic coercion about which Murray, among others, expresses concern. Another line of reply to Schellenberg’s hiddenness argument might put emphasis on the collective nature of our epistemic responsibility. I find philosophically attractive the idea that in the same way that we can be collectively morally responsible for commons tragedies like environmental degradation, we can also be collectively responsible for epistemic commons tragedies. This is to say that God may plausibly be thought to have an interest not just in the moral development of individuals but also in the collective moral development of human beings as agents who will cooperate together over time and across cultures to achieve important moral outcomes. Obviously, any attractions to this view could help in response to versions of the problem of evil that point to those specific evils that look to be generated by failures of human cooperation.31 But my point here is to be directed at the hiddenness problem specifically. Suppose, then, that human beings are collectively responsible for the epistemic environment in which all of us will form our beliefs about, for example, whether God exists. Under these conditions, it may be that a person can quite reasonably come to believe that God does not exist—​or come to be in doubt about whether God exists—​and not because the person has made some moral or epistemic mistake in the use of her libertarian free will, but rather because her 29 

I ignore a great many responses to the argument here but I will be explicit, in particular, that I am ignoring all of those responses to the hiddenness argument that are based on the view that, appearances notwithstanding, there are no nonresistant nonbelievers. For my own part, I simply do not find it plausible to suppose that every nontheist is at some level willfully resisting what would otherwise be perfectly sufficient evidence that God exists. 30  Michael Murray, “Deus Absconditus,” in Divine Hiddenness: New Essays, edited by Howard-​Snyder and Moser (Cambridge University Press, 2002). 31  On this point, see Ted Poston’s “Social Evil” in Oxford Studies in the Philosophy of Religion.

642   Daniel Speak epistemic environment has been poisoned by countless apparently inconsequential such choices by other libertarian agents. If her unbelief is because of this poisoned environment, then it does not seem that her locally reasonable unbelief should constitute grounds for thinking there is no such being as God. Of course, this line of argument would need a good deal of development that it cannot be given here. What I want to emphasize, however, is that this would be a form of theistic reply to the hiddenness argument that would depend on further controversial views about morally responsible agency, in particular as it is related to epistemic states and collective action.

6.  Conclusions We have covered a lot of ground here; but we should conclude by returning to the main aims of this chapter. First and foremost, my aim has been to make salient some of the more influential ways in which theistic belief can be seen to put distinctive pressure on how we develop our theories of moral responsibility. One thing to see is that the ways are, in fact, legion; and this has been true even after imposing any number of restrictions on the scope of the inquiry. Thus, a second point to take away is that there are bound to be countless other ways in which broader religious commitments and outlooks will shape the theory of moral responsibility. To see some of the untapped potential here, consider the fact that this chapter hasn’t discussed original sin, inherited guilt, theories of atonement, divine forgiveness, or the many features of and theories regarding individual religious formation—​each of which will clearly have further implications for theorizing about moral responsibility. It is tempting to look for a through-​thread or overarching moral to the story of theistic influence on debates about moral responsibility. Perhaps there isn’t one, since many of the influences, challenges, and philosophical dispositions pull in competing directions. However, it might be unfair to ignore a recent bit of sociological data. The impressive PhilPapers survey of professional philosophers conducted in 2009 produced some striking results regarding the relationship between theistic commitment and views about human free will.32 Indeed, it can be argued that one of the strongest correlations unearthed by the survey was between theism and libertarianism regarding free will. For example, only 14.6% of all the philosophers surveyed either accepted or leaned toward theism. However, isolating just respondents who endorsed libertarianism, a full 50.8% of these were inclined toward theism, while only 39.8% were inclined toward atheism. In addition, a full 50% of those respondents who identified as theists also endorsed libertarianism. To give some context, only 4.6% of respondents who endorsed physicalism about the mind also endorsed libertarianism, whereas physicalists claimed

32 See https://​philpapers.org/​surveys/​. Total respondents numbered more than 800 professional

philosophers.

Responsibility and Religion    643 to reject libertarianism at a rate of 85%. In other words, at least among contemporary philosophers, it seems the impact of theistic commitment has pushed forcefully in the overall direction of the view that, though human beings are indeed morally responsible, they would not be if the thesis of determinism were true. One possibility that cannot be dismissed out of hand, however, is that the influence in the direction of libertarianism is not philosophically innocent. Taking note of the PhilPapers data just mentioned, Manuel Vargas has suggested that the attractions to libertarianism of many theists is a product of motivated reasoning rather than of straightforward and probative philosophical inquiry.33 Vargas is particularly attuned to the theistic need to exonerate God for causing or allowing the world’s evils by way of appeal to the incompatibilist conception of human agency. The Vargas challenge thus muddies the water quite generally with respect to the nature of the influence of theism on our responsibility theories, since his concern about motivated reasoning can be extended across the kinds of pressures on theory that we have been considering in this chapter. With these concerns in mind, it will be appropriate to close by noting that, whatever philosophical pressures religious thought and practice have in fact exerted on theorizing about moral responsibility, the important goal of evaluating these influences for philosophical legitimacy remains. My hope is that our terse treatment of the five areas of unique theistic influence can ultimately contribute to the ongoing evaluative project.34

33  “The Runeberg Problem: Theisim, Libertarianism, and Motivated Reasoning,” in Free Will and Theism, eds. Kevin Timpe and Daniel Speak. (Oxford, 2016). 34  Many thanks to Ajay Ravichandran for thoughtful discussion of this topic and to Derk Pereboom and Manuel Vargas for extremely useful comments on an earlier draft.

PA RT X I

C A SE ST U DI E S

CHAPTER 31

Moral Resp onsi bi l i t y in the C on t e xt of Addict i on Doug McConnell

1.  Introduction Addicted people often cause significant harm to others in addition to the harm they do to themselves. To ensure a continual drug supply,1 addicted people often resort to theft, fraud, spending family savings, and so on. Furthermore, the intoxicating effect of some drugs increases the risk of violent offending. In the United States, roughly 20% of all prisoners have been sentenced for crimes committed to obtain drugs or money for drugs, while 14% of those incarcerated for violent crimes report that they committed their most serious offense for drug-​related reasons (Bronson et al. 2017: 6). Addicted people also harm those close to them by prioritizing drug use over being a good parent, partner, or friend. The children of addicted parents, for example, are much more likely to suffer abuse and neglect which undermines their life prospects (Barnard 2007: 14–​17). Friends and family are also hurt simply to see the self-​destruction of someone they love. Herein, I attempt to specify the extent to which addicted people are morally responsible for these actions and for failing to recover given the harm their addiction is causing to others.2 The wide range of positions on moral responsibility and addiction make it impossible to cover everything in a single chapter, therefore, I will narrow my scope and assume 1 

For ease of exposition, I will refer to drug addiction throughout, but much of what I say can be generalized to behavioral addictions. 2  Some people might maintain serious addictions without causing any harm, in which case there is no moral pressure to recover. But these cases will be very rare; the addicted person would need to have a steady supply of drugs and no friends or family that cared much for him or required much from him.

648   Doug McConnell the degree of control an agent has over their actions is the crucial factor in whether they are responsible for those actions. I will also adopt an “accountability” conception of moral responsibility whereby a person is accountable for an action if they were in sufficient control of it and, therefore, they can be fairly treated with sanctions (or benefits) (Watson 2004). Finally, I assume that the folk are generally correct when they hold each other to be morally responsible and I use this folk standard as a baseline. If one has less than typical control over an action, as addicted people appear to in certain circumstances, then one is less accountable for that action and deserving of lesser sanction (or benefit) (views that fall roughly into this category include Sinnott-​Armstrong 2013; Watson 1999; Wallace 1999; Pickard 2017, 2015; Levy 2014, 2011).3 There are, of course, many theorists who prefer various self-​expressive views of moral responsibility. On these views, people are responsible for actions that are produced in the right kind of way from the agent’s beliefs, desires, and/​or values (even when the agent is not in control of those actions). Such views are often paired with an “attributability” conception of moral responsibility, whereby an action is attributable to someone if it expresses who they are which then justifies moral appraisal of the person on the basis of the act. However, depending on the account, further conditions may be required to justify sanctioning or benefiting that person as a consequence of their act (Scanlon 1998; Watson 2004). The question of whether addicted people are responsible for their actions then depends on the extent to which addicted actions are self-​expressive. If the motivations behind addicted actions are not self-​expressive, or not as self-​expressive as the motivations behind non-​addicted actions, then those actions may be less attributable to them (accounts that fall into this category include, Sripada 2017; Schroeder and Arpaly 2013; Arpaly and Schroeder 2013). This chapter should still be valuable independent of one’s preferred account of moral responsibility because it sets out the distinguishing features of addiction which will be relevant either as factors that potentially undermine control or that potentially undermine self-​expression. The chapter is structured as follows. I outline what counts as “addiction” and specify three important, semi-​independent etiological factors—​dopamine dysregulation, detrimental environments, and detrimental self-​conceptual content. I then explore the extent to which these factors (and interactions between them) might mitigate addicted people’s responsibility by undermining their agential control. I conclude that addicted people’s responsibility is mitigated in certain situations but only to a moderate extent at most. Finally, I consider and largely reject arguments that addicted people should be held responsible for becoming addicted or for failing to recover.

3  Hanna Pickard takes addiction to involve an impairment of control, but she develops a conception of moral responsibility that I take to be a variation of accountability whereby people with such impairments should face sanction for acting with ill will or negligence but, rather than be exposed to blame and hostile feelings (which may worsen their impairment), they should be treated with compassion and empathy even when sanctions are applied (Pickard 2017: 175).

Moral Responsibility in the Context of Addiction    649

2.  What Is Addiction? If we are to understand the moral responsibility of addicted people, we need to specify who counts as addicted. For present purposes we are interested in clinically significant addiction. Clinically significant addiction is conceptually linked to a broader, folk conception of addiction. The folk use “addiction” to refer to behavior motivated, in part, by a recurring strong desire to use a substance and/​or a desire to counter the unpleasant effects of withdrawal, for example, “I’m definitely addicted to caffeine, It’s the first thing I do every morning and if I don’t have a cup of coffee in the morning, I get headaches and my day never really gets started.” The folk also use a slightly different conception of “addiction” (not mutually exclusive of the first) to refer to pleasurable behavior that is imprudent. For example, I might say, “I’m addicted to these Twinkies,” with the clear implication that, although I enjoy eating them, I regret the amount I am eating due to the costs to my health. In contrast, regular caffeine use usually doesn’t have this imprudent feature because it doesn’t typically involve trading off anything of importance.4 There are two related continua here. Some prior substance use may only generate relatively weak physiological pressures to use again while other patterns of use may generate much stronger pressures. Regarding the prioritization of short-​term pleasures, some behavior, such as the occasional Twinkie binge, does not involve breaking social norms, while some patterns of drug use involve the consistent breaking of social norms and involve some severely self-​destructive and immoral behavior. Clinically significant addiction is usually thought of as a pattern of imprudent, physiologically driven behavior at the more serious end of these continua. The dominant method of setting the threshold of clinical significance in Western Psychiatry is the American Psychiatric Association’s Diagnostic and Statistical Manual (the DSM 5) which diagnoses patients with Substance Use Disorder using 11 criteria: (1) cravings and urges to use the substance; (2) needing more of the substance to get the effect you want; (3) development of withdrawal symptoms, which can be relieved by taking more of the substance; (4) spending a lot of time getting, using, or recovering from use of the substance; (5) not managing to do what you should at work, home, or school, because of substance use; (6) continuing to use, even when it causes problems in relationships; (7) giving up important social, occupational, or recreational activities because of substance use; (8) using substances again and again, even when it puts you in danger; (9) continuing to use, even when you know you have a physical or psychological problem that could have been caused or made worse by the 4 

The folk sometimes use “addiction” even more broadly to refer to a behavior that someone particularly likes and engages in regularly while “addictive” games or foods may mean simply that they are pleasurable. This conception of addiction doesn’t entail breaking social norms so it is not relevant to clinically significant addiction.

650   Doug McConnell substance; (10) taking the substance in larger amounts or for longer than the time you meant to, and; (11) wanting to cut down or stop using the substance but not managing to. The severity of substance use disorder is gauged by how many of the criteria the patient has met over a 12-​month period. If two or three criteria are met the disorder is mild, if four or five are met the disorder is moderate, while if six or more are met then the disorder is severe. These criteria build on the narrower folk-​psychological conceptions of addiction described before. Criteria (1), (2), and (3) aim to capture the experience of the physiological features of some kinds of addiction, while criteria (4) to (11) aim to capture the seriously self-​destructive, imprudent aspect of addiction where drug use undermines values that nearly every person would judge to be central to a good life, for example, relationships, health, and self-​sufficiency. We might quibble over where the threshold of clinical significance should be set (is it really enough to meet just two of these criteria to be mildly addicted?). We might also question whether the physiological features are relevant given that, if a person is successfully pursuing their values, why would it matter if they have to manage withdrawals and cravings? Nevertheless, it is the consistent inability of the addicted agent to meet a variety of social norms that is the fundamental problem, that is to say, addiction is a disorder that affects a person in a social context (Levy 2013). The DSM criteria may not be perfect, but they are one way of distinguishing those who break norms severely enough to be the focus of clinical expertise. The problem with the DSM criteria in general is that, because they focus on symptoms, they say nothing about what causes or constitutes mental disorders, for example, indicative brain pathologies. However, one of the reasons the DSM criteria are still in use is that, despite decades of searching, we have been unable to find brain pathologies or etiological factors whose presence guarantee the presentation of addiction (or any other mental disorder) (Levy 2019). When I discuss the various etiological factors involved in addiction in this chapter, I will briefly comment on their implications for the constitution of addiction but, for the moment, we can note that something appears to have gone awry with clinically addicted people’s agency, given that they can’t pursue core human values, and this lack of capacity might mitigate responsibility to some extent. It has been objected, however, that addiction isn’t a mental disorder at all; addicted people simply value drug use very highly and so they should be held to typical standards of moral responsibility. There are two main variants in this family of views: the moral view, on which drug use itself is morally wrong and, therefore, so too are any acts in service of drug use; and the liberal view, on which drug use is a morally permissible preference that has been wrongly pathologized for political ends (Foddy and Savulescu 2010). On the latter view, once we remove the prejudicial barriers to drug access, we wouldn’t expect drug-​using people to behave any more immorally than anyone else. Moral and liberal views are supported by research which shows the control that people

Moral Responsibility in the Context of Addiction    651 classified as addicted exercise over their lives. Roughly 80% of people who have ever met the DSM criteria for addiction recover without treatment, usually in their early 30s (Heyman 2009; Carlson 2006). This appears to indicate that when people find other things that they value more than drug use, they simply change their behavior accordingly. Furthermore, addicted people engage in sophisticated planning to obtain drugs, such as using false identities to obtain scripts from physicians and moving around different areas so as not to use the same doctor twice (Addenbrooke 2011). These plans are flexible in the face of unexpected contingencies, for example, the addicted person will change their plan if there is a police car outside the dealer’s house. Addicted people don’t just respond to serious and obvious risks but also relatively modest incentives, for example, refusing doses for financial incentives of $5–​$20 and adjusting how much one purchases according to price (Hart 2014; Neale 2002; Fingarette 1988). In other words, addicted people’s behavior remains responsive to a variety of non-​drug-​using reasons such as staying out of jail and non-​drug opportunity costs. The diachronically ordered, flexible nature of addicted action indicates that it involves executive reasoning and oversight; such behavior would be impossible if the agent was unconscious of their behavior or attempting to inhibit that behavior (Levy 2011: 92). Despite the evidence for a high degree of control in relatively synchronic assessments of addicted action, there remain strong reasons for thinking addiction involves a deficit in control. When we consider a more diachronic time frame, we see highly self-​ destructive, chronic ends-​ends inconsistency which goes well beyond typical agential failures. The corollary of the claim that 80% of people recover without treatment is that 20% of people struggle to recover to the point that they seek treatment. Those who seek treatment have a rocky prognosis. Most rounds of treatment don’t lead to permanent recovery; addicted people often relapse and some never recover (Brandon, Vidrine, and Litvin 2007; Stanford, Banerjee, and Garner 2010; Dennis et al. 2005). These addicted people spend much more time and effort attempting to recover than would be necessary if it were merely a façade to deflect opprobrium. If they really valued drug use that highly, then why would they thwart their own pursuit of it? Holders of the liberal view” or, if no space, just “Some might claim that these addicted people actually prefer an ends-​ends inconsistent life and we shouldn’t insist that people have diachronically stable goals. But this is implausible. Some drug use might feature in a good life but not so much that it comes at the cost of all those things that require more diachronically stable action, for example, close relationships, social and recreational activities, good health, and the capacity to work and learn. Furthermore, many addicted people’s lives appear miserable, living in squalor; suffering from painful concomitant medical problems, such as peptic ulcers; periodically scrambling for money to pay for the next dose; and committing suicide at a high rate (Darke and Ross 2002). The first-​person reports of shame, hopelessness, and social rejection, especially from more severely addicted people, corroborate this interpretation of their lives (Flanagan 2013; Kennett, McConnell, and Snoek 2018). So, what explains this self-​destructive action?

652   Doug McConnell

3.  Etiology 3.1  Dopamine Dysregulation An influential account of how addiction undermines agency claims that addictive drugs cause pathologically strong incentive salience desires, that is, the visceral, appetitive desires that typically develop in response to rewards such as food, water, and sex (Berridge, Robinson, and Aldridge 2009; Robinson and Berridge 2003; Dill and Holton 2014). Non-​addictive incentive salience desires are created by mesolimbic dopamine flushes in relation to stimuli that an animal experiences as pleasurable. The release of dopamine is proportional to the degree of pleasure experienced. When that reward (or a cue that regularly precedes the reward) appears, the incentive salience desire is activated, motivating the animal to get the reward and further reinforcing the dispositional motivation for the reward. Importantly, when a familiar reward is as pleasurable as expected, there is no increase in dopamine and the strength of incentive salience desire remains steady while, if it is less pleasurable than expected, there is a decrease in dopamine and the incentive salience desire is weakened. Through this system, the strength of an incentive salience desire tends to be proportional to the actual size of the reward based on past experiences. Now, addictive rewards, such as nicotine, opiates, cocaine, alcohol, and perhaps stimuli involved in gambling, artificially generate large mesolimbic dopamine flushes (or simulate the downstream effects of dopamine) every time they are accessed. The usual attenuation of dopamine when the reward is as pleasurable as expected, or less pleasurable than expected, cannot occur. Therefore, addictive rewards create disproportionately strong and persistent incentive salience desires that indicate the addictive reward will always be better than last time no matter how pleasurable the dose actually was. Holton and Berridge (Holton 2009; Holton and Berridge 2013) claim that, over time, these artificial dopamine surges change the addicted person’s brain so that, when he perceives a cue for drug use, he experiences a motivation or “want” for the drug that is disproportionate to both the amount he “likes” the drug (i.e., the pleasure of using) and the amount he values the drug. This distinguishes addictive desires from non-​addictive desires.5 Although addictive incentive salience desires are distinct from normal incentive salience desires because of their unusual strength, they threaten self-​governance through the same channels as non-​addictive temptations (Dill and Holton 2014). First, they can distort judgment by drawing the agent’s attention to the attractive aspects of a reward (e.g., the delicious taste of a dessert) and making it more difficult to attend to the

5  On Arpaly and Schroeder’s (2013, c ­ hapter 11) self-​expressive view it is this difference which entails that addictive desires are not as self-​expressive as non-​addictive desires. Pursuit of addictive desires is only attributable to agents to the extent that they intrinsically desire (i.e., “like”) those ends and not to the disproportionate extent to which they “want” them.

Moral Responsibility in the Context of Addiction    653 negative aspects of such rewards (e.g., heart disease); or the positive aspects of alternative, more abstract and temporally distant, rewards (e.g., feeling healthy). When judgment distortion is strong enough it can lead to “judgment shift” where the agent changes their evaluation of the temptation to see it as the most valuable option after all (Holton 2009). Judgment shifts are distinguished from rational changes of mind because they are typically short lived, regretted, and occur without the agent having deliberated over new information that would justify the change in evaluation. Second, incentive salience desires might motivate the adoption of intentions in pursuit of the rewards they target unless the agent can find sufficient motivation to adopt a different intention. They can, therefore, motivate akratic action, that is, actions contrary to the agent’s synchronic all-​ things-​considered judgment.6 This explains why addicted people “can know perfectly well that continued consumption would destroy everything that they hold dear. [But] That does nothing to stop the rush of desire that is triggered by the thought or sight of the drug, or, more broadly, of the people, places or paraphernalia that have surrounded its consumption” (Dill and Holton, 2014: 3). In judgment shift and akrasia, the action in pursuit of the temptation appears (synchronically) sophisticated, planned, and flexible because it is guided by intentions. In akrasia, agents might confess that they don’t think they’re doing what is best for them, but otherwise there will be no synchronic clue that they are not fully controlling their actions. In the case of judgment shift, the action seems just as controlled as any self-​governed act because the agent will also claim that they are doing what they judge to be best. Contra Holton, Levy (2014) argues that the loss of control in addiction is not driven by desires but, rather, the dopamine surges of addictive rewards lead the agent to erroneously update his beliefs about the value of the drug.7 He draws on the prediction error minimization model of the brain to argue that addictive rewards undermine self-​ governance by causing unconscious changes in doxastic states. On this view, the brain creates models that it guesses will match the world. To the extent that the world fails to fit the model, there is an error. The brain attempts to minimize these errors either by updating the model or changing the world. The modeling process is hierarchical; when lower levels of modeling/​behavior cannot eliminate the error, the error is passed up to higher levels to eliminate. Most modeling happens at lower, unconscious, levels; conscious remodeling only occurs when errors are passed to the highest levels. An addicted person trying to recover is normally operating with a set of models consistent with drugs being unrewarding. However, when cued by an addictive reward, the dopamine flush suggests that the drug will be much more rewarding than the model expects. This is a large error that can be eliminated by switching to a different and readily available model of the world in which the drug is as rewarding as it seems (the new model change involves judgment shift). The agent changes to this model to eliminate the error, 6  This akratic form of agential failure depends on there being a volitional step in agency, where one commits to an intention, that is in addition to judging what is best. For a defense of this view and a summary of supporting evidence, see Dill and Holton (2014). 7  See Yaffe (2013) for a similar view but without using the prediction error minimization framework.

654   Doug McConnell and then uses the drug. Subsequently, evidence begins to mount up that contradicts the model in which drugs are highly rewarding. Eventually the agent eliminates those errors by reverting to the model that drug use is not valuable. The important difference between this account and the incentive salience account is that this switch between models can happen at a subconscious level so that the agent experiences the world as if drugs really are that rewarding. Such unconscious doxastic shifts are not like normal judgment shifts because they cannot be executively resisted at the time of the cue; therefore, the only way to avoid them is to avoid the cues. If Levy is correct, then the only way to maintain self-​governance in the face of addictive temptation is to minimize exposure to cues, because cues cause unconscious model change. To avoid cues, one must recognize which cues are dangerous and be able to plan accordingly, for example, taking the long way home rather than walking through an area where one used to source drugs. Planning draws on limited cognitive resources and the world is somewhat unpredictable, so the addicted agent will usually find it is impossible to consistently avoid all cues for drug use. If Dill and Holton are correct, then, in addition to avoiding addictive temptations, the agent can resist synchronically using the same executive resources that can be used to resist everyday temptation. For example, exerting executive control over attention to avoid focusing on the attractive parts of a temptation and focusing instead on the unattractive aspects of ceding to the temptation or the attractive aspects of more valuable goals (Dill and Holton 2014). Despite the potential efficacy of normal self-​control resources, the frequency and strength of addictive temptations relative to “everyday” temptations make it much more difficult to consistently maintain self-​control. When we consider that typical agents cede to “everyday” temptations fairly often, it should come as no surprise that addicted people occasionally cede to their much stronger and more persistent temptations. These stronger and more consistent addictive temptations draw more heavily on limited cognitive resources and, almost inevitably, the agent will happen to be low on cognitive resources at some point and then they will cede to the temptation.8 Another significant difficulty in addiction is that a lapse refreshes the disproportionate strength of the addictive incentive salience desire. Therefore, giving in to an addictive temptation is more likely to lead to a full relapse, while ceding to an extra slice of cake, say, won’t lead to a runaway increase in cake consumption. However we understand the threat to agential control posed by dopamine dysregulation, we might be tempted to say that it is this dysregulation that constitutes addiction. This is the brain pathology that indicates whether the disease of addiction is present or not. Unfortunately, the situation is not so straightforward because some people with these brain changes do not exhibit addicted behavior. People manage to recover from, or fail to develop addiction despite (presumably) suffering from the dopamine dysregulation that regular use of addictive substances entails (Levy 2013). There seems 8  Furthermore, Sripada (2018) has argued that we probably make errors in judgment at a certain rate. Even if this rate is infrequent, the regularity of addictive temptations will ensure that addicted people lapse occasionally just through error.

Moral Responsibility in the Context of Addiction    655 to be a range of additional factors that either make the agent more vulnerable to addiction, protect against addiction onset, or support recovery, including agential capacities, mental content, and environmental factors. I address some of these below. Dopamine dysregulation may not even be a necessary feature of addiction. Not all addictions seem to involve dopamine dysregulation (e.g., cannabis) and different substances and addictive behaviors seem to involve quite different degrees of dysregulation (Nutt et al. 2015). This might lead us to claim that some of these problematic patterns of behavior are not “true” addictions or that they belong to different classes of addiction, but I won’t pursue this idea here.

3.2  Environmental Factors Some environments make consistent drug use more attractive relative to recovery or not using drugs in the first place. The classic case used to make this point is that of the US soldiers who were using opiates consistently during the Vietnam war. When they returned from service, the large majority of these soldiers simply stopped using (Robins et al. 2010). Most of those who continued to use drugs upon their return were returning to environments in the United States where opiates were more readily available. The explanation for the soldiers’ drug use in Vietnam is that the effects of drug use were highly valued given the environment. The war was extremely stressful, life threatening, and increasingly seemed to be pointless, while drugs provided some escape and were readily available. Drug use was comparatively less valuable in their home environments where most had loving families, employment, their lives weren’t at risk, and drugs weren’t easily accessible. Whatever dopamine dysregulation they suffered from was not sufficient to maintain the addiction. War zones are not the only environments that can make drug use relatively attractive. Unfortunately, even in relatively wealthy societies, people can find themselves in hopeless environments that lack opportunities, such as addiction treatment services, education, training, and employment, and/​or that are miserable, such as being homeless or socially isolated. Such hopeless and miserable environments render drug use attractive as a pleasurable way to kill time or to help ignore discomfort and loneliness. In these cases, abstaining from drug use may just make life worse. People in these environments may correctly judge (i.e., even when not affected by judgment shift) that regular drug use is part of the best life available to them, despite also judging that a better life, should it become available, would involve a loving family, friends, a job, a house, etc. When people in these detrimental environments attempt to recover it can be much more difficult than in other environments because drugs and cues for drug use are more common. Resisting these more frequent temptations saps cognitive resources, leading to more frequent lapses. This is especially problematic if dopamine dysregulation has rendered the addicted person highly vulnerable to addictive temptations. Such environmental factors currently drive the opioid epidemic in the United States, where

656   Doug McConnell pharmaceutical companies successfully lobbied to increase the availability of opioid painkillers. An environment where opioids were readily available (and endorsed by healthcare professionals) initiated the epidemic across the country and, despite the subsequent decreases in prescribing, overdoses continue to rise in low socioeconomic areas, presumably because these areas make ongoing drug use more attractive (Dasgupta, Beletsky, and Ciccarone 2018). Perhaps most addicted people’s environments aren’t so miserable that abstaining from drugs would inevitably result in a more miserable life; there may be some ambitious, difficult paths to recovery that come with a high risk of failure. Nevertheless, the poor cost/​benefit analysis of attempting to recover can still entail that it would be unreasonably ambitious to try, that is, likely to be a waste of one’s time and effort. This will be the case if the steps to recovery are very costly, such as leaving one’s partner, and/​or if those steps only offer a low chance of success, and if recovery isn’t that much better than using anyway. Sufficiently miserable, hopeless environments can be an initial cause of addiction, but there is also a feedback mechanism whereby being addicted (whatever the initial etiology) tends to create such environments. A criminal history, health problems, and lack of housing, education, and work experience, can make it much more difficult to succeed in typical life plans than it is for non-​addicted people. Addicted people with unemployment and housing problems are less likely to complete treatment programs than those without those problems (Saloner and Cook 2013). Employers don’t want to take on the risk of engaging with a recovering person when they can employ someone less risky. By the time the addicted person “hits rock bottom,” and is desperate enough to try and recover, they may have already made their circumstances such that the cost/​benefit analysis of recovery is unfavorable.

3.3  Self-​Conceptual Factors In the chronic absence of options and/​or repeated relapses due to addictive temptations, agents can understandably give up on recovery. Such resignation tends to go hand in hand with the development of fatalistic self-​conceptual content, for example, “This is just who I am.” Arguably, self-​conceptual content has self-​fulfilling motivational force and so this kind of content supports ongoing drug use. Narrative self-​constitution provides one way of understanding the motivational effects of self-​conceptual content (McConnell 2016; McConnell and Snoek 2018). On this view, agents self-​narrate to make sense of their lives. An obvious way to make sense of repeated relapse or consistent drug use in a world lacking better opportunities is that one is living out a narrative of hopeless addiction. Self-​narration is socially mediated; others co-​author the agent’s self-​narrative by contributing content and verifying or challenging the content that she claims for herself. In the case of addiction, this co-​authoring is often stigmatic and

Moral Responsibility in the Context of Addiction    657 reinforces the fatalistic narrative of someone who is worthless and hopeless (Kennett, McConnell, and Snoek 2018). Self-​narratives have motivational force because to act and interpret the world in ways contrary to the existing self-​narrative generates cognitive dissonance. Therefore, addiction self-​narratives channel attention and behavior in directions that make sense given that narrative, that is, toward drug-​using actions, people, and environments. This creates a feedback loop where the addictive environment and the addicted self-​ narrative reinforce each other. Detrimental self-​narratives can be changed but that takes narrative work, especially when one is attempting to change long-​established self-​narrative aspects that have been inter-​subjectively verified for years. New goals and characteristics that don’t readily make sense within the existing self-​narrative feel alien to the agent and appear implausible to her and her peers. Therefore, the agent has a battle to convince herself and her peers that, contrary to long-​standing belief, the changes are true (McConnell and Snoek 2018). The struggle to make significant changes to an established detrimental self-​narrative will typically take months if not years. Addiction self-​narratives don’t just tend to entrench a lack of belief in one’s self-​ efficacy, they also tend to focus on shameful memories which support other damaging self-​regarding attitudes, that is, low self-​respect, self-​esteem, and self-​trust. As a result, the addicted person can feel less than a moral equal to others and believe that they don’t deserve to recover or have a good life. The addicted agent can therefore become less able or willing to notice and pursue more valuable opportunities should they present themselves and overly pessimistic about her chances of ever meaningfully connecting her life with more valued outcomes. In contrast, continuing the established self-​narrative makes sense despite it being disvalued; this is who one is. In summary, a range of etiological factors can lead to and entrench addiction and the above list is certainly not exhaustive.9 Initially, one of these etiological factors might be dominant in causing addiction, but, over time, feedback mechanisms tend to ensure that additional etiological factors develop. Each of these etiological factors can resolve the apparent paradox of addicted agency being sophisticated and flexible yet self-​destructive at the same time. Addictive temptations can generate much more significant end-​ends inconsistency than everyday temptations. Detrimental environments can mean that apparently self-​destructive behavior is in fact a way of making the best of a miserable situation. Entrenched detrimental self-​concepts can prevent the agent from noticing paths to recovery or believing they deserve to recover.

9  Other significant etiological factors include various genetic risk factors and their interactions with particular environmental factors (Kendler et al. 2012), the influence of context-​dependent automatic habits (Schroeder and Arpaly 2013), desensitization to more temporally distant rewards (Bechara 2005) and, perhaps most seriously, an inability to even develop a conception of a good life (Kennett 2013: 157)

658   Doug McConnell

4.  Moral Responsibility To work out the situations in which addicted people’s responsibility is mitigated, we need to focus on when and where addiction influences agency. To this end, it is helpful to describe the effects of addiction according to how they fall on the global-​local spectrum and the static-​episodic spectrum for mental disorders (King and May 2018). Mental disorders that affect a larger proportion of the agent’s psychological processes are more global, for example, loss of consciousness and mood changes. Phobias and obsessive-​compulsive impulses are more local. Mental disorders are static when they affect the agent constantly, for example, schizophrenia or generalized anxiety disorder, and episodic to the extent that they manifest in discrete instances, for example, narcolepsy. Cases of addiction fall differently on these spectra depending on their etiology. Therefore, I will divide cases of addiction into several etiological groups: (1) addiction caused purely by dopamine dysregulation; (2) addiction caused purely by detrimental environments, (3) addiction caused by dopamine dysregulation and detrimental environments, and (4) addiction caused by a combination of dopamine dysregulation, detrimental environment, and detrimental self-​conceptual content. The etiological combinations represented by these groups are not exhaustive, but they illustrate a good proportion of the different effects that addiction has on moral responsibility.

4.1  Group 1—​Dopamine Dysregulation Only Some addicted people are vulnerable to addictive temptations due to dopamine dysregulation but have not yet developed detrimental self-​concepts and live in relatively good environments (i.e., that offer support for recovery, aren’t too heavily polluted with cues for drug use, and can support a good life if abstinent). For these people, addiction has specific and episodic effects because it is only when cued for drug use that they experience a disproportionately strong temptation that drives judgment shift or akrasia. Outside of such episodes, these people are no different to typical agents and so addiction has no mitigating influence on their responsibility. Furthermore, these people’s vulnerability is highly specific—​it generates a disproportionate motivation for drug use (and things that have come to be associated with drug use). So even during an episode of addictive temptation, the influence of addiction wouldn’t mitigate responsibility for immoral actions that weren’t targeted at accessing drugs, for example, stealing to buy a new television. The situation where these agents might have mitigated responsibility is when they are under addictive temptation and we are judging moral responsibility for an act in pursuit of drugs, for example, stealing money to pay for the next dose or using a drug despite knowing that their addiction is harming others. The extent to which addictive temptation mitigates responsibility depends on how much more judgment shift

Moral Responsibility in the Context of Addiction    659 and akratic motivation it causes relative to typical temptations.10 Here we face the difficult empirical issue of knowing how strong a specific addictive temptation is for an individual. We may have to rely on folk psychology, that is, making educated guesses based on the agent’s behavioral history and comparing their behavior to other agents who seem similar in the relevant ways. Working out the degree to which an agent’s responsibility is mitigated also depends on the extent to which they try to use their resources of cognitive control. An addictive temptation would not mitigate responsibility at all if the agent happened to have enough cognitive control that he could easily avoid or resist it but chose not to bother. Alternatively, if the agent did all they possibly could but failed, then they should be excused. So mitigation of responsibility is proportional to the effort put in to act morally but, again, we will probably have to rely on folk psychology to estimate the extent of an agent’s self-​control resources and effort at any time.11 (For a similar idea, see Yaffe 2011.) The degree of mitigation provided by judgment shift also depends on the agent placing proper weight on others’ interests while not under judgment shift. If the agent gives others’ interests less weight than he should then he allows those interests to be overridden too easily under judgment shift and so responsibility is less mitigated. For example, if an agent thinks that stealing from others is wrong but not such a big deal then he will be vulnerable to judging that stealing is acceptable even under relatively weak temptations. For the same reason, judgment shifts provide less mitigation for extremely wrongful actions, for example, a large judgment shift should be insufficient to deceive someone into judging that it was permissible to kill in pursuit of drugs given the very high weight that should be placed on others’ lives. We can say something analogous for akratic motivations driven by addictive temptations. The more immoral the pursuit of drugs would be, the more easily one should be able to summon motivation for an alternative intention. So when addicted agents are exposed to addictive temptations, their responsibility is mitigated according to the size of the addictive temptation relative to the self-​control resources available, the effort they make, and the weight that should be placed on others interests.12 The behavior of most people with dopamine dysregulation suggests that the typical size of addictive temptation might lead us to excuse some lesser immoral actions, such as failing to attend appointments, and it might mitigate the punishment for some more serious immoral acts, such as theft in service of drug use. However, if the immoral act is very serious, for example, murdering someone to access a drug, then the degree of mitigation will have a negligible impact given the magnitude of appropriate blame. 10 

Some people might claim that even strong non-​addictive temptations mitigate responsibility to some extent. I remain agnostic on that issue. 11  For an extended discussion of how degrees of blameworthiness and praiseworthiness can depend on the degree of difficulty and degrees of sacrifice required, see Dana Nelkin (2016). 12  Addicted agents might also have mitigated responsibility when they cede to typical temptations if their resources of cognitive control have been excessively drained through resisting addictive temptations.

660   Doug McConnell

4.2  Group 2—​Environmental Factors Only People in Group 2 are addicted because, given their environment, it is arguably prudent to continue using drugs. This is the case when the risk/​reward ratio of recovery is unfavorable, either because a lack of opportunities entails that abstinent life would be worse, or the costs of trying to recover are too high given the chance of success. People in Group 2 are addicted to drugs that don’t, or have not yet, caused dopamine dysregulation (or perhaps they have protective physiological of psychological features that enable them to resist addictive temptations as well as “everyday” temptations). For these people, addiction is relatively static and global rather than episodic and specific. Indeed, it is because the miserable, opportunity-​poor features of their environments are unchanging that it is rational to continue using drugs. Detrimental environments also have relatively global effects on psychology because they reduce the best realistically achievable life available to pursue. This puts a restriction on the agent’s long-​term plans and expectations, which frames how they think, act, and feel from day to day. In these situations, for example, people often don’t form mid-​to long-​term plans, rather their lives tend to be structured by short repeating patterns focused on finding safe spaces, shelter, warmth, food, and drugs. Do the global and static effects of these environments mitigate or excuse immoral acts, including acts that maintain an addiction that is harming others?13 Moral responsibility might be mitigated for immoral acts when people have little other way to access value in their lives and the cost to others is not excessive. That’s to say, we might not sanction someone for stealing to get high for the same reason that we might not sanction a starving person who steals food (however, we could still maintain the stealing was wrong in both cases). Exactly how much an addicted person’s responsibility is mitigated depends on the extent to which their environment is unjustly devoid of opportunities. If everyone was living in an equally miserable environment, then that environment would not mitigate the actions of someone who treated another citizen with ill will merely to alleviate his own misery. But when some people enjoy an environment replete with opportunities while others live in misery, the well-​off should be prepared to make some sacrifices to the extent that they tolerate, and so are complicit in maintaining, those unjust conditions. Obviously, which environments count as unjust will depend on one’s theory of distributive justice, but I cannot pursue that here. Given that most individuals who might be harmed by the addicted person will be only slightly responsible for the addicted person’s detrimental environment, it wouldn’t be fair for any one individual to pay too dearly for the addicted person’s drug use. In cases where we judge that the addicted person should not be required to compensate a victim by way of sanction, we might think society should compensate the victim. As with addictive temptations, it seems likely that miserable environments will only be

13  Ideally, of course, public money would be spent to improve the environment so that the cost/​benefit ratio of recovery improves, but such changes are often frustratingly slow and significant changes might not even occur within the addicted person’s lifetime.

Moral Responsibility in the Context of Addiction    661 moderately mitigating. They might significantly mitigate sanctions for failing to meet lesser obligations to friends and partners, and perhaps small thefts. They could not mitigate more serious wrongs such as serious assault or theft of large sums from individuals because such acts are unnecessary to relieve the addict’s suffering and disproportionately harm citizens who are (in most cases) only slightly complicit in maintaining the addicted person’s miserable environment. Here one might say that addiction is not the mitigating factor because it is the unjust environment that is doing the work. However, for individuals in Group 2, the environment is a constitutive feature of the addiction; if we improved the environment, they would recover.14 Do detrimental environments mitigate people’s responsibility for the harm they cause through failing to recover? The degree of mitigation depends on how much more sacrifice and effort is required of the addicted person to recover compared to what we would expect of someone in a more typical environment. The costs of not using drugs in a miserable environment may be much higher than in an environment with a reasonable range of opportunities, so it is more understandable that people don’t want to take them on and their failure to do so is more mitigated. However, as the harm their drug use is doing to others increases, the more risk and cost they should be prepared to take to recover. The situation is complicated by the fact that trying to recover can also increase harm to others due to abstinence violation effects (e.g., see Collins and Lapp 1991) and the subsequent reduced belief in one’s powers of agency (discussed in more detail later). In these cases, it may even be morally obligatory to maintain one’s addiction (for now) despite the fact that it is causing harm to others rather than risk causing even greater harm by trying to recover in unfavorable circumstances. A further complication is created by the feedback mechanism whereby addiction tends to erode the opportunities in one’s environment. This entails that it is usually worth taking greater risks to try and recover early in addiction. Later in addiction, when one has already lost one’s social standing, there is less cost in waiting until circumstances better support recovery. At the other end of the environmental spectrum, some addicted agents have access to resources which make it much easier to recover or, at least, mitigate the harm their addiction causes, and their prospective recovered life may be very good. Doctors and airplane pilots, for example, recover from addiction at a much higher rate than others, which is presumably because these people get access to better support, there is a demand for their skills, and they are familiar with a valuable non-​addicted life that they can (at least early in the addiction) realistically hope to return to (Heyman 2009). The rich can ensure a steady supply of drugs without acting immorally and they can also pay others to meet some of the obligations they have to others, for example, hiring help to look after children. Therefore, addicted agents in favorable circumstances can be held to higher standards than those in worse circumstances. These people are obliged to try and recover if their addiction is harming others.

14 

For an argument in favor of externalist conceptions of how mental disorders are constituted, see Davies (2016).

662   Doug McConnell

4.3  Group 3—​Brain and Environmental Factors Combined The majority of clinically significant addictions probably fall into Group 3, where people suffer from both the specific and episodic vulnerability to addictive cues and the global and static effects of detrimental environments. These two etiological factors interact in various ways. First, addictive temptations reduce the chances of those in Group 3 successfully recovering relative to those in Group 2. Therefore, those in Group 3 are justified in being resigned to their addiction, despite the harm it causes others, in some cases where those in Group 2 would be obliged to attempt recovery. Second, addictive judgment shifts will periodically cause those in Group 3 to overestimate how much harm they can justifiably do to service their addiction so, in these instances, their responsibility is more mitigated than those in Group 2. Third, if living in a miserable environment mitigates responsibility in ways discussed earlier, then the responsibility of Group 3 is also more mitigated than someone in Group 1. However, Group 3’s greater mitigation of responsibility relative to Group 1 will be offset to some extent by the fact that those in Group 3 should be better prepared for the effects of addictive temptations. This is because those in Group 1 might be trying to abstain and so might act more desperately to access drugs when an episode of addictive temptation strikes. Those in Group 3, in contrast, are resigned to their addiction and so have more time to ensure a drug supply while minimizing harm to others.

4.4  Group 4—​Brain, Environmental, and Self-​Conceptual Factors Detrimental self-​conceptual contents are a further etiological factor that tend to develop in long-​term addictions. Due to space constraints, I will just analyze the influence of detrimental self-​concepts on moral responsibility when they operate in addition to brain and environmental factors (but they might equally combine with either factor alone and perhaps, in some rare cases, a detrimental self-​concept may be the sole etiological factor maintaining an addiction). Detrimental self-​concepts have a combination of static, episodic, global, and specific effects. In long-​term addictions, the beliefs that one is hopeless and worthless addict can become entrenched within established self-​narratives. These detrimental beliefs are static to the extent that they take significant narrative work to change and may never be far from one’s attention given the centrality of addiction and drug use in the person’s life. Self-​narratives have relatively global effects when the content of the narrative involves general descriptions of oneself, for example, I’m untrustworthy, unemployable, etc. Such general beliefs have wide-​ranging effects because they inform many decisions about which plans and values the agent believes they can or should pursue. Self-​narratives can also have more specific and episodic effects according to the specific content of the self-​narrative and the specific situations in which it becomes relevant.

Moral Responsibility in the Context of Addiction    663 For example, one might have established a personal narrative ritual of celebrating with drug use when one has a win betting on the horses. This narrative ritual only influences thought and action when one is considering betting on horses and how to behave after winning such bets. To the extent that detrimental self-​narratives motivate continuing drug use, and make recovery feel alien and implausible, they further undermine chances of success in recovery. Therefore, they decrease the cost/​benefit ratio of attempting to recover. So the moral responsibility of those in Group 4 is typically further mitigated than those in Group 3. Furthermore, detrimental self-​narratives blind the agent to opportunities to recover making him less able to recognize when his environment happens to improve. So detrimental narratives mitigate responsibility even when the cost/​benefit of recovery has improved to a point that those in Groups 2 and 3 would no longer have their responsibility mitigated.

5.  Historical Control One might reply that we should still hold addicted people to the typical standard of moral responsibility despite all these factors because these people had control before addiction began. They should have recognized the danger and avoided addiction developing. However, in my view, the connections to prior control are insufficiently strong in most cases of addiction to hold addicted people to the same levels of moral responsibility as typical agents. First, people often begin their regular drug taking when they are young and have not yet developed adult agential powers. Furthermore, even once typical agency has developed, non-​addicted adults aren’t that good at taking long-​term consequences into account and we shouldn’t hold addicted people to higher standards than the non-addicted. Second, many non-​addicted people use drugs quite regularly without becoming addicted, so it is not clear how much weight a rational agent should put on the risks of addiction developing. It seems that those who develop addiction are better characterized as victims of bad luck rather than having been particularly reckless. Third, for those who live in miserable environments with relatively few opportunities, the risks of addiction seem negligible relative to the benefits of making today manageable. So we could really only hold someone responsible for becoming addicted if they had exceptional foresight and/​or reason to think themselves highly susceptible to addiction, and the opportunity cost of foregoing drug use was not high. Given the objections with tracing conditions to distant historical control, Stephen Morse (2011: 191) develops what he takes to be a more convincing variation which traces responsibility to the periods of relatively high control during addiction. Once severe addiction has developed, but during intervals between intoxication, the addicted person should recognize that their drug use is driving immoral action. If they fail to take some steps toward recovery during these intervals, then we can hold them responsible for their subsequent moral failures.

664   Doug McConnell But, as I have already shown, there are many situations in which addicted people should not be held to the same levels of moral responsibility as typical agents despite recognizing the harm they are causing. Most obviously, those in Groups 2 to 4 have their responsibility mitigated because they live in detrimental environments that make the cost/​benefit ratio of attempting recovery very poor. The degree of mitigation varies according to how poor the environment is, how strong the addictive temptations are (Group 3), and how dominating the detrimental self-​narrative is (Group 4). People in these groups can even be excused for deliberately continuing their drug use as long as they avoid causing too much harm to any one person. Morse’s argument applies most directly to Group 1 because these people’s environment provides a reasonable opportunity to recover. But even in Group 1, responsibility remains mitigated according to the strength and frequency of addictive temptations which may periodically cause the agent to act immorally in pursuit of drugs despite them making reasonable efforts to recover. Only once their vulnerability to addictive temptations has faded away can they be held to typical levels of moral responsibility at all times. Morse’s argument has some weight for Groups 2–​4 in cases where the environment happens to significantly improve. If such improvements occur, those in Group 2 can immediately be held to typical levels of moral responsibility and those in Group 3 can be treated identically to those in Group 1 (because now they only suffer from vulnerability to addictive temptations). Those in Group 4, however, continue to have their responsibility mitigated even when the environment improves because their self-​narrative hinders the recognition of opportunities. Even once the environment becomes so favorable to recovery that even they believe it is possible to recover, their responsibility remains mitigated because they remain vulnerable to addictive temptations and their self-​narratives continue to render recovery alien and implausible until revised through narrative work.

6.  Conclusion People with addiction have mitigated responsibility in a variety of situations and to varying degrees depending on the specific etiology of their addictions. I have argued that the relevant etiological considerations include vulnerability to addictive temptation; living in environments which make a life without drug use miserable; and fatalistic, stigmatic self-​narratives that have become entrenched over the years. Dopamine dysregulation mitigates responsibility to the extent that it causes abnormally large judgment shift and akratic motivation. Detrimental environments mitigate to the extent that giving up drug use in such environments imposes much greater costs on addicted people than when attempting to recover in environments with a typical range of opportunities. Addiction self-​narratives increase the magnitude of mitigation to the extent that they make recovery directed actions feel alien and implausible. They also tend to make the agent slow to recognize opportunities to recover, so they extend the period of

Moral Responsibility in the Context of Addiction    665 mitigation. Even when these factors combine, the magnitude of mitigation is unlikely to be huge; nevertheless, it is probably greater than society typically grants addicted people. Holding addicted people to an unfairly high standard of moral responsibility probably exacerbates addiction (Kennett, McConnell, and Snoek 2018), so it would be in the best interests of addicted people and society more broadly if we applied lower standards of moral responsibility according to the specific etiological factors involved and the situations in which those factors influence agency.

Acknowledgments This work benefited from helpful comments by Gabriel De Marco and Dana Nelkin, and was made possible by funding from the Uehiro Foundation on Ethics and Education and the Wellcome Trust grant “Responsibility and Healthcare” (WT104848).

References Addenbrooke, M. 2011. Survivors of Addiction: Narratives of Recovery. Hove: Routledge. Arpaly, N., and T. Schroeder. 2013. In Praise of Desire. Oxford: Oxford University Press. Barnard, Marina. 2007. Drug Addiction and Families. London: Jessica Kingsley Publishers. Bechara, A. 2005. “Decision Making, Impulse Control and Loss of Willpower to Resist Drugs: A Neurocognitive Perspective.” Nature Neuroscience 8 (11):1458–​1463. Berridge, K. C., T. E. Robinson, and J. W. Aldridge. 2009. “Dissecting Components of Reward: ‘Liking,’ ‘Wanting,’ and Learning.” Current Opinion in Pharmacology 9: 65–​73. Brandon, T., J. Vidrine, and E. Litvin. 2007. “Relapse and Relapse Prevention.” Annual Review of Clinical Psychology 3: 257–​284. Bronson, J., J. Stroop, S. Zimmer, and M. Berzofsky. 2017. “Drug Use, Dependence, and Abuse Among State Prisoners and Jail Inmates, 2007–​2009.” Carlson, B. E. 2006. “Best Practices in the Treatment of Substance-​Abusing Women in the Child Welfare System.” Journal of Social Work Practice in the Addictions 6 (3): 91–​115. Collins, R. L., and W. M. Lapp. 1991. “Restraint and Attributions: Evidence of the Abstinence Violation Effect in Alcohol Consumption.” Cognitive Therapy and Research 15 (1): 69–​84. Darke, S., and J. Ross. 2002. “Suicide among Heroin Users: Rates, Risk Factors and Methods.” Addiction 97 (11): 1383–​1394. Dasgupta, Nabarun, Leo Beletsky, and Daniel Ciccarone. 2018. “Opioid Crisis: No Easy Fix to Its Social and Economic Determinants.” American Journal of Public Health 108 (2). American Public Health Association: 182–​186. Davies, Will. 2016. “Externalist Psychiatry.” Analysis 76 (3): 290–​296. Dennis, Michael L., Christy K. Scott, Rodney Funk, and Mark A. Foss. 2005. “The Duration and Correlates of Addiction and Treatment Careers.” Journal of Substance Abuse Treatment 28 (2 suppl.). Dill, Brendan, and Richard Holton. 2014. “The Addict in Us All.” Frontiers in Psychiatry 5 (Oct.): 1–​20.

666   Doug McConnell Fingarette, H. 1988. Heavy Drinking: The Myth of Alcoholism as a Disease. Berkeley: University of California Press. Flanagan, Owen. 2013. “The Shame of Addiction.” Frontiers in Psychiatry 4 (October):120–​131. Foddy, B., and J. Savulescu. 2010. “A Liberal Account of Addiction.” Philosophy, Psychiatry and Psychology 17 (1):1 –​22. Hart, Carl. 2014. High Price: A Neuroscientist’s Journey of Self-​Discovery That Challenges Everything You Know About Drugs and Society. New York: Harper Perennial. Heyman, Gene. 2009. Addiction: A Disorder of Choice. Harvard, MA: Harvard University Press. Holton, Richard. 2009. Willing, Wanting, Waiting. Oxford: Oxford University Press. Holton, Richard, and K. C. Berridge. 2013. “Addiction Between Compulsion and Choice.” In Addiction and Self-​Control: Perspectives from Philosophy, Psychology, and Neuroscience, edited by Neil Levy, 239–​268. Oxford: Oxford University Press. Kendler, Kenneth S, Kristina Sundquist, Henrik Ohlsson, Karolina Palmér, Hermine Maes, Marilyn A Winkleby, and Jan Sundquist. 2012. “Genetic and Familial Environmental Influences on the Risk for Drug Abuse: A National Swedish Adoption Study.” Archives of General Psychiatry 69 (7). NIH Public Access: 690–​697. Kennett, Jeanette. 2013. “Just Say No? Addiction and the Elements of Control.” In Addiction and Self-​Control, edited by Neil Levy, 144–​164. Oxford: Oxford University Press. Kennett, Jeanette, Doug McConnell, and A Snoek. 2018. “Reactive Attitudes, Relationships, and Addiction.” In Routledge Handbook of the Philosophy and Science of Addiction, edited by Hanna Pickard and Serge Ahmed, 440–​451. London: Routledge. King, Matt, and Joshua May. 2018. “Moral Responsibility and Mental Illness: A Call for Nuance.” Neuroethics 11 (1). Springer Netherlands: 11–​22. Levy, Neil. 2011. “Addiction, Responsibility, and Ego Depletion.” In Addiction and Responsibility, edited by Jeffrey Poland and George Graham, 89–​111. Cambridge, MA: MIT Press. Levy, Neil. 2013. “Addiction Is Not a Brain Disease (and It Matters).” Frontiers in Psychiatry 4 (April): 24. Levy, Neil. 2014. “Addiction as a Disorder of Belief.” Biology & Philosophy. Springer Netherlands: 1–​19. Levy, Neil. 2019. “The Truth in Social Construction.” In Rethinking the Biopsychosocial Model, edited by Will Davies, Julian Savulescu, and Rebecca Roache. Oxford: Oxford University Press. McConnell, Doug. 2016. “Narrative Self-​ Constitution and Recovery from Addiction.” American Philosophical Quarterly 53 (3): 307–​322. McConnell, Doug, and Anke Snoek. 2018. “The Importance of Self-​Narration in Recovery from Addiction.” Philosophy, Psychiatry, and Psychology 25 (3):31–​44. Morse, Stephen. 2011. “Addiction and Criminal Responsibility.” In Addiction and Responsibility, edited by Jeffery Poland and George Graham, 159–​200. Cambridge, MA: MIT Press. Neale, J. 2002. Drug Users in Society. New York: Palgrave. Nelkin, Dana Kay. 2016. “Difficulty and Degrees of Moral Praiseworthiness and Blameworthiness.” Noûs 50 (2): 356–​378. Nutt, David J., Anne Lingford-​Hughes, David Erritzoe, and Paul R. A. Stokes. 2015. “The Dopamine Theory of Addiction: 40 Years of Highs and Lows.” Nature Reviews Neuroscience 16 (5): 305–​312. Pickard, Hanna. 2015. “Psychopathology and the Ability to Do Otherwise.” Philosophy and Phenomenological Research 90 (1) :135–​163. Pickard, Hanna. 2017. “Responsibility Without Blame for Addiction.” Neuroethics 10: 169–​180.

Moral Responsibility in the Context of Addiction    667 Robins, L, John E. Helzer, Michie Hesselbrock, and Eric Wish. 2010. “Vietnam Veterans Three Years after Vietnam: How Our Study Changed Our View of Heroin.” American Journal on Addictions 19 (3): 203–​211. Robinson, T. E., and K. C. Berridge. 2003. “Addiction.” Annual Review of Psychology 54: 25–​53. Saloner, B., and B. L. Cook. 2013. “Blacks And Hispanics Are Less Likely Than Whites to Complete Addiction Treatment, Largely Due to Socioeconomic Factors.” Health Affairs 32 (1): 135–​145. Scanlon, T. 1998. What We Owe to Each Other. Cambridge, MA: Harvard University Press. Schroeder, T., and N. Arpaly. 2013. “Addiction and Blameworthiness.” In Addiction and Self-​ Control, edited by Neil Levy, 214–​238. Oxford: Oxford University Press. Sinnott-​Armstrong, W. 2013. “Are Addicts Responsible?” In Addiction and Responsibility, 122–​ 143. Oxford: Oxford University Press. Sripada, Chandra. 2017. “Frankfurt ’s Unwilling and Willing Addicts.” Mind 126: 781–​815. Sripada, Chandra. 2018. “Addiction and Fallibility.” The Journal of Philosophy 115 (11): 569–​587. Stanford, Mark, Kakoli Banerjee, and Robert Garner. 2010. “Chronic Care and Addictions Treatment: A Feasibility Study on the Implementation of Posttreatment Continuing Recovery Monitoring.” Journal of Psychoactive Drugs Suppl 6 (September): 295–​302. Wallace, R. Jay. 1999. “Addiction as Defect of the Will: Some Philosophical Reflections.” Law and Philosophy 18 (6): 621. Watson, Gary. 1999. “Excusing Addiction.” Law and Philosophy 18: 589–​619. Watson, Gary. 2004. “Two Faces of Responsibility.” In Agency and Answerability, 260–​288. Oxford: Clarendon Press. Yaffe, Gideon. 2011. “Lowering the Bar for Addicts.” In Addiction and Responsibility, edited by J. Poland and G. Graham, 113–​138. Cambridge, MA: MIT Press. Yaffe, Gideon. 2013. “Are Addicts Akratic?” In Addiction and Self-​Control: Perspectives from Philosophy, Psychology, and Neuroscience, edited by Neil Levy, 191–​214. Oxford: Oxford University Press.

CHAPTER 32

Moral Resp onsi bi l i t y f or Im plicit Bias a nd the Im pact of S o c ia l Categori z at i on Maureen Sie

“Implicit bias matters, and anyone concerned with fairness and justice should want to know who’s responsible for it.” (Brownstein 2015: 765)1 “[...] in abnormal moral contexts, our entitlement to respond with moral reproach is independent of the blameworthiness of individuals.” (Calhoun 1989: 400) Suppose a young woman tackles a teenager running through a store with an expensive Gucci bag in his hands. The teenager falls over, hurts his knees badly, and becomes very angry with her. He was picking up the bag his mother forgot in the dressing room and was running through the store because a cab was waiting outside. Witnessing a scene like this, we might understand immediately why the woman acts as she does; in her eyes, the teenager might look like a thief. “Clearly,” we might think, “she tackled him to prevent a theft.” When the teenager belongs to an ethnic minority that is often depicted as “criminal,” we might also immediately understand why he responds with such anger. In our eyes he might not look like a thief at all and, moreover, he did not do anything wrong. Further, we might fully understand his frustration because, for him, the incident is bound to be one of many like it.

1  I thank Alf Archer, Huub Brouwer, Amanda Cawston, Matteo Colombo, Tim Klaasen, Ben Matheson, and Naftali Weinberger for helpful comments and discussion on an earlier version of this paper. I thank Neil Levy for his detailed criticism and suggestions and the editors of this volume for their helpful feedback on earlier versions of this chapter. I also thank the participants of the LMU research colloquium of Monika Betzler and to the Antwerp seminar organized by Katrien Schaubroeck for their helpful discussion of an earlier draft version of this chapter.

Moral Responsibility for Implicit Bias    669 According to a large, debated, body of research in psychology, the woman of our example might have acted under the influence of implicit attitudes with respect to people from this particular ethnic minority.2 In psychological literature, attitudes refer to broad positive or negative evaluative states that can be directed to all kinds of things. Implicit attitudes, in the domain of social categorization, are referred to as implicit biases; they allegedly bias our behavior in stereotypical and prejudiced ways.3 The woman’s attitude is called implicit (as opposed to explicit) because of the way it affects her. The implicit nature of her bias makes it the case that she could behave in a biased way even when she does not explicitly believe that the stereotypical depiction of people from this minority is correct.4 If the woman’s action is indeed caused by implicit bias, the question arises whether she is morally responsible for it. At first sight, implicit biases seem to escape our awareness and control; hence how can we be morally responsible for them? This chapter elaborates on the question of our moral responsibility for implicit biases and the philosophical discussion addressing it. It aims to bring across why the empirical literature on implicit bias (hereafter shortened to the IB-​literature) is unsettling and relevant when reflecting on our practices of moral responsibility. This is a difficult task. First of all, the IB-​literature is heavily debated and it is controversial what it establishes exactly. Secondly, it is at least as debated how we should understand the concept of moral responsibility, what the conditions of moral responsibility are, and whether we are in fact morally responsible beings. Nevertheless, even when we steer clear of the most disputed controversies in both debates, there are important lessons contained in the IB-​literature, highly relevant to our practices of moral responsibility.

2  Recently a critical meta-​analysis was published of the body of literature allegedly showing the causal influence of implicit bias on our behavior, more particularly the efficacy of changing our behavior and explicit attitudes via changing implicit biases. See Forscher et. al. 2017. This study suggests that we need to be careful with drawing any confident and final conclusions about the exact nature of the relation between implicit biases and our explicit attitudes and behavior. With a few exceptions, this corresponds with what many philosophers have been arguing for with regard to this literature, though the popular reception has been different. Big companies like Google and Facebook have implemented programs to raise awareness of implicit bias and the literature has found its way to many other professional communities as well. Whenever it is appropriate I use the term “allegedly” to make clear that the empirical results are debated. 3  The term stereotype was coined by Lippmann to describe the (simplified and culturally informed) “little pictures” we carry around inside our heads about various types of people (Lippmann 1922: 63). Stereotypes are commonly understood to be generalizations or, in terms of Valian, “schemas” (Valian 2005). These generalizations are simplified and standardized images that ascribe specific (psychological or other) characteristics to all members of a specific social category, without individuating between them (see, for example, Allport 1954; Jost and Banaji 1994, and many others). They are (some of) the cognitive underpinnings of prejudice. Prejudices are the often affective, evaluative application of triggered stereotypes. 4  In that case, if we would ask her why she behaved in this biased way she might fiercely object to this depiction of her action. She might even be offended, for example, because she belongs to the same ethnic minority, is often bothered by the stereotypes about this minority and believes the suggestion that her own behavior is biased is outrageous.

670   Maureen Sie In this chapter, I broadly follow the Strawsonian understanding of what it means to hold someone morally responsible, that is, that we respond to one another with one of the moral sentiments of blame, resentment, indignation, or its positive counterparts (Strawson 1962). I will also sometimes refer to these moral sentiments as “appraising responses,” as is common in the debate on our moral responsibility for implicit bias. Following this Strawsonian understanding allows me to do justice to the widespread appeals to forward-​looking reasons in the philosophical debate on our moral responsibility for implicit bias. Forward-​looking reasons appeal to the aims served by our holding one another morally responsible, such as morally improving our practices. Backward-​looking reasons, conversely, focus on what makes an individual deserve the moral sentiments that come her way. The broader philosophical controversies concerning moral responsibility tend to center around backward-​looking reasons. In the first section of this chapter, I discuss that part of the IB-​literature concerned with what I call the “mechanisms of implicit bias”. Many take this literature to show that the nature of implicit biases problematizes our moral responsibility for the actions they cause or influence. Nevertheless, most philosophers also believe that some kind of appraising response is needed, if only to counteract the widespread and structural harm implicitly biased actions allegedly cause. In the second section, I discuss this widespread and structural harm, that is, the impact of implicit bias. I argue that we need to distinguish between what I call “localized” and “general” harm. Examples of the former are phenomena such as shooter and evaluation bias; examples of the latter are phenomena such as microaggressions and inequities and epistemic injustice. Although it is clear that we can and should take measures to counteract localized harms, I argue this is unclear with respect to general harm. This is not to claim that the general harm caused by implicitly biased behavior is not worrying. I argue that the phenoma discussed under the heading of this general harm make it clear that our inclination to categorize one another socially is deep-​seated and widespread. The fact that implicitly biased behavior potentially causes harm in ways we might not have realized beforehand, I argue, might be the most important lesson to learn from the empirical literature on implicit bias. For one thing, it teaches us that we might not know yet what it is our so-​called egalitarian commitments commit us to exactly.5 In the third section I come back to the distinction between forward and backward-​looking reasons to hold someone morally responsible. I argue that a backward-​looking approach to the issue of our moral responsibility for implicit bias pays insufficient attention to the harm done by these biases and the fact that holding one another morally responsible might be the only way we counteract such harm. I use an argument articulated by Cheshire Calhoun to do so. I also argue that a forward-​looking approach that accepts the argument articulated by Calhoun fails to do justice to the possible and existing controversies about the need to counteract what I have referred to as the “general harm” done by implicit biases. In the discussion on

5 

With “egalitarian commitments,” I refer to our commitment to non-​discrimination, where discrimination is understood to broadly encompass all practices with discriminatory outcomes.

Moral Responsibility for Implicit Bias    671 our moral responsibility for actions caused or influenced by implicit biases, it is often taken for granted that our egalitarian commitment entails a commitment to counteract all harm allegedly done by implicit biases. A reflection on the differences between localized and general harm caused by implicit biases shows what is problematic about that assumption. I conclude that this leaves us in a situation where we need to engage in a discussion with one another about the potential harm done by implicit biases and the extent to which we are responsible for improving our moral practices in this respect. That observation brings a last worrying feature to the surface with which the IB-​literature confronts us: If our inclination to categorize socially is as ingrained and widespread as this literature on their impact suggests, it is bound to affect this discussion itself, as well. Hence, I agree with Brownstein that “implicit bias matters and anyone concerned with fairness and justice should want to know who’s responsible for it.” However, I also believe the IB-​literature shows us that simple and straightforward answers to that question are unavailable. We can only do our best to try and answer the responsibility question by engaging with one another as morally responsible individuals. Unfortunately, even that process itself is bound to be affected by our deep-​seated and widespread inclination to categorize one another socially.

1.  The Agential Dimension: The Mechanisms of Implicit Bias In this section, I discuss the mechanisms of implicit biases and the question of whether the literature gives us cause to doubt whether we are ever morally responsible for harboring or manifesting them. Although the idea of the existence of implicit attitudes is not new in philosophy or psychology, several techniques have only recently been developed and widely applied to measure them scientifically (Brownstein 2015; Brownstein and Saul 2016a: 3–​9). What these techniques allegedly have brought to the fore is their pervasiveness across populations and the many effects they have on our behavior, either indirectly, by influencing our evaluations and judgments; or directly, by steering our overt interactions with one another.6 I illustrate this in more detail in the next section. Also, numerous experiments have established that implicit biases are resistant to direct efforts to change or control them via force of will or mere intention; suppressing negative or stereotypical associations can lead to a rebound effect (Galinsky and Moskowitz 2000; Macrae et al. 1994; Monteith et al. 1998), or a stronger manifestation after the suppression (e.g., Holroyd 2012: 283, Plant and Devine 2001). Some other results indicate

6  Although critics of the literature on implicit bias point out that we can question whether we can generalize the findings, since many of the experiments were conducted on a select group of participants: that is, young, Western, educated subjects (mostly students). See, for example, Forscher 2017. For critical discussion of Forscher et al’s meta-​study, see the brains blog roundtable, M. Brownstein et al. 2017.

672   Maureen Sie that implicit biases, for example with respect to gender, are manifested to an equal degree by individuals who measure low or high on explicit sexist beliefs (Banaji and Hardin, 1996: 139, referred to by Holroyd 2012: 301). Philosophers have identified a host of issues that need to be addressed in the psychological literature on implicit biases,7 but in the context of harmful stereotypes, implicit biases are often understood to refer to (1) automatic association processes that function largely outside our awareness, (2) which affect how we “perceive, evaluate, or interact with people from the groups that our biases ‘target’ ” (Saul 2013) on the basis of the particular “stereotype and prejudice” combination that is at work in a specific situation (e.g., Kang and Lane 2010; Kelly and Roedder 2008).8 Often, the way implicit biases are discussed implies that the stereotypes and prejudices involved are negative, as we shall see in the next section.9 By presenting them in this way, that is, by emphasizing their negative effects, the philosophical question of whether we are morally responsible for harboring or manifesting implicit biases naturally follows (hereafter, I will use “for implicit bias” as shorthand).10 Let me explain this. Adult human beings are widely considered to be morally responsible in a way that young children and nonhuman animals are not. Many philosophers believe that adult human beings possess certain capacities that justify evaluating them morally in a deeper, more substantial sense than is justified in the case of young children and nonhuman animals. Some philosophers discuss the general worry that implicit biases undermine our status as morally responsible individuals and argue that the literature gives us reason to reconsider the general concept of moral responsibility (e.g., Faucher 2016) or the way we apply it in the domain of harmful social categorizations. Robin Zheng, for example, argues that we should distinguish between holding people “attributively responsible” for their actions and holding them accountable. She argues that even in those cases in which moral sentimental responses are not warranted, we can still be held accountable

7 

Such as how to understand the opposition between implicit and explicit attitudes exactly (e.g., Frankish 2016), whether we can make sense of a distinction between explicit and implicit attitudes, (Machery 2016), whether the “implicit” of “implicit attitudes” refers to the “automaticity” or the “lack of awareness” and, if the latter, how this lack of awareness needs to be understood exactly (e.g., Brownstein 2015; Kelly and Roedder 2008; Saul 2012). Also, whether phenomena commonly identified as implicit bias all bear the same characteristics. If they do not, as Holroyd and Sweetman have argued, this is likely to have considerable significance for the sorts of normative recommendations that are made regarding the mitigation of undesirable effects of the different classes of implicit bias (Holroyd and Sweetman 2016). 8  For an excellent overview of empirical research and theoretical perspectives in the area of implicit intergroup bias, how they are represented in the mind, and how they are expressed in behavior; see Amodio & Mendoza (2010). 9  Shooter bias, for example, refers to the tendency to misidentify objects carried by African Americans and the resulting disastrous actions that might result from this. Although the name might suggest that this bias explains why African Americans are more often being shot as a result of the suspicion of carrying a gun, more mechanisms might be at play. Like shooter bias, evaluation bias is also discussed mainly in the context of an unfair evaluation, for example, of women in an academic setting. 10  I discuss why this negativity is one-​sided, in the next section.

Moral Responsibility for Implicit Bias    673 for implicit biases (Zheng 2016: 74–​76).11 According to Zheng, when we reflect on the harm done by implicit biases we might think that even when we do not deserve blame or resentment for this, we can still be expected to compensate or take measure for the harm done by them. Other philosophers argue that the empirical literature on implicit bias gives us no reason to doubt that at least on some occasions, responding to them with moral sentiments is appropriate. Jules Holroyd, for example, identifies several conditions that are influential in the philosophical discussion on moral responsibility and argues that for each of these conditions, the empirical literature does not give us reason to think that we are never responsible for harboring or manifesting implicit bias. In the remainder of this section, let me follow the structure of her paper. One could think, Holroyd suggests, that we are not morally responsible for implicit biases because they are solely caused by the culture we live in (Holroyd 2012: 279–​280). On this view our implicit biases are the result of growing up and living in our society and not something for which we can be held morally responsible. However, the empirical literature shows us that the degree of implicit bias manifested differs among individuals (Holroyd 2012: 280). This could perhaps be explained by the fact that each of us has been exposed to stereotypes and prejudices in varying degrees. If that were the case, individual differences in manifesting or harboring implicit biases might still be unavoidable because they are caused by circumstances we have no control over, that is, the individual exposure to stereotypes and prejudices. However, as Holroyd points out, the empirical data suggests that it is not only our personal attitudes (as opposed to our general knowledge of the stereotypes and prejudices) that matter to the degree to which we manifest implicit biases. The degree to which we manifest implicit biases is also sensitive to the reasons for which we are committed to egalitarianism. Those individuals who are intrinsically motivated egalitarians (who care about being non-​prejudiced in itself and not only because it goes against the social norms or might lead to social sanctions) display less bias in tests than those who are not (Devine and Plant 2002; Holroyd 2012: 281).12 However, although the findings are robust, they are also small, and we need to realize that there is a lot variation across individuals when we hold constant the intrinsic motivation.13 The three other conditions Holroyd discusses in detail—​control, awareness, and reasons-​responsiveness—​all have also been the topic of critical scrutiny in the more general philosophical debate on moral responsibility (See, e.g., Clark 2013; Pettit, 2007, Schlosser 2013, Sie 2013, and Sie and Wouters 2008). This scrutiny is partly a response to general worries triggered by a plethora of findings in cognitive science about the role

11 

Note that the distinction drawn here between accountability and attributability is the opposite of what is common in the debate on moral responsibility and free will. In that debate, many use “accountability” to refer to our being an apt candidate for the reactive attitudes. 12  This does not seem to require effort or conscious control, hence, the hypothesis for this is that in these individuals there are “preconscious regulatory systems at work,” that it is not in response to awareness of implicit bias (Holroyd 2012: 281). 13  As rightfully pointed out to me by Neil Levy.

674   Maureen Sie of deliberative awareness, prior awareness of the reasons for which we act (cf. Sie 2009). As many philosophers have pointed out, the idea that we are only morally responsible when we act on reasons we are aware of prior to our actions is too simplistic and should be abandoned in favor of a more sophisticated one (cf. Arpaly 2004; Hieronymi 2004; Pettit 2007; Sie 2009, 2014; Smith 2008). This more sophisticated idea should take into account at least the fact that we can be held morally responsible for actions that somehow result from prior actions, for example, when we cause an accident due to alcohol consumption. The fact that we were not fully or sufficiently in control at the time of the action due to the alcohol, most people believe, does not mitigate our blame for it. In such cases our moral responsibility for the accident is called “indirect”: It traces back to the decision to drink (and drive). The reason we feel people are blameworthy for the wrong they do when drunk is that they are responsible for drinking too much. The philosophical debate on exactly how to conceptualize the idea of indirect moral responsibility is technical and complex. However, for the purposes of this chapter, it will suffice to bear in mind the intuitively plausible idea that a lack of control, awareness, and reasons-​responsiveness at a certain moment in time does not automatically undermine our moral responsibility for our actions at a later point in time. That observation already opens up a wide variety of ways in which it can be argued that we are morally responsible also for actions that are influenced by implicit biases. Furthermore, besides the intuitively plausible idea of the existence of indirect moral responsibility, there are also reasons to doubt that the implicit nature of certain mental states (such as implicit biases) undermines our moral responsibility for the actions that result from it. I come back to that issue more elaborately in section 3, but before that, let me roughly lay out the territory to explain why we need a more sophisticated picture of what makes us moral agents when discussing the IB-​literature by discussing the conditions of control, awareness, and reasons-​responsiveness. First of all, people are able and constantly expected to consider their knowledge about the way in which they function. Just like we are expected to know about the potential effects of drinking on our subsequent behavior, the fact that we have difficulty waking up, for example, is not considered a justification for being late to work. We expect people to use an alarm clock. In a similar vein, once we are informed about the existence and ubiquitous presence of implicit biases, this drastically changes our epistemic position with respect to harmful stereotypes and prejudices (cf. Kelly and Roedder 2008: 534). We can be expected to consider that knowledge and, for example, make use of psychological or institutional measures to counteract their influence. In the case of humans, self-​control cannot be simply taken to refer to the control we exert at the moment of our actions (see, e.g., Hörstkotter 2014).14 This is an example of what we referred to as indirect responsibility.

14 

And there are many other views in philosophy that have pointed this out, for example, virtue ethical theories.

Moral Responsibility for Implicit Bias    675 Second, people are able and, at least to a certain extent, expected to shape their views on the world, the things they care about, their attitudes, and their habits or ways of life. These views, cares, attitudes, and habits reflect on who we are as moral agents. Likewise, our implicit biases and their influence on our behavior might also reflect on who we are as moral agents (cf. Brownstein 2015; Smith 2008). If we think that we are responsible for being the kind of moral agents that we are, and our implicit biases reflect this, then their implicit nature (our not being aware of their influence) does not undermine our moral responsibility per se.15 If we think people are morally responsible for the kind of agent they are—​kind, open, hostile, frustrated—​then they could very well be morally responsible for implicit bias as well. Moreover, the fact that their influence on our behavior is dependent on our other motivational states also adds to the plausibility of our moral responsibility for implicit biases. As we already noted (with the caveat mentioned), people who are highly intrinsically committed to responding without bias “manifest significantly less negative race bias across a range of tests for implicit biases” (Devine and Plant 2002; Holroyd 2012: 288; Rees 2016).16 Hence, even in case there is something about the nature of implicit biases that somehow undermines our awareness and direct control over them, the fact that they are related to our other attitudes and commitments brings them back under our self-​control more broadly understood. That is, once we understand the notion of human agency in the sophisticated way that seems required to capture it adequately. Moreover, the empirical literature suggests that we manifest less bias when the aim or goal of treating people fairly is activated but not when we think that goal has been achieved.17 As soon as we think we have succeeded or reflect on an occasion in which we have treated people fairly, the bias manifests itself again (Holroyd 2012: 289–​290).18 Hence, the empirical findings suggest that besides the nature of our commitments, the other aims that we happen to have also influence the manifestation of our implicit biases. Lastly, studies suggest that informing people about their implicit biases leads to a negative self-​directive affect, people feel bad about this, and this has a behavioral effect in subsequent tasks. When making implicit race bias salient, people showed less racially biased responses (Monteith et al. 2002).19 Again, even in case it would be firmly established that the nature of implicit biases itself undermines our direct control over them because they escape our awareness, knowledge about their 15 

As announced, I come back to this issue in the third section of this paper, shortly introducing the view of Angela Smith and that of her main opponent in this debate, Neil Levy. 16  It is good to bear in mind that “significantly” is a technical term. It is unclear whether these measured effects make a difference to how we act in the real world. 17  This, as Holroyd points out, recommends a humble attitude with respect to our ability to treat others fairly (Holroyd 2012: 290; Sie 2013: 285), because such a humble attitude is more conducive to achieving the goal of actually treating other people fairly. 18  Holroyd bases herself on the findings of Moskowitz and Li (2011) and Park, Glaser, and Knowles (2008). 19  This salience has an effect on people which suggests they develop what Monteith et. all call a “behavioral cue,” a cue that the situation they confront is one in which they behave in a manner that deviates from their personal standards. This effect is substantially higher in people who measure low on prejudice (Monteith et al. 2002).

676   Maureen Sie existence and their influence on our behavior appears to have an effect on us, as well as on the manifestation of implicit biases. Third, not being deliberatively aware of what makes us act in the way we do—​not being able to report on it, for example, or not having articulated our reasons prior to acting—​does not mean that we are not in some way in control, that we do not “respond to reasons” or that we are not rightfully regarded as the “agent of the action” (Pettit 2007: 81, Cf. Smith 2005: 252). When individuals possess agential capacities, capacities to steer their actions in ways that correspond to their values and norms, we can and often do ask them to explain and justify their actions, even when they are not considered to be deliberatively aware of the reasons for which they acted (Cf. Pettit 2007: 83). It makes perfect sense, for example, to ask our woman of the opening example why she did what she did, even though it is clear that she acted in the spur of the moment. We often expect people to be able to come up with explanations and justifications for their actions, even when they did not deliberate on them beforehand: expectations that are regularly met. When I ask you why you forgot my birthday, for example, the reply that I did not think it was important to you makes perfect sense. Clearly the reason for that need not be that, prior to forgetting, I thought about whether me remembering matters to you. Also, there are several ways in which we can become and be aware of what makes us act in the way we do. This too is illustrated by several “implicit bias” experiments. While people in an experimental setting often respond with surprise to findings that show their implicit attitudes do not track their explicit ones, there are also findings that show that people can reliably predict discrepancies between their implicit and explicit attitudes when asked to predict differences between how they believe they should act and how they would act (Holroyd 2012: 292).20 Also, there are studies that show that actually people are quite capable of predicting their scores on tests measuring implicit attitudes toward different social groups (see, e.g., Hahn et al. 2013). Apparently, most of us, though we are not introspectively aware of harboring and being influenced by implicit bias in our everyday interactions with one another, are knowledgeable about the fact that we often fall short of our egalitarian commitments. Moreover, when put in certain experimental settings, we even seem able to become aware of our implicit attitudes and use this awareness to accurately predict them (Hahn et al. 2013). When we endorse a more sophisticated picture of what it entails to be a capable moral agent, we cannot interpret the literature on the mechanisms of implicit bias as establishing that we are not morally responsible for actions influenced by implicit biases. Also, other empirical findings show that the manifestation of implicit biases partly depends upon the activation of other goals, such as the need to enhance self-​ esteem (Holroyd 2012: 296; cf. Van Voorst Vader-​Bours In preparation).21 This shows, once again, that there is a relation between the manifestation of our implicit attitudes and other features of our motivational makeup. If our motivational makeup is related to 20  Holroyd discusses the work of Monteith and Voils (1998); their findings also show that even those who do not predict a discrepancy between how they should and how they would act manifest some bias. 21  Holroyd refers to the work of Uhlmann, Brescoll, and Machery 2010.

Moral Responsibility for Implicit Bias    677 our explicit attitudes, we cannot simply infer that we are completely passive regarding our implicit biases; hence it is wrong to think that we can never be morally responsible for them. On the basis of a combination of arguments like the above, Holroyd, like many other philosophers, conclude that there are at least some occasions in which individuals are “liable to blame for the extent to which they are influenced in behavior and judgment by implicit biases” (Holroyd 2012: 274; also see Brownstein 2015; Washington and Kelly 2016; Zheng 2016). To be sure, none of these authors argues that we are morally responsible for all manifestations of implicit biases and/​or that we are morally responsible to the same degree for their manifestation when we would reject the implicit biases underlying them if we were reflectively aware of them. The intuition that there is a moral difference between behavior such as the woman’s in our opening example and explicit racist acts is also widespread (cf. Zheng 2016: 68; see Levy 2014). What people disagree about is the extent to which we are morally responsible for implicit biases we do not identify with and/​or that we are not aware of and/​or that are not integrated at a deliberative level with other norms, values, and principles we accept (e.g., Levy 2014). To summarize, the literature on implicit biases, according to many, raises a question about our moral responsibility for them, especially when we would or do not endorse them upon reflection. However, that is not to say that the empirical literature shows, or is considered to show that we are never morally responsible for implicit biases. Moreover, some of those who argue that we are not attributively morally responsible also argue that we need other ways to address one another to correct for the harm done. They argue, for example, that we need to hold one another accountable,22 or subject one another to more nuanced forms of moral criticism to make progress toward more social equality (Saul 2013: sec. 4.2.7; Zheng 2016: 76).23 I come back to this view more elaborately in the third section. Note that when these authors claim that we might not be attributively morally responsible, though, what they mean is that our moral sentiments are not justified in the backward-​looking sense commonly required to establish our moral responsibility in the philosophical debate on whether we truly deserve the resenting and blaming responses of other agents. As I said in the introduction to this chapter, it is common in philosophical discussions to make a distinction between backward-​looking and forward-​looking reasons to hold someone morally responsible. The insight articulated just now, that we need ways to address one another for the harm done to make progress toward more social equality, typically is a forward-​looking reason. In the philosophical debate, forward-​looking reasons to justify holding another morally responsible are the characteristic of so-​called consequentialist views. Consequentialists views have a bad press, because they justify ascriptions of responsibility to individual agents unrelated to whether they actually 22  See previous note 11 on how this use of the distinction between attributability and accountability differs from what is common in the debate on moral responsibility and free will. 23  Zheng mentions Smiley 1992; Moody-​Adams 1994; Fricker 2007; and Springer 2007 (Zheng 2016: footnote 25).

678   Maureen Sie deserve them. According to many, it is the question of individual desert that makes us worry about the legitimacy of moral responsibility in the first place. If the fact that it promotes social equality or some other goal is a reason to hold someone morally responsible for her or his behavior, then why do we tend to believe it is unfair to hold someone morally responsible when she is not? It is beyond the purposes of this chapter to address these issues, but it is good to bear in mind that, unpopular as they are, forward-​looking reasons are regularly appealed to when discussing our moral responsibility for implicit biases. It is for forward-​looking reasons, I suspect, why the empirical literature finds its way to our society in the manner it does: IBs are often regarded as an explanation for the persistency of certain harmful stereotypes and prejudices that we have been combatting for decades. Hence, raising awareness of them, training ourselves to prevent their influence is considered required for finally winning the battle and finally realizing our egalitarian goals. Why this is considered urgent in society and philosophy can be found in that part of the IB-​literature that deals with their impact. I will turn to that next.

2.  The Social Dimension: The Impact of Implicit Bias In this section I distinguish between what I call “localized” and “general” harm. Examples of the former are phenomena such as shooter and evaluation bias, examples of the latter are phenomena such as microaggressions and inequities and epistemic injustice. Not much hangs on this distinction, except that whereas it is clear that we can and should take measures to counteract what I capture under the heading of “localized harms,” this is not the case for what I capture under the heading of “general harm” for several reasons. Hence, the only point I want to make with distinguishing these harms is that the lessons to draw from the literature on the harm of IBs as a whole are more unclear than might seem at first sight. That, in turn, has an effect on how to interpret what we can learn from the literature on the mechanisms of IB.

2.1  General Harm In the discussion of our moral responsibility for implicit bias, much attention is paid to phenomena that identify the harm done to people that belong to groups that have traditionally suffered from structural injustice and disadvantage, caused by, for example, stereotype threat, microaggressions, and micro-​inequities. Microaggressions are remarks, gestures, and responses much more minor than the tackling in our opening example, but which are nevertheless derogatory in some way.24 For example, asking people 24  For a critical discussion of the empirical research on the phenomenon of microaggressions and its application in our society, mainly the United States, see Lilienfeld 2017.

Moral Responsibility for Implicit Bias    679 with a non-​white appearance where they are “really from,” avoiding eye contact, and clutching one’s purse in interactions with people from certain social groups are cases in point. Although the question “Where are you really from?” can be intended as friendly interest, and the occasional awkward interactions in and of themselves do no physical harm, when people are constantly singled out (as possibly “not belonging here”), this might affect them negatively in numerous ways, especially if this goes together with being approached in ways that affirm the stereotypes and prejudices of the social group to which they belong. An example of the latter would be when a Dutch student of Moroccan descent is constantly praised for being well-​adapted in certain respects (speaking Dutch fluently, performing excellent), affirming the stereotype of that social group as not well-​adapted in these respects.25 This is also the case for the closely related phenomenon of micro-​inequities (e.g., Brennan 2016; Rowe 2008; Sue 2010). Micro-​ inequities are seemingly innocent ways in which people are treated differently on the basis of the social category to which they belong. For example, on average, women wait 20 seconds longer when they order a coffee than men, just as African Americans wait longer than white Americans (Myers et al. 2008, quoted by Brennan 2016: 241). The point of drawing attention to these inequities is not that longer waiting times are particularly harmful in themselves, but that differential treatment can turn up in very unexpected contexts and that, taken together and across all aspects of one’s life, microaggressions and inequities can add up. As a result, people who belong to a minority in certain contexts experience the world differently than those who belong to the dominant group (cf. Brennan 2016: 237) because of the structure of such interactions. Microaggressions and inequities, which are both debated phenomena as well, are discussed in the context of implicit bias because they are allegedly caused by implicit attitudes that influence our actions in ways that conform to certain stereotypes and prejudices. For example, when we need help with an administrative issue at a philosophy department, we might direct our question to the first female employee we see. We do so automatically and without prior deliberation, not because we explicitly believe that “women cannot be philosophers,” and hence she must be someone from the administrative office, but as a sort of reflex triggered by the feeling this is the right person to ask for help. Or, when preparing coffee for our customers, we might speed up just a little when the customer is a white male, not because of any manifest thoughts such as “This guy’s important, let’s not make him wait,” but because it feels slightly more important to be quick and adequate. Our implicit attitude toward white males might steer us toward serving them more promptly, for example. Microaggressions and inequities allegedly make it the case that people who belong to different social groups are interacted with in different ways.26 In certain situations, such as professional contexts, these differences can accumulate and cumulatively lead 25  Some might argue that this is a prejudice even, because it does not align with the facts about people belonging to this social group. 26  See aforementioned Lilienfeld 2017 for a critical review of the empirical research on microaggressions and the conceptual tensions the concept harbors.

680   Maureen Sie to a disadvantage for members that belong to the group that is stereotypically viewed as less fit for a certain profession (Cf. Valian 2005). In this sense, microaggressions and inequities might contribute to another phenomenon also often discussed in regard to implicit bias: that stereotypes can also influence our performance psychologically in a negative way (e.g., Saul 2013). This, controversial, phenomenon is called “stereotype threat”27 and refers to the finding that group awareness can negatively affect performance.28 For example, the existing reality of women underperforming in mathematics seems to be the result of the stereotype that they are bad at it, rather than a reflection of the stereotype’s initial legitimacy.29 Stereotype threat may also cause underperformance due to racial or age stereotypes. African Americans, for example, underperform in standardized tests when racial information is solicited prior to taking the test, thus making race salient to them (Steele and Aronson 1995).30 Likewise, the elderly underperform in memory tests when they are made to be aware of being elderly (Chasteen et al. 2005).31 Stacey Goguen has argued that stereotype threat has many more detrimental effects than just on our performance. For example, it can also cause things such as general anxiety, disengagement, and self-​doubt (Coguen 2016: 225–​259).32 Such effects contribute to a phenomenon Miranda Fricker raised awareness of in her influential book Epistemic Injustice (Fricker 2007). She lays bare how knowledge and power relate to one another at the level of the subject’s ability to understand herself, as well as at the level of her being understood by others.33 The phenomenon of epistemic injustice refers to power imbalances in our society that have led to discriminatory outcomes at both levels, offering unequal resources to different groups to understand themselves, what Fricker calls “hermeneutical injustice,” and appointing a different status to the knowledge contributions of people from different groups, which she calls “testimonial injustice” (Fricker 2007).34 I will come back to the considerable importance of the phenomenon 27  Some stereotypes have also been found to positively boost performance, see Shih et al. (2002). Since I focus on harm in this section, I restrict attention to stereotype threat. 28  It is controversial whether there exists a single phenomenon to which we can refer to as “stereotype threat.” The literature discussing this phenomenon actually postulates quite different underlying cognitive, affective, and motivational mechanisms to explain the effects of stereotypes on our performance. See for an helpful overview of the literature, e.g., Pennington et al. 2016. 29  Research suggests that gender differences in mathematics performance in the general population are actually trivial; see Hyde et al. 2008. 30  The underperformance only occurred in so-​called threat condition (in which Black participants in the test were told that the test was diagnostic of ability). In a so-​called non-​threat condition (in which they were told that the test was simply problem solving, not diagnostic of ability) there was no apparent risk of fulfilling the racial stereotype about their intellectual ability, and underperformance did not occur. 31  This paragraph was taken from Sie and Van Voorst Vader-​Bours 2016: 98–​99. 32  For a critical discussion of “Stereotype Threat–​paradigm,” see Blum 2016. 33  Also see Charles Taylor who has argued that because of the fact that we develop our identity in dialogue with (our significant) others, people that belong to a cultural minority face serious problems of recognition (Taylor 1994). 34  See Alcoff et al. (2010) for a discussion of epistemic injustice and Bohman and McCollum (2012) for a discussion of epistemic injustice in new domains.

Moral Responsibility for Implicit Bias    681 of epistemic injustice when discussing moral responsibility, in the last section of this chapter. Although the empirical literature on phenomena such as stereotype threat, microaggressions and inequities, and epistemic injustice is debated, it managed to unsettle many by its suggestion that stereotypes and prejudices cause harm in ways we have not realized beforehand.35 The widespread patterns of behavior they appear to bring to the surface are perceived to be all the more worrying because they tend to disadvantage people who belong to social groups that have suffered a history of abuse and systematic injustice. I come back to this in section 3. Before turning to that, let me first continue with discussing some phenomena generally discussed alongside those hitherto discussed, but distinct from them because they are relatively easy to pinpoint and because it is much more clear what our egalitarian commitments recommend when faced with them. I refer to the harm done by these phenomena as “localized harm.”

2.2  Localized Harms In the case of localized harms, it is relatively clear that we should do all we can to correct for the harm done, provided that we agree on the findings.36 Take, for example, the much discussed “shooter bias” related to the stereotype that Black people are aggressive and white people are not.37 This bias allegedly leads to the misidentification of the objects Black/​white males are carrying.38 In a simulated shooting task in a laboratory setting in which participants saw videotaped faces of men who were either holding a gun or a neutral object, such as a wallet or a cell phone, more unarmed Black men were “shot” because it was mistakenly supposed that the objects in their hands were guns. White people who were actually holding a gun were more often perceived as carrying a wallet and, hence, were not shot (e.g., Correll et al. 2002; Greenwald et al. 2003; Payne 2001). Clearly, shooter bias might lead to disastrous consequences, mainly for Black people, as is unfortunately illustrated outside the lab by numerous incidents of unarmed people of color39 being shot in streets, parks, train stations, and so on, because

35  See the debated nature of the literature the aforementioned critical studies, e.g., Forscher et. al 2017 and Lilienfeld 2017. 36  The lines on shooter bias that follow are taken from Sie and Van Voorst Vader-​Bours 2016: 100. 37  Shooter bias, as is the case for many of the biases discussed in the literature, is a bias disclosed by experimental studies in the United States. It is not clear whether the findings from these studies generalize to other parts of the world, though there are also European studies investigating shooter bias in relation to, for example, Muslim headgear; see note 39 and also Unkelback et al. 2009. 38  Although the findings on the shooter bias are often taken as settled in the philosophical discussion, recently some articles were published reporting on findings contradicting the original studies. See, for example, James et.al. 2016. 39  As noted above, it is not only color of skin that might put you in danger. Islamic appearance and Islamic head-​dress appear to have the same effect (also see Unkelbach et al. 2008).

682   Maureen Sie of a misidentification of what they were actually holding in their hands.40 What is also clear is that this phenomenon runs counter to our egalitarian commitments, our commitments to non-​discrimination. Another, much discussed, example of a localized harm that is less dangerous but still seriously worrying is that implicit bias influences our perceptions and evaluations in a negative way when people do not fit the stereotype that belongs to certain professions. According to some debated studies, the ethnicity or gender suggested by the name on a letter of application or CV, for example, influences its evaluation even when the evaluation is done by those who are committed egalitarians and/​or belong to the same social group (Saul 2013: sec. 1.1; Steinpreis et al. 1999).41 Other findings in the same domain suggest that publication tracks and teaching performance within academia are also evaluated more favorably in the cases of male than of female professors (e.g., Moss-​ Racusin et al. 2012). Several case studies allegedly show that implementing anonymous reviewing or evaluating procedures leads to a more balanced representation. One often-​ cited study is taken to show that the percentage of women in major symphony orchestras increased 20% since it became standard to audition anonymously (Goldin and Rouse 2000).42 In the literature, other experiments and case studies are also discussed, arguing that phenomena like implicit bias and stereotype threat might be a partial explanation for the underrepresentation of certain groups, for example, women in philosophy (Saul 2013).43 What distinguishes these localized harms from the more general ones (besides the fact that they are more easy to pinpoint and address) is that it is clear how we should respond to them. People can disagree about whether the findings do indeed establish these practices (of policing and professional evaluation) to be discriminatory, but once we agree on the findings, our egalitarian commitments entail taking institutional and other measures to correct for them.44 For example, we can make use of institutional and psychological measures identified as successful in counteracting the manifestation of implicit biases, such as using anonymity and/​or provide training for those at risk of causing the most harm (e.g., people on evaluation committees or those patrolling our 40 

Cf. the Amadou Diallo case in New York City in 1999—​see, for example, Kelly and Roedder (2008: 256)—​or the Jean Charles de Menezes case in the aftermath of the London bombings of July 2005. 41  Sean Hermanson criticizes Saul and other philosophers who have put IB on the philosophical agenda for misrepresenting what we can conclude from the empirical literature and critically discusses some of the studies that keep recurring in the discussion, such as the Steinpreis et al. study. According to him the results of this study are not at all convincing, more-​over, they have not been corroborated by other studies. Hermanson 2017: p 6. 42  Whether we can actually infer that can be doubted when we look at the details of this study, because there are “some odd patterns in the data” and some “anomalous results” as well. See Hermanson, 2017: 4–​5; according to Hermanson, conceded by the authors of that study itself on p. 734. 43  For criticism of this claim, see Ceci and Williams 2011 and the aforementioned Hermanson who actually accuses philosophers of pushing a political agenda in their discussion of the IB-​literature. For a reply to the objection formulated by Ceci and Williams, see Lee 2016. 44  Although we might also disagree on which measures should be taken, especially if these measures itself require us to transgress other egalitarian principles or values.

Moral Responsibility for Implicit Bias    683 streets).45 Moreover, believing such measures to be called for does not require us to solve the underlying vexing issue of whether we are morally responsible for implicit biases in the backward-​looking sense. Now let me explain why I believe it is so much less clear what our egalitarian commitments entail in the case of the more general harm discussed in the previous subsection.

2.3  Counteracting General Harm In the previous subsection, I distinguished between localized and more general harm. I observed that with respect to localized harm, it is clear that our egalitarian commitments obligate us to do what we can to make our practices less discriminatory. However, that message is limited to clear and well-​established cases, cases established beyond a shadow of reasonable doubt. It is unclear how to counteract or remedy the harm allegedly done by such phenomena as micro-​inequities, aggressions, stereotype threat, and epistemic injustice, even if we are convinced that the scientists did establish the existence of these phenomena as well as their harm. These phenomena are hypothesized to be the result of the myriad ways in which implicit biases affect our face-​to-​face interactions. That observation raises many difficulties with respect to determining how to respond to them (cf., Antony 2016: 166). The problem is that these phenomena relate awkwardly to one another and/​or perhaps even with those of more localized harm, such as evaluation bias. Part of the harm allegedly caused by implicit biases is underperformance in certain domains. This means that we cannot simply correct for these biases as if they were only epistemically distorting (Antony 2016: 167). Perhaps, implicit biases (partly) cause harm via undermining the performance of people belonging to certain groups. Explaining more precisely what implicit biases are, how they operate, and what their role is in our cognitive architecture raises questions about whether it is possible to single out those we should and could get rid of to make our practices more egalitarian (cf. Antony 2016; Huebner 2016; Sie and Van Voorst Vader-​Bours 2016: 93–​102).46 Moreover, although the discussion about our moral responsibility for implicit bias understandably focuses on the wrong done to people who belong to traditionally underrepresented and disadvantaged groups, implicit biases affect everyone, not only those who belong to those groups. For example, people who take up or aspire to a role that does not fit the stereotype or prejudices of the social group they belong to might also be harmed by interactions steered by implicit biases (see Van Voorst Vader-​Bours and Sie 2016: 99–​100). A full-​time father might be responded to differently than a full-​time mother. In this sense, stereotypes and prejudices affect us all, steering us toward roles 45 

Whether such training is effective is subject to controversies, see Forscher 2017. As mentioned several times already, the aforementioned meta-​study suggests that we need more studies to establish the exact relation between implicit biases and our explicit attitudes and behavior before we can draw any clear conclusions (Forscher 2017). This corresponds with what many philosophers have been arguing for with respect to the literature on implicit bias. 46 

684   Maureen Sie and ways of behaving that fit the stereotypes and prejudices of the group we belong to. Also, it is not only people from dominant and privileged groups that harbor and are influenced by implicit biases. Tragically, if the literature is correct, also people for underprivileged groups harbor and are influenced by implicit biases.47 In this sense, with respect to the general harm done by implicit biases there is no simple opposition between victims and perpetrators (cf. Van Voorst Vader-​Bours and Sie 2016: 101–​102). This is not to deny that it is clear who are harmed most, but to point out that even when that is the case it is unclear who to address exactly.48 Lastly, the general harm done by implicit biases arises at the aggregate level. Taken as a single event, complimenting someone with their excellent command of the Dutch language, for example, or perhaps even tackling someone to prevent a theft, do not stand out as moral wrongdoings. Becoming aware of the fact that at an aggregate level they might become harmful might be the reason why the literature on the impact of implicit biases makes such an impression on us. Suddenly we learn that (1) our inclination to categorize one another socially might thoroughly affect all our interactions with one another and (2) might cause harm at an aggregate level that is substantial, even though the individual actions we perform at an everyday level, in and of themselves, seem innocent enough. If I am right in this, that it is that lesson that unsettles us and that should unsettle us, this might also shed a new light on how to interpret the literature on the mechanisms of implicit bias discussed in the first section of this chapter. Let me explain this. As we saw, many philosophers take this literature to suggest that we are less egalitarian than we think because it suggests, notwithstanding our egalitarian commitments, we are thoroughly influenced by implicit biases. But when the literature on the impact of implicit biases unsettles us or we conclude that it should unsettle us, this shows that we have not been aware of (1) how thoroughly implicit biases affect even the most mundane of our everyday interactions and (2) that this can cause considerable harm at the aggregate level. Hence, this can also be taken to show that we have never realized exactly what it entails to be “committed egalitarians” if this is taken to include non-​discrimination at all levels of human interaction. Moreover, if this is what it is taken to entail, some might rethink their commitment because the requirement to police even the tiniest of our mundane interactions seems like a very high price to pay for a more egalitarian society. When there is room for disagreement about our egalitarian commitments or disagreement on what it requires of us, this also has consequences for what to learn from current studies on the mechanisms of implicit biases. As we saw in the section on this literature, it suggests that the degree of implicit biases we manifest, as well as their influence on our behavior, is related to things such as the nature of our commitment to egalitarianism (extrinsic or intrinsic) and our other mental (motivational and cognitive) states. If we 47 

Hence, the woman in our opening example, when she learns about the IB-​literature might come to understand that her anger of being accused of biased behavior is not warranted. See footnote 4. 48  An observation that becomes highly relevant when research results are used to devise and implement policies. It would be a mistake, for example, to think that it is only people from dominant and privileged groups that should be addressed by these policies.

Moral Responsibility for Implicit Bias    685 were not fully aware of the impact of implicit biases and/​or their widespread nature, our commitment to egalitarianism might have been somewhat superficial to begin with and/​or might be a commitment we can and do not want to live up to once we realize what it would require of us. Even though the jury is not in on whether the empirical research establishes that (1) implicit biases exist, (2) what they are exactly, and (3) whether we have indeed found ways to measure them, what clearly surfaces from the literature as a whole is that we are motivationally complex beings. Our implicit and explicit attitudes interact with one another in interesting ways, an insight illustrated by other empirical literature on the many influences that escape our awareness as well.49 Under the influence of the literature on the impact of implicit biases discussed in this section, the nature of our commitments and other mental states might change. Once we become convinced that the gravity and harm the tiniest of our interactions can lead to, or think it likely that this is the case, we might start to believe that we should change our ways. If that would be the case, this might affect future measurements of our implicit attitudes. Moreover, it might also have consequences for whether we believe our appraising responses to one another are justified. In the next section, let me draw attention to an argument of Cheshire Calhoun to explain why people might think this is the case and contrast it with what has been the dominant approach in the discussion of our moral responsibility for implicit bias.

3.  Abnormal Moral Contexts and the Point of Moral Responsibility In the previous two sections, I have discussed the IB-​literature as in part concerned with their mechanisms and in part concerned with their impact. A backward-​looking approach, an approach that focuses on backward-​looking reasons for moral responsibility, has been dominant in the discussion of the IB-​literature. Nevertheless, as we saw in the first section, many eventually appeal to forward-​looking reasons as well, to defend that we need to respond to biased behavior. In the second section, I elaborated on the alleged harm the IB-​literature draws our attention to. I have suggested that it is the structural and considerable harm that explains the appeal to forward-​looking reasons in the debate on our moral responsibility for implicit biases. In this section, I shortly elaborate on the two most influential views that take the backward-​looking approach and continue with sketching a forward-​looking approach to the issue of our moral responsibility for implicit bias, that is, that of Cheshire Calhoun. I end with adding my own reflections to this forward-​looking approach. 49  With “the other insights illustrated by other empirical literature,” I refer to the large body of research that is often discussed under the heading of “dual process theory of cognition” and/​or the literature on the “adaptive unconscious.”

686   Maureen Sie

3.1.  Attributionism and Volitionism In the first section, I outlined several arguments for thinking that we are morally responsible for biased behavior regardless of their implicit nature, all of which connect up with more general arguments endorsed in the literature with respect to our moral responsibility for actions that are not the result of explicit prior deliberative processes. Several philosophers have elaborated on these arguments in the past decade. For example, a much discussed view is that of Angela Smith, who has defended a view called attributionism, appealed to by, for example, aforementioned Brownstein and Zheng (Smith 2005, 2008; Brownstein 2015, Zheng 2016). According to attributionism, actions or attitudes can be attributed to an agent whenever they reflect their rational judgment, and they can reflect these judgments even without prior deliberation or choice. The reason for this is that most of our emotions, desires, and beliefs are sensitive to our general evaluations, “judgment dependent” as she calls it, without being under our direct volitional control. What we notice and are aware of and what not, “reflects our judgments about what things are important or significant” (Smith 2008, 370). According to Smith, we rightfully expect one another to be able to justify and explain actions that disclose attitudes, desires, and beliefs when they can be attributed to us. And that they can be attributed to us does not require us to be in direct control of them. It is sufficient that they are indicative of our “relatively stable evaluative tendencies” (Smith 2008: 251). Smith elaborates her view in part as a response to what is called “volitionism.” Volitionism, in contrast to attributionism, holds that conscious choice is a necessary condition for moral responsibility.50 In his turn, Neil Levy, one of the defenders of volitionism in the discussion on our moral responsibility for implicit bias (Levy 2005), has responded to Smith’s attributionism by arguing that it cannot account for the distinction between expressing an attitude and “merely reflecting an attitude.” According to Levy, it is only in the case that our actions express our attitude that we can claim that it discloses our evaluative tendencies in a way that legitimizes ascriptions for responsibility (Levy 2016). In order for us to make a distinction between expressions of our attitudes and mere reflections, according to Levy, we need to revisit the role of what I have called “deliberative awareness” in the first section of this chapter. Levy’s objection to Smith makes sense, even if we believe deliberative awareness, control, and reasons-​responsiveness cannot simply be taken for granted as conditions for moral responsibility. When we revisit the example with which we opened this chapter, how the women responds to allegations of the teenager that her behavior is biased is bound to make a difference to our evaluation of her behavior corresponding to the distinction drawn just now between reflecting and merely expressing her attitudes. Suppose that she just shrugs her shoulders in response to the teenager’s anger, communicating 50 

For criticism of volitionism, also see aforementioned Pettit 2007.

Moral Responsibility for Implicit Bias    687 that she does not really understand or care about his allegations. This suggests her behavior does not only express an attitude, but also reflects it. Would she, alternatively, respond with anger or horror to his allegations, this suggests her behavior might perhaps express her attitudes, but not reflect it. If her initial action reflects her attitude rather than merely express it, this is bound to increase our resentment, blame, or moral indignance about it, that is, when we believe that the initial action is harmful and that we are under some prima facie obligation to prevent such harms. This, as I have discussed in the second section of this chapter, might not be as uncontroversial as it is often taken to be. Before discussing where that observation leaves us, let me first sketch the forward-​ looking approach to our moral responsibility for implicit bias with the help of an argument by Cheshire Calhoun.

3.2  Abnormal Moral Contexts In her essay “Responsibility and Reproach,” Calhoun argues in favor of ascriptions of moral responsibility in the case of racist and sexist acts, even when we might doubt that the perpetrators are in fact morally responsible for those acts (Calhoun 1989: 389). In that essay, Calhoun discusses the feminist predicament as one in which the two approaches, distinguished above, pull us in opposite directions. Feminists, according to Calhoun, understand that the social acceptance of certain ways of behaving impede an individual’s awareness that what she or he is doing is wrong. This observed lack of awareness causes us to judge that we cannot really hold the individual agent morally responsible. However, especially in contexts in which behavior that is in fact morally wrong is widely accepted—​contexts she refers to as “morally abnormal”—​we need to hold people morally responsible because it educates them about the wrong done, motivates them to change their ways, and communicates to those who suffer from the wrongdoing that they are right to feel wrongfully treated (Calhoun 1989: 405).51 In these contexts, Calhoun argues, “the point of moral responsibility” and its justification come apart (cf. Zheng 2016: 76).52 In those situations there are people that might not be deserving of blame, but who are appropriately held morally responsible in a forward-​looking sense, for forward-​looking reasons.53 Whether or not an 51 

In the broader philosophical discussion, so-​called scaffolding accounts emphasize the communicative, motivating, and educating aspects of our practices of moral responsibility. They use those aspects to argue against the focus on the metaphysical issue of free will. (See, e.g., Vargas 2013; McGreer 2010, 2015). 52  What Calhoun refers to as the “point of responsibility” is close to what I tend to refer to as the “communicative and coordinative” purposes of our practices of moral responsibility (see, e.g., Sie 2013 and 2014). Thanks to Catriona McKenzie for drawing my attention to this fact. Unlike Calhoun I do not think this point can be distinguished from its justification only in abnormal moral contexts, as I explain in Sie 2018. 53  Thanks to Dana Nelkin for suggesting this formulation of Calhoun’s point.

688   Maureen Sie individual agent can (fully) help acting in the way she or he does is less relevant in such situations, if at all, because responding to her or him with the moral sentiments (Calhoun mainly discusses reproach) is required to raise awareness of the unjust situation that we are in and to combat social oppression, structural injustices, and disadvantages. Part of the discussion on our moral responsibility for implicit bias, the reason why it gets the amount of public and philosophical attention it gets and why it worries people when they are first confronted with it, might be informed by the suspicion that our situation is still abnormal in Calhoun’s sense. The suspicion that our situation is abnormal gains support from the phenomena discussed in the previous section under the heading of the general harm implicit biases cause. As we have seen, these harms allegedly thoroughly affect all our interactions, thus suggesting that it is not only in rare situations and contexts that implicit biases make us act in ways contrary to our egalitarian commitments. Discussing localized and general harms under one heading of “harm done by implicit biases” suggests that implicit biases not only incidentally lead to grave and clear harm, but that this harm is widespread and deeply engrained in our everyday ways of interacting with one another. Thus the suspicion can easily arise that the implicit bias literature shows our society to be far removed from the egalitarian society that some of us imagine it to be and that the reason for this is that each of us is less egalitarian than we might have hoped. As mentioned in the previous section, I suspect it is partly as a result of this that a lot of the, still debated, IB-​literature discussed in this chapter has already found its way to our society in training programs and other institutional policies.54 However, it is good to bear in mind that whether our situation is actually abnormal in Calhoun sense, that is, a situation in which behavior that is in fact morally wrong is widely accepted, is also debated. It is beyond the purposes of this chapter and my expertise to address that issue. I believe that addressing it properly requires us to turn to feminist philosophy, disability studies, and the philosophy of race and gender (See, e.g., Anderson 2010; Haslanger 2015). It also requires us to address and discuss more thoroughly the details of the empirical studies presented in this chapter and appealed to in the philosophical discussion. We should not overlook or hide the fact that the IB-​ literature is debated and several meta-​studies and more critical discussions of the results have been put forward in the past decade. That being said, even when we do not believe our context is morally abnormal in Calhoun’s sense and agree that the IB-​literature is contestable in a number of ways, there is ample reason to take an interest in it and to consider our moral responsibility in its light. It is in order to make this point that I have structured this chapter in the way I did. Before concluding, let me make explicit why we have a reason to take interest in this literature independent of our stance on the many controversial issues it is riddled with.

54 

For examples of such measures and policies in the domain of microaggressions see, for example, Lilienfeld 2017.

Moral Responsibility for Implicit Bias    689

3.3  The Point of Moral Responsibility There are two reasons why the IB-​literature should interest philosophers interested in moral responsibility, despite the controversies about how to interpret it exactly. Firstly, the literature on the mechanisms of implicit biases, when considered in tandem with the more general discussion on conditions of moral responsibility in response to the plethora of findings in cognitive science about the role of deliberative awareness, prior awareness of the reasons for which we act, clearly establishes that we are motivationally complex beings. That is, it establishes that our explicit attitudes, our motivations, and aims interact in interesting ways with measurable implicit attitudes. Whether these implicit attitudes testify to the existence of implicit biases, biases that enable us to predict behavior and that can be targeted to change it as well (debated in the aforementioned critical meta-​study Forscher et.al. 2017), is an important and interesting issue, but not the only thing that makes the studies referred to in the first section of this chapter relevant to our reflections on moral responsibility. The insight that our explicit attitudes interact in several ways with our implicit ones, in itself, is an important insight that should rightfully make us nervous about countering specific allegations of prejudiced and stereotyping behavior too readily and with too much confidence on the basis of our explicit egalitarian commitments and/​or intentions. Secondly, although the literature on the impact of implicit biases is debated as well, and does not offer us guidelines on how to respond to them in the case of the general harm they allegedly cause, it does establish that our inclination to categorize one another socially is deeply rooted and might cause considerable harm in ways that are not immediately clear. This means that we have a lot to sort out with respect to what we can and want to expect of one another where this behavior is concerned, such as that of our opening example.55 That is, we have to sort out (1) how to weigh the considerable harm possibly done by implicit biases and (2) whether and to what extent people will be able and expected to steer their behavior in line with their evaluations in this respect. Calhoun’s argument makes clear that holding one another morally responsible might be required for doing so. Our moral sentiments, or their watered-​down equivalents of moral criticism, frowns, resenting or indignant irritation and frustration, communicate that we believe or suspect such actions to transgress legitimate expectations in the domain of social interactions. In the case of implicit biases, as should by now be clear, such communication is required because it might easily escape our attention that single actions can contribute to serious harm and that our explicit egalitarian attitudes are no guarantee for treating others in a fair and just manner. Hence, I agree with Brownstein (see opening quote at the beginning of this chapter) that everyone concerned with

55  Moreover, simplifying what the experimental work establishes or jumping to conclusions not established by this research is risky, attractive as it might be for those of us who finally see their philosophical intuitions and views backed up by psychological research. Such a strategy might backfire when the research fails to withstand further critical examination.

690   Maureen Sie fairness and justice should take an interest in the literature on implicit bias. However, I believe that figuring out what our egalitarian commitments should entail exactly, and who’s responsible for transgressing them, is something that we need to do in our everyday practices, that is, by interacting with one another as morally responsible agents. Familiarity with the IB-​literature might considerably smoothen that process because it brings to our attention, first of all, that tiny seemingly innocent interactions can contribute to considerable harm. Secondly, that all of us, at times, might behave in implicitly biased ways. When institutions take measures to counteract the harm I have referred to as “localized,” they believe, established on the basis of empirical research, that they show their egalitarian commitments in practice and their ability to change their ways. Similarly, the way in which we respond to criticism, resentment, and blame as a response to actions that are suspected to contribute to considerable harm will show the extent of our egalitarian commitments and/​or our ability to change our ways. However, we should bear in mind that in the case of actions that allegedly contribute to what I have referred to as “general” harm, it is unclear what these commitments entail or should entail. Public discussions on the topic of stereotypes and prejudices, more generally, suggest that many people believe we are too uptight and obsessed with what has become known under the general label of “political correctness.” Hence, it is not only the case that our ideas on what our egalitarian commitments entail exactly is unclear, it might turn out that we disagree on what they should entail as well. It might be partly due to the possibility and real existence of such disagreements that the process of changing our practices is such a difficult, painstaking, and slow process. However, there is more that we can take from the IB-​literature as I have discussed it. If I am right that with respect to implicit biases we need to figure out together what we can and should expect of one another in the domain of social categorization, the IB-​literature as I discussed it in this chapter provides yet a further reason for concern. A concern related to the phenomenon of epistemic injustice discussed in section 2.1. In order for this “figuring out together” to lead to an outcome that fairly represents all involved, everyone must be regarded as an equal participant in this process. What the literature on the impact of implicit biases suggests, though, is that this equality is not realized in most settings. This adds a further harm to those who are not regarded and responded to as equal participants to our moral practices: it will be more difficult for them to leave their mark on our shared moral practices by changing these practices in ways they consider desirable. The phenomenon of epistemic injustice has skyrocketed in many philosophical discussions, as has the phenomenon of implicit biases, because many intuitively grasp that such phenomena affect our practices of moral responsibility at their very core, potentially distorting their whole point. When someone who objects to being treated in a certain manner, or responding to it with anger, is not taken seriously in the same way or to the same degree as others are taken, her or his objection or anger will not contribute to the same extent in changing these practices as those of others. Moreover, even our appraising responses itself, considered by many

Moral Responsibility for Implicit Bias    691 indispensable to improve our practices, are bound to be influenced by all kinds of biases as well (cf. Nadelhoffer 2008).56 Given the literature on the impact of implicit bias hitherto discussed, it is only to be expected that whether an outcome of a certain action or chain of actions is evaluated as bad, unwanted, or sad, itself will also be influenced by implicit biases in the domain of social categorization, that is, that it will depend on what social group the perpetrator and/​or victim belong to. The empirical literature not only gives us cause to consider the harm implicit biases do and the need to figure out how to weigh that harm exactly, it should also cause us to be worried that they interfere with the proper functioning of our moral responsibility practices, or, to phrase it as Calhoun does, with the very point these practices serve.

4.  Conclusion The literature on implicit bias gives us ample reason to be extremely vigilant with respect to our widespread inclination to categorize one another socially and shows there is much to sort out with one another regarding what it is that our egalitarian commitments commit us to exactly, vis-​à-​vis the literature on the mechanisms and impact of implicit bias. Responding to one another with the moral sentiments, including the more nuanced forms of moral criticisms, such as frowns, and communicating our observations on the possibly biased nature of our perceptions and evaluations, as Calhoun has pointed out, might be crucial to start doing so. In the cases of localized harms, it seems clear what our egalitarian commitments entail, but this is not so for general harms. In the case of actions contributing to more general harms, we need our moral appraisals to express that we think wrong has been done, that we believe it is reasonable to expect one another to behave better in the situation concerned, and to figure out whether the person appraised agrees with us that this is the case. What we consider reasonable to expect from one another in those cases will partly depend on how we evaluate the harm discussed in the second section of this chapter and the extent to which our situation is still abnormal in Calhoun’s sense. A further complication of sorting out what we consider to be reasonable to expect from one another is that it requires us to treat all involved as equal participants in this process. Since the literature on implicit bias establishes that our inclination to categorize one another socially is deeply rooted and widespread, that is a much more challenging task than we might have thought.

56 

For example, when we perceive the outcome of a certain action or chain of actions as bad, unwanted, or extremely sad, we will judge the agent more harshly than we would if that exact same action or chain of actions would have had, in our eyes, a less bad, unwanted, or sad outcome.

692   Maureen Sie

References Alcoff, L., S. Goldberg, C. Hookway, and M. Fricker. 2010. “Book Symposium: Miranda Fricker’s Epistemic Injustice: Power and the Ethics of Knowing.” Episteme: A Journal of Social Epistemology 7 (2): 128–​179. Allport, G. W. 1954. The Nature of Prejudice. Cambridge, MA: Addison-​Wesley. Amodio, D. M. and S. A. Mendoza. 2010. “Implicit Intergroup Bias: Cognitive, Affective, and Motivational Underpinnings.” In Handbook of Implicit Social Cognition, edited by B. Gawronski and B. K. Payne, 353–​375. New York: Guilford Press. Anderson, E. 2010. The Imperative of Integration. Princeton, NJ: Princeton University Press. Antony, L. M., 2016. “Bias: Friend or Foe? Reflections on Saulish Skepticism.” In Implicit Bias and Philosophy: Volume I, Metaphysics and Epistemology, edited by M. Brownstein and J. Saul. Oxford: Oxford University Press. Arpaly, N. 2004. Unprinciples Virtue: An Inquiry Into Moral Agency. New York: Oxford University Press. Banaji, M. and C. Hardin. 1996. “Automatic Stereotyping.” Psychological Science 7: 136–​141. Blum, L. 2016. “The Too Minimal Political, Moral, and Civic Dimension of Claude Steele’s ‘Stereotype Threat’ Paradigm.” In Implicit Bias and Philosophy: Volume 2, Moral Responsibility, Structural Injustice, and Ethics, edited by M. Brownstein and J. Saul. Oxford: Oxford University Press Bohman, J. and J. McCollum (eds.). 2012. “Special Issue: Epistemic Injustice.” Social Epistemology: A Journal of Knowledge, Culture and Policy 26 (2). Brennan, S. 2016. “The Moral Status of Micro-​inequities: In Favor of Institutional Solutions.” In Implicit Bias and Philosophy: Volume 2, Moral Responsibility, Structural Injustice, and Ethics, edited by M. Brownstein and J. Saul. Oxford: Oxford University Press. Brownstein, M. 2015. “Implicit Bias.” In Stanford Encyclopedia of Philosophy, edited by E. Zalta, M. Brownstein, and J. Saul. Oxford: Oxford University Press Brownstein, M. and J. Sauls (eds.). 2016a. Implicit Bias and Philosophy: Volume I, Metaphysics and Epistemology. Oxford: Oxford University Press. Brownstein, M. and J. Saul (eds.). 2016b. Implicit Bias and Philosophy: Volume 2, Moral Responsibility, Structural Injustice, and Ethics. Oxford: Oxford University Press Brownstein, M, N. Beard, K. Frankish, J. Holroyd, N. Levy, E. Machery, A. Madva, S. Spaulding, and C. Sripada. 2017. What Can We Learn from the Implicit Association Test, A Brains Blog Roundtable, http://​philosophyofbrains.com/​2017/​01/​17/​how-​can-​we-​measure-​implicit-​ bias-​a-​brains-​blog-​roundtable.aspx, consulted on 2 July 2018. Calhoun, C. 1989. “Responsibility and Reproach.” Ethics 99: 389–​406. Ceci, S. J. and W. M. Williams. 2011. “Social Sciences—​Psychological and Cognitive Sciences: Understanding Current Causes of Women’s Underrepresentation in Science.” PNAS 108 (8): 3157–​3162. Chasteen, A. L., S. Bhattacharyya, M. Horhota, R. Tam, and L. Hashe. 2005. “How Feelings of Stereotype Threat Influence Older Adults Memory Performance.” Experimental Aging Research 13 (3), 235–​260. Clark, T. W. 2013. “Experience and Autonomy: Why Consciousness Does and Why It Doesn’t Matter.” In Exploring the Illusion of Free Will and Moral Responsibility, edited by G. Caruso. New York: Lexington Books, Rowman & Littlefield. Coguen, S. 2016. “Stereotype Threat, Epistemic Injustice, and Rationality. In Implicit Bias and Philosophy, Volume 1, Metaphysics and Epistemology, edited by M. Brownstein and J. Saul, 216–​237. Oxford: Oxford University Press.

Moral Responsibility for Implicit Bias    693 Correll, J., B. Park, C. M. Judd, and B. Wittenbrink. 2002. “The Police Officer’s Dilemma: Using Ethnicity to Disambiguate Potentially Threatening Individuals.” Journal of Personality and Social Psychology 83 (6): 1314–​1329. Devine, P. G. and E. A. Plant. 2002. “The Regulation of Explicit and Implicit Race Bias: The Role of Motivations to Respond Without Prejudice.” Journal of Personality and Social Psychology 82 (5): 835–​848. Faucher, L. 2016. “Revisionism and Moral Responsibility.” In Implicit Bias and Philosophy: Volume 2, Moral Responsibility, Structural Injustice, and Ethics, edited by M. Brownstein and J. Saul. Oxford: Oxford University Press, 115–​146. Forscher, P. S., C. K. Lai, C. R Ebersole, M. Herman, P. G. Devine, and B. A. Nosek. 2019. A meta-​analysis of procedures to change implicit measures. Journal of Personality and Social Psychology, 117(3), 522–​559 Frankish, K. 2016. “Playing Double: Implicit Bias, Dual Levels, and Self-​Control.” In Implicit Bias and Philosophy: Volume I, Metaphysics and Epistemology, edited by M. Brownstein and J. Saul, 23–​46. Oxford: Oxford University Press. Fricker, M. 2007. Epistemic Injustice: Power and the Ethics of Knowing. Oxford: Oxford University Press. Galinski, A. D. and G. B. Moskowitz. 2000. “Perspective-​Taking: Decreasing Stereotype Expression, Stereotype Accessibility, and In-​Group Favouritism.” Journal of Personality and Social Psychology 78 (4): 708–​724. Glaser, J. and E. D. Knowles. 2008 “Implicit Motivation to Control Prejudice.” Journal of Experimental Social Psychology 44: 164–​172. Goldin, C. and C. Rouse. 2000. “Orchestrating Impartiality: The Impact of ‘Blind’ Auditions on Female Musicians.” The American Economic Review 90 (4): 715–​741. Greenwald, A. G., M. R. Banaji, and B. A. Nosek. 2003. “Understanding and Using the Implicit Association Test: I. An Improved Scoring Algorithm.” Journal of Personality and Social Psychology 85 (2): 197–​216. Hahn, A., C. M. Judd, H. K. Hirsh, and I. V. Blair. 2013. “Awareness of Implicit Attitudes.” Journal of Experimental Psychology: General 143(3), 1369–​1392. Doi:10.1037/​a0035028. Haslanger, S. 2015. “Social Structure, Narrative, and Explanation.” Canadian Journal of Philosophy 45 (1):1–​15. DOI: 10.1080/​00455091.2015.1019176. Hermanson, S. 2017. Implicit Bias, Stereotype Threat, and Political Correctness in Philosophy. Philosophies 2: 12. DOI: 10.3390/​philosophies2020012. Hieronymi, P. 2004. “The Force and Fairness of Blame.” Philosophical Perspectives 18 (1): 115–​148. Holroyd, J. 2012. “Responsibility for Implicit Bias.” Journal of Social Philosophy 43 (3): 274–​306. Holroyd, J. and J. Sweetman. 2016. “The Heterogeneity of Implicit Biases.” In Implicit Bias and Philosophy: Volume 1, Metaphysics and Epistemology, edited by M. Brownstein and J. Saul. Oxford: Oxford University Press. Horstkötter, D. 2014. “Self-​Control and Normativity: Theories in Social Psychology Revisited.” Theory and Psychology 25 (1): 25–​44. Huebner, B. 2016. “Implicit Bias, Reinforcement Learning, and Scaffolded Moral Cognition.” In Implicit Bias and Philosophy, Volume 1, Metaphysics and Epistemology, edited by M. Brownstein and J. Saul, 47–​79. Oxford: Oxford University Press. Hyde, J. S., S. M. Lindberg, M. C. Linn, A. B. Ellis, and C. C. Williams. (2008). “Gender Similarities Characterize Math Performance.” Science 231 (5888): 494–​495. James, I., S. M. James, and B. J. Villa. 2016, “The Reverse Racism Effect. Are Cops More Hesitant to Shoot Black Than White Suspects?” Criminology and Public Policy15 (2): 457–​479.

694   Maureen Sie Jost, J. T. and M. R. Banaji. 1994. “The Role of Stereotyping in System-​Justification and the Production of False Consciousness.” British Journal of Social Psychology 33 (1): 1–​27. Kang, J. and K. Lane. 2010. “Seeing Through Colorblindness: Implicit Bias and the Law.” UCLA Law Review 58: 465–​520. Kelly, D. and E. Roedder. 2008. “Racial Cognition and the Ethics of Implicit Bias.” Philosophy Compass 3 (3): 522–​540. Lee, C. K. 2016. “Revisiting Current Causes of Women’s Underrepresentation in Science.” In Implicit Bias and Philosophy, Volume 1, Metaphysics and Epistemology, edited by M. Brownstein and J. Saul, 265–​282. Oxford: Oxford University Press. Levy, N. 2005. “The Good, the Bad and the Blameworthy.” Journal of Ethics and Social Philosophy 1 (2): 1–​16. Levy, N.. 2014. “Consciousness, Implicit Attitudes, and Moral Responsibility.” Noûs 48: 21–​40. Levy, N. 2016. “Implicit Bias and Moral Responsibility. Probing the Data.” Philosophy and Phenomenological Research 93 (3):3–​26. Lilienfeld, S. O. 2017. “Microaggressions: Strong Claims, Inadequate Evidence.” Perspectives on Psychological Science 12 (1): 138–​169. Lippmann, W. 1922. Public Opinion. New York: Free Press. Machery, E. 2016. “De-​Freuding Implicit Attitudes.” In Implicit Bias and Philosophy: Volume I, Metaphysics and Epistemology, edited by M. Brownstein and J. Saul, 104–​129. Oxford: Oxford University Press. Macrae, C. N., G. V. Bodenhausen, A. B. Milne, and J. Jetten. 1994. “Out of Mind but Back in Sight: Stereotypes on the Rebound.” Journal of Personality and Social Psychology 67 (5): 808–​817. McGeer, V. 2010. “Co-​reactive Attitudes and the Making of Moral Community.” In Emotions, Imagination and Moral Reasoning, edited by C. MacKenzie and R. Langdon, 299–​326. New York: Psychology Press. McGeer, V. 2015. “Building a Better Theory of Responsibility.” Philosophical Studies 172 (10): 2635–​2649. Monteith, M. J., L. Ashburn-​Nardo, C. I. Voils, and A. Czopp. 2002. “Putting the Brakes on Prejudice: On the Development and Operations of Cues for Control.” Journal of Personality and Social Psychology 83(5), 1029–​1050. Monteith M. J., J. W. Sherman, and P. G. Devine. 1998. “Suppression as a Stereotype Control Strategy.” Personality and Social Psychology Review 2: 63–​82. Monteith, M. J. and C. I. Voils. 1998. “Proneness to Prejudice Responses: Toward Understanding the Authenticity of Self-​Reported Discrepancies.” Journal of Personality and Social Psychology 75: 901–​916. Moody-​Adams, M. M. 1994. “Culture, Responsibility and Affected Ignorance.” Ethics 104 (2): 291–​300. Moskowitz, G.B. and P. Li. 2011 “Egalitarian Goals Trigger Stereotype Inhibition: A Proactive Form of Stereotype Control.” Journal of Experimental Social Psychology 47 (1): 103–​116. Moss-​Racusin, C. A., J. F. Dovidio, V. L. Brescoll, M. J. Graham, and J. Handelsman. 2012. “Science Faculty’s Subtle Gender Biases Favor Male Students.” PNAS 109 (41): 16474–​16479. Myers, C.K., M. Bellows, H. Fakhoury, D. Halle, A. Hall, and K. Ofman. 2008. “Ladies First?: A Field Study in Discrimination in Coffeeshops.” Applied Economics 42 (14): 1761–​1769. Nadelhoffer, T. 2008. “Bad Acts, Blameworthy Agents, and Intentional Actions: Some Problems for Juror Impartiality.” In Experimental Philosophy, edited by J. Knobe and S. Nichols, 149–​ 170. New York: Oxford University Press.

Moral Responsibility for Implicit Bias    695 Payne, B. K. 2001. “Prejudice and Perception: The Role of Automatic and Controlled Processes in Misperceiving a Weapon.” Journal of Personality and Social Psychology 81 (2): 181–​192. Pennington, C.R., D. Heim, A. R. Levy, and D. T. Larkin. 2016, “Twenty Years of Stereotype Threat Research: A Review of Psychological Mediators.” PLOS One 11 (1): 1–​25. Pettit, P. 2007. “Neuroscience and Agent-​Control.” In Distributed Cognition and the Will: Individual Volition and Social Context, edited by D. Ross, D. Spurrett, K. Harold, and G. L. Stephens, 77–​91. Cambridge, MA: MIT Press. Plant, E. and P. G. Devine. 2001. Responses to Other-​Imposed pro-​Black Pressure: Acceptance or Backlash? Journal of Experimental Social Psychology 37: 486–​501 Rees, C. F. 2016. “A Virtue Ethics Response to Implicit Bias.” In Implicit Bias and Philosophy, Volume 1, Metaphysics and Epistemology, edited by M. Brownstein and J. Saul, 191–​214. Oxford: Oxford University Press., Rowe, M. 2008. “Micro-​affirmations and Micro-​inequities.” Journal of the International Ombudsman Association 1 (1): 45–​48. Saul, J. 2012. “Skepticism and Implicit Bias.” Disputatio Lecture 5 (37): 243–​263. Saul, J. 2013. “Implicit Bias, Stereotype Threat, and Women in Philosophy.” In Women in Philosophy: What Needs to Change?, edited by F. Hutchinson and K. Jenkins, 39–​61. Oxford: Oxford University Press. Schlosser, M. E. 2013. “Conscious Will, Reason-​Responsiveness, and Moral Responsibility.” The Journal of Ethics 17 (3): 205–​232. Shih, M., J. A. Richeson, M. Ambady, K. Fujita, and H. M. Gray. 2002. “Stereotype Performance Boosts: The Impact of Self-​Relevance and the Manner of Stereotype Activation.” in Journal of Personality and Social Psychology 83(3): 638–​647. Sie, M. 2009. “Moral Agency, Conscious Control, and Deliberative Awareness.” Inquiry 52 (5): 516–​531. Sie, M. 2013. “Free Will an Illusion? An Answer from a Pragmatic Sentimentalist Point of View.” In Exploring the Illusion of Free Will and Moral Responsibility, edited by G. D. Caruso, 273–​ 289. New York: Lexington Books. Sie, M. 2014. “Self-​Knowledge and the Minimal Conditions of Responsibility: A Traffic-​ Participation View on Human (Moral) Agency.” The Journal of Value Inquiry 48 (2): 271–​291. Sie, M. 2018. “Sharing Responsibility: The Importance of Tokens of Appraisals to Our Moral Practices.” In Social Dimensions of Moral Responsibility, edited by M. Oshana, K. Hutchison, and C. Mackenzie. Oxford: Oxford University Press. Sie, M. and A. Wouters. 2008. “The Real Neuroscientific Challenge to Free Will.” Trends in Cognitive Science 12 (1): 3–​4. Sie, M. and N. M. Van Voorst Vader-​Bours. 2016. “Stereotypes and Prejudices: Whose Responsibility? Indirect Personal Responsibility for Implicit Bias.” In Implicit Bias and Philosophy, Volume 1, Metaphysics and Epistemology, edited by M. Brownstein and J. Saul, 90–​114. Oxford: Oxford University Press. Smiley, M. 1992. Moral Responsibility and the Boundary of Community. Chicago: University of Chicago Press. Smith, A. 2005, “Responsibility for Attitudes: Activity and Passivity in Mental Life.” Ethics 115(2): 236–​271. Smith, A. 2008. “Control, Responsibility, and Moral Assessment.” Philosophical Studies 138: 367–​392. Springer, E. (2013). Communicating Moral Concern. An Ethics of Critical Responsiveness. Cambridge, MA, MIT Press.

696   Maureen Sie Steele, C. M. and J. Aronson. 1995. “Stereotype Threat and the Intellectual Test Performance of African Americans.” Journal of Personality and Social Psychology 69 (5): 797–​811. Steinpreis, R. E., K. A. Anders, and D. Ritzke. 1999. “The Impact of Gender on the Review of the Curricula Vitae of Job Applicants and Tenure Candidates: A National Empirical Study.” Sex Roles 41 (7/​8): 509–​528. Strawson, P. F. (1962) 1982. “Freedom and Resentment.” Reprinted in Free Will, edited by G. Watson, 59–​81. Oxford: Oxford University Press. Sue, D. 2010. Micro-​Aggressions in Everyday Life: Race Gender and Sexual Orientation. New York: Wiley. Taylor, Charles (1994). “The Politics of Recognition.” In Multiculturalism, edited by I. Gutman, 25–​73. Princeton, NJ: Princeton University Press. Uhlmann, E. L., V. L. Brescoll, and E. Machery. 2010. “The Motives Underlying Stereotype-​ Based Discrimination Against Members of Stigmatized Groups.” Social Justice Research 23 (1): 1–​16. Unkelbach, C., L. Goldenberg, N. Müller, G. Subbe, and N. Spannaus. 2009. “A Shooter Bias in Germany Against People Wearing Muslim Headgear.” Revue Internationale de Psychologie Sociale No ¾: 181–​201. Unkelbach, C., J. P. Forgas, and T. F. Denson. 2008. “The Turban Effect: The Influence of Muslim Headgear and Induced Affect on Aggressive Responses in the Shooter Bias Paradigm.” Journal of Experimental Social Psychology 44 (5): 1409–​1413. Valian, V. 2005. “Beyond Gender Schemas: Improving the Advancement of Women in Academia.” Hypati, 20: 198–​213. Van Voorst Vader-​Bours, N. M. In preparation. Making the Invisible Visible: Implicit Bias, Harm and Responsibility. Dissertation. Vargas, M. 2013. Building Better Beings: A Theory of Moral Responsibility. Oxford: Oxford University Press. Washington, N. and D. Kelly. 2016. “Who’s Responsible for This? Moral Responsibility, Externalism, and Knowledge About Implicit Bias.” In Implicit Bias and Philosophy: Volume 2, Moral Responsibility, Structural Injustice, and Ethics, edited by M. Brownstein and J. Saul. Oxford: Oxford University Press, 10–​36. Zheng, R. 2016. “Attributability, Accountability, and Implicit Bias.” In Implicit Bias and Philosophy, Volume 1, Metaphysics and Epistemology, edited by M. Brownstein and J. Saul, 62–​89. Oxford: Oxford University Press.

CHAPTER 33

Skep ticism ab ou t Ev i l : Atro cit y and t h e L i mi ts of Resp onsi bi l i t y Dominic Murphy and John M. Doris 1

In what follows, we argue for an untoward conclusion: There’s relatively little evil in the world. This will likely strike readers as manifestly implausible, and with good reason: There undeniably exists an abundance of horror perpetrated by human beings on one another, and it is natural to attribute this horror to the moral turpitude—​the moral evil—​of the perpetrators. Our contention to the contrary, however implausible or even repugnant it may seem, is in fact derived from a highly plausible supposition: An act is not attributable to a person in a way that justifies a negative moral assessment of that person unless they are morally responsible for the act. If so, for a reprehensible deed to be attributable to the evil of the perpetrator, the perpetrator must be morally responsible for that deed. We believe, however, that it is not common for people to be morally responsible for such conduct. We will discuss extreme wrongdoing both by lone individuals and by individuals acting in institutional contexts, and argue for two presumptions: The Presumption of Morally Relevant Impairment,2 for individual dispositional explanations of evil; and The Presumption of Extenuating Circumstances, for situational explanations of evil. We argue that taken together, these presumptions motivate a third presumption, The Presumption of Non-​Responsibility: Whether one favors dispositional or individual explanations of 1 

Author order determined randomly. We’ve chosen “impairment” here, but there’s no unproblematic selection, since most words in the vicinity, like “deficit,” may invoke discriminatory ablism. We intend something like the “adverse departure from normal functioning,” sometimes used to characterize illness, while recognizing the drawbacks with this approach, not least of which is the infamous difficulty of perspicuously characterizing “normal” (see Woolfolk 1999). Another obvious choice is “disability,” but the issues here are, if anything, even more sensitive and complex; Barnes (2016), for example, questions the conventional association of disability with impairment. For criticism of academic philosophy’s treatment of disability, see Tremain (2017). 2 

698    Dominic Murphy and John M. Doris atrocity, perpetrators tend to not be in a psychological condition consistent with the warranted attribution of moral responsibility. Then, if moral responsibility is typically necessary for the attribution of evil, it follows that the attribution of evil is typically unwarranted. We call this conclusion Skepticism about Evil. In motivating this skepticism, we address two questions that have been of considerable concern to philosophers working on the theory of moral responsibility: Are psychopaths morally responsible? Are perpetrators of atrocity in warfare morally responsible? Our answer in both cases is No. Or at least, Most usually not. These answers have not been universally endorsed, so part of our burden here will be defending our replies. We also shoulder a more practical burden having to do with appropriate responses to extreme wrongdoing. Skepticism about evil appears to counsel an obnoxious lenience, if individual perpetrators are not morally responsible for their appalling deeds, then punishing them may be prohibited. Unfortunately, this lenience is not easily avoided, because existing expedients for justifying punitive responses in the cases we consider are themselves problematic. In what follows, after we develop skepticism about evil, we consider how we might respond to extreme wrongdoing, and reluctantly endorse strict liability. None of the available responses, we find, are particularly appealing. But this, unfortunately, is exactly what one should expect: The moral casualties of extreme wrongdoing are less visible than the physical casualties, but are nevertheless catastrophic.

1.  What Is Evil? To begin, consider the rhetorical uses to which “evil” is put. Here is George W. Bush, in the aftermath of the September 11, 2001, terrorist attacks on the United States: We’ve been warned there are evil people in this world. We’ve been warned so vividly. And we’ll be alert. Your government is alert. The governors and mayors are alert that evil folks still lurk out there. My administration has a job to do and we’re going to do it. We will rid the world of the evil-​doers.3

If a person is evil, this attribute cannot be outweighed or counterbalanced by other attributes they possess: The moral evaluation of an evil person must always be on balance exceptionally negative. One might, on bringing one’s latest flame home to meet the family, say things like, “He doesn’t dress very well, but he’s a good person”; or “She’s not very punctual, but she’s a good person”; or even, “He sometimes shades the truth, but he’s a good person.” Attributions of evil admit of no such nuance. You would not, we submit, introduce your date by straight-​facedly saying, “He’s evil, but we have a lot of fun.”

3 

http://​edition.cnn.com/​2001/​US/​09/​16/​gen.bush.terrorism/​

Atrocity, Evil, and Responsibility    699 To designate someone as evil is to demand their exile from the moral community. Evil is not to be engaged, but eradicated; we must “rid the world of the evil-​doers.” One has to have done something truly terrible to get designated evil: The extremity of the offense justifies the extreme response. Odd to say we’re evil if we go in for a bit of teasing about your funny hat, even if we meant to be a little mean; conversely, evil inflicts intense suffering. And the infliction must be unnecessary, unjustified, or “wanton”; the suffering your dentist causes you, even where it reaches considerable levels, isn’t evil. Thus, evil deeds are supposed to be “breathtakingly horrible” and “incomprehensible” to ordinary people (Stone 2009: 22). Additionally, attributions of evil typically assume psychologism; for an actor to be counted evil, the actor’s deeds must be appropriately sourced in deplorable psychological states.4 (Our skepticism is skepticism about evil persons, but may extend to evil actions, if evildoers are required for evil deeds.) Numerous specifications of the requisite psychological states have been suggested: It has been claimed the evildoer must act with pleasure (Steiner 2002: 190; McGinn 1997: 61ff.), or indifference (Haybron 2002: 269), or intent to dehumanize (De Wijze 2002: 218; Garcia 2002: 201). Our guess is that while some evil deeds may be aptly described according to each of these accounts, none of the proposals successfully identifies a necessary condition on evil: For instance, one evildoer might be unfeelingly callous, and another actively sadistic. In any event, we’ll do no more than propose, as we’ve already suggested, the following minimal condition on psychologism: To be appropriately designated evil, the perpetrator must be morally responsible for the conduct in question (cf. Perrett 2002: 304; Thompson 2002: 240). Beyond this, we won’t commit to a theory of morally responsible agency, which, done properly, would require immersion in a huge literature of impressive sophistication and complexity (as the contributions to the present volume well illustrate).5 We can offer a bit of framing, in terms of the philosophical commonplace that desert presupposes responsibility: For someone to deserve some response, they must be morally6 responsible for whatever that response targets. It follows that a perpetrator does not deserve to be designated evil if they are not responsible for their offense. On occasion, the desert-​responsibility commonplace has been questioned; Feldman (1995: 68) thinks it part of the “received wisdom” about desert, but insists it is “clearly false.”7 And indeed, it is easy to adduce apparent counterexamples: One might deserve compensation for

4  There’s also the problem of “Natural Evil” for various theisms (Swinburne 1978): If there exists a sufficiently powerful deity, the suffering caused by earthquakes, hurricanes, illness, and the like is at least partly due the deity’s agency or omission. Those skeptical about the existence of such a deity will of course be skeptical about the existence of Natural Evil, but in any event, we won’t take up that issue here; the skepticism we’re contemplating is skepticism about evil human beings. 5  One of us (Doris 2015a) has attempted a theory of morally responsible agency, but that theory is not presupposed here. 6  Not all formulations of the doctrine include the “morally” qualifier, but it is required; for instance, in many cases mere causal responsibility is insufficient for desert. 7  Feldman cites Rachels (1978: 157), Sadurski (1985: 187), and (more tentatively) Rawls (1971: 104) as receivers of the wisdom.

700    Dominic Murphy and John M. Doris some quite accidental misfortune, or deserve some forms of respect merely by virtue of being a human being. We gain some clarity by marking off “deserves” from nearby linguistic territory such as “merits,” “is due,” “is owed,” and “is entitled”: For example, it might be said that someone is owed, but does not deserve, compensation for injury; or that someone merits, but does not deserve, appreciation for their naturally good looks. If you’re reluctant to let 1950s analytic thuggery decide substantive philosophical questions, instead note an apparent asymmetry: While Feldman might have it right about benefits (which are his central examples), the responsibility-​desert connection may be tighter for burdens: While the authors certainly appreciate our fellow flyers’ dismay when we join them in their row, it would be bizarre to say we deserve to be punished for our large dimensions. The diagnosis here is obvious: We don’t deserve punishment because our inconvenient size is not our fault—​that is, we’re not responsible for it. So desert may exhibit asymmetry across the negative and positive cases, as indeed, may responsibility itself, if responsibility conditions differ for wrongdoing and rightdoing differ, as Nelkin (2011: ch. 2) argues. Indeed, the variety of relevant phenomena may counsel pluralism about desert (Sher 1987). We presume that, for one class of phenomena, desert presupposes moral responsibility, and our thesis is closely circumscribed around the case of moral evil: No one deserves the designation “evil” who is not morally responsible for the transgression in question. Another attempt to sever the responsibility-​desert connection might be derived from Vargas’s (2013: 250, 265) account, where desert is justified by the good it does its subjects. In as much as responsibility attributions look backwards (or sideways) at the condition and circumstances of actors, a forward-​looking, teleological approach sunders desert and responsibility. But this appears a radical revision of thinking about desert, if it is a way of thinking about desert at all (Doris 2015b), and Vargas (2015: 2666–​2671) himself doesn’t endorse it, since he thinks it is desert practices, and not ordinary attributions of desert themselves, that are justified teleologically. As will emerge, we think teleological considerations have a role in appropriate responses to extreme wrongdoing, but we will adhere to the commonplace in denying them a role in the attribution of desert. A final striking possibility is that no one is deserving in the ordinary sense we adopt here, because no one is morally responsible.8 But this way of proceeding does not sever the responsibility-​desert connection, but assumes it, by arguing from skepticism about responsibility to skepticism about desert. Our position, again, is circumscribed: We are not arguing for skepticism about responsibility, but arguing for non-​responsibility regards the preponderance of extreme wrongdoing, a position we take to entail skepticism about evil. Once again, our rendering of responsibility is, for the purposes here, ecumenical, requiring only another philosophical commonplace, a familiar understanding of excuses. Excusing conditions

8 

This broadly skeptical possibility recalls Pereboom’s (2001, 2014) “hard incompatibilism,” though we are unable to explore the nuances of Pereboom’s position here.

Atrocity, Evil, and Responsibility    701 transiently block particular attributions of responsibility: central classes of excuse are force (e.g., acting under credible threat), ignorance (e.g., non-​culpably misperceiving consequences of one’s action), and impairment (e.g., an adverse drug reaction destabilizing cognitive capacities).9 When impairments are persisting and global rather than transient and local, it is common to term them as exemptions rather than excuses (e.g., Wallace 1994: 118); severe mental illness is a standard example. We’ve here another case of a useful, but unstable, philosophical distinction: Both temporal duration and range of affected domains vary on a continuum, and there may be little gained, in many cases, by arguing over whether a given affliction is an excuse or exemption (cf. Brink 2013a: 9). Another useful-​but-​ unstable way of demarcating matters, which runs closer to the strategy we shall adopt here, locates some responsibility-​undermining conditions in the person, and others in the situation: An individual’s attributes may destabilize agency, and so may her circumstances. The important point is that when the capacities requisite for morally responsible agency are absent or substantially impaired (through no fault of the person’s own), attribution of moral responsibility is typically inappropriate.10

2.  The Presumption of Morally Relevant Impairment: Dispositional Explanations of Evil Often, horrific deeds appear attributable to the personality or character of the perpetrator: Evildoers do what they do because of who they are; it seems that regular folks just don’t do stuff like that. We could, unfortunately, note countless ugly examples, but we’ll recount just one, Kent Kiel’s (KK) interview with an incarcerated felon named “Shock Ritchie” (SR):11 Richie smiled as he told me a story of a prostitute he had killed for pissing him off. He actually seemed proud when he described wrapping her up in the same blanket

9  Fischer and Ravizza’s (1998: 12) classic discussion divides excuses into those based in ignorance and those based in force. Impairment might represent a distinct category, or it might reduce to the other two: For example, a compulsion might be thought of as force, and a hallucination as ignorance. 10  A suggestive way to understand this is by means of the influential family of “reasons-​ responsiveness” theories of responsibility (e.g., Fischer & Ravizz 1998; Nelkin 2011; Vargas 2013): Mitigations, excuses, or exemptions obtain when the capacity to appropriately respond to reasons is sufficiently impaired. We will not depart our professed theoretical neutrality, but the reader is free to do so. For some ecumenical thoughts on reasons-​responsiveness and competitors, see Doris (2018), and McKenna and Van Schoelandt (2015). 11  Adapted from Kiehl (2014: 84–​85). No material has been added, but for ease of reading, ellipses have not been marked.

702    Dominic Murphy and John M. Doris he had suffocated her with so he could keep all the forensic evidence in one place. He put her in the trunk of his car and drove out to a deserted stretch of road bordered by a deep forest. SR: “I had all these great plans to carry the body miles into the woods and bury it really deep so nobody would ever find it. But it’s fucking hard to carry a body. You ever tried to carry a body?” KK: “No, I don’t have any experience carrying dead bodies.” SR: “Well it’s a lot of work, let me tell you. So, I only got about a hundred yards off the road and just into the trees before I was exhausted. Then I went back and got the shovel from the car. I started digging a huge hole. You know how hard it is to dig a hole big enough to bury a body?” KK: “No, I don’t have any experience digging holes to bury bodies.” SR: “Well it’s harder than you might think.”

It won’t surprise you to learn that Shock Ritchie is a psychopath. Our focus here is “clinical psychopaths”—​conventionally, those who score over 30 on Hare’s Psychopathy Checklist, revised (PCL-​R), an interview and case-​history based measure that is arguably the diagnostic state of the art (Hare 2003; for commentary, see Forth, Bo, and Kongerslev 2013). The average North American non-​incarcerated male scores 4 on the PCL-​R, and the average inmate a 22 (Kiehl 2014: 10), so individuals scoring 30 or above are pretty dramatically different from the norm. While matters are both complicated and contentious, we contend that psychopaths are not typically morally responsible for their moral transgressions. (We leave it open whether they are responsible for other things they do; if you think serial killer John Wayne Gacey deserved credit for his paintings of clowns, which apparently sell for thousands of dollars,12 we won’t argue). As we’ll see, other commentators take the contrary position, and still others (e.g., Maibom 201813) caution, not unreasonably, that the state of clinical science is presently too uncertain to reach a confident conclusion one way or the other. There are even those who doubt that psychopathy is a legitimate clinical category: In a trenchant critique, Jalava, Griffiths, and Maraun (2015: 8–​9) contend that psychopathy became a prominent diagnosis “despite the lack of compelling data” and insist that there is “no convincing evidence that psychopathy is a biologically based disorder.”14 We’ll side with prevailing scientific opinion in supposing there’s a there there in the vicinity of what scholars talk about when they talk about psychopathy, but even if Jalava and colleagues are right, what’s important for the Presumption of Morally 12 

http://​metro.co.uk/​2017/​10/​09/​chilling-​paintings-​by-​killer-​clown-​john-​wayne-​gacy-​ expected-​to-​sell-​for-​7000-​each-​6987388/​ and https://​spottedcouchartcrimeblog.com/​2014/​02/​18/​ bring-​in-​the-​clowns-​the-​nefarious-​paintings-​of-​john-​wayne-​gacy/​ 13  Maibom here seems to depart Maibom (2008), where she seemed pretty firmly in the “psychopaths are responsible” camp. 14  In the course of castigating philosophers working on psychopathy for their sloppy and credulous attempts at interdisciplinarity, Jalava and Griffiths (2017) argue for a more modest conclusion: The evidence regarding psychopathy and responsibility is so far inconclusive.

Atrocity, Evil, and Responsibility    703 Relevant Impairment is the existence of deficits implicated in moral functioning, not whether they should be thought of as organized into a clinical or natural kind. Our focus here will be moral responsibility rather than criminal liability; although we range over literature treating both questions, we should note a striking discontinuity: While numerous contemporary theorists argue that psychopaths should be excused from criminal liability, the practice of criminal law uniformly rejects this suggestion (Luna 2013: 359). We’ll have a little (too little) to say about this when we consider, later on, how best to respond to atrocity. For now, we’ll canvass considerations speaking to moral responsibility. To begin, we should avoid over-​pathologizing wrongdoing: The fact that someone behaves in an aberrant or wildly counter-​normative fashion cannot be treated as sufficient reason for attributing psychopathology to them. Such conduct may well be evidence of mental illness, but we should take seriously the hypothesis that some people who do bad things, even incomprehensibly bad things, suffer no infirmity, and are best understood simply as awful people. However, psychopaths present abnormally independently of criminality or wrongdoing.15 And this abnormality clearly involves impairment; psychopathy is “associated with emotional deficits that are linked to impairments in nearly all domains of life” (Kiehl 2014: 234).16 Psychopaths are familiarly, and not inaptly, characterized as “without conscience” (Hare 1993; Kiehl 2014: 37). Simply put, they don’t care a whit about anyone else, and they have little hesitation, and no remorse, about harming others when it suits their purposes. Contrary to the impression one gets from popular culture, the archetypal psychopath is not a serial killer like Ted Bundy. Such individuals are probably psychopaths who also exhibit some paraphilia, like sexual sadism; this very dangerous conjunction is also, mercifully, very rare (Kiehl 2014: 11). Although confident estimates are difficult to arrive at, the FBI puts the percentage of murders committed by serial killers at less than 1%; the comprehensive Radford Serial Killers Data Base lists around 3,300 such perpetrators from around the world.17 If many of these murderers are psychopathic, than what we say about “garden variety” psychopaths, who don’t offend in quite the horror movie fashion of serial killers, applies here as well. Again, psychopathy involves impairment independent of immorality. The PCL-​R includes such morally relevant items as callous/​lack of empathy, juvenile delinquency, revocation of conditional release, and criminal versatility (Hare 2003). But it also contains 15  One unexpected example: Psychopaths may be “unable to appreciate metaphors”; their brains do not respond differently to abstract and concrete words, as neurotypical brains do (Kiehl 2014: 164). 16  It might be argued, on broadly evolutionary grounds, that psychopathy is not a disability, because psychopaths—​perhaps unsurprisingly, given their promiscuity—​may be “reproductively successful” (Rice and Harris 2013: 238–​240). There are famously difficult questions about what should count as a mental illness (e.g., Murphy 2006), but we don’t think the question should be decided by counting offspring. 17  https://​www.newyorker.com/​magazine/​2017/​11/​27/​the-​serial-​killer-​detector http://​skdb.fgcu.edu/​ info.asp

704    Dominic Murphy and John M. Doris items that extend beyond moral functioning: poor behavioral controls, lack of realistic, long-​term goals, and impulsivity. The latter tendencies have been noted in the literature since Cleckley’s The Mask of Sanity, the founding text of modern psychopathy research, which notes these psychopathic characteristics: poor judgment and failure to learn by experience, specific loss of insight, failure to follow any life plan (1941/​1964: 362–​363). Consistent with a lack of insight, psychopaths typically don’t feel there’s anything wrong with them (Kiehl 2014: 42). This insouciance persists when they are in objectively horrible circumstances, like prison; psychopaths “don’t get bothered by much of anything,” and are not prone to rumination or depression (Kiehl 2014: 28). Particularly striking, if you’re a bit of a checker, like the current authors, is that psychopaths never worry about things like leaving the stove on when they go out; according to Kiehl (2014: 164), they “have no concept of what worrying like this is like.” One can overcheck, of course, but a bit of worry is probably required to safely navigate the world. Difficulty in long-​ term planning is suggested by the overrepresentation of psychopaths among workers at temporary employment agencies; in one sample of temp workers, 13.5% scored 30 or more, and 30.3% scored 25 or more (the PCL-​R cutoff for psychopathy used in some studies; Raine 2013: 122–​123). Often psychopaths’ lives are “chronically unstable and aimless,” and their behavior, even momentous behavior like killing, seems bizarrely unmotivated (Hare 1993: 57–​58; cf. Litton 2008; Morse 2013: 331): it doesn’t, to use the Gibbardian marker of rationality, “make sense” (Gibbard 1990). All of this speaks to global difficulties in practical reasoning; indeed, it has been asserted that psychopathy is “essentially a disorder of decision making” (Koenigs and Newman 2013: 93). In addition to deficits in practical reasoning are emotional deficits; noted in the PCL-​ R are lack of remorse or guilt, shallow affect, callous, and lack of empathy (Hare 2003). These deficits are implicated in how psychopaths perceive, and fail to perceive, the social world: When one psychopathic prisoner was shown a picture of a fearful face, he said, “I don’t know what you call this emotion, but it’s what people look like just before you stab them.” Such emotion pictures are a staple of psychopathy research, and psychopaths often respond atypically (Blair, Mitchell, and Blair 2005: 54–​55).18 Psychopaths may be aware that they have broken a rule, but it is not clear that they genuinely appreciate the significance of this, since their understanding of what they have done lacks emotional charge. In some studies, psychopaths are much less aroused than non-​psychopaths by the distress of others (Blair, Jones, Clark, and Smith 1997). As Kiehl describes it, psychopaths “know the words but not the music.”19 Without music, songs fall flat. One celebrated, though controversial, finding is that psychopaths have difficulty understanding the difference between moral and merely conventional distinctions (Blair 1995; Blair, Jones, Clark, and Smith 1995); if that’s true, for them, 18  https://​www.theatlantic.com/​magazine/​archive/​2017/​06/​when-​your-​child-​is-​a-​psychopath/​ 524502/​ 19  https://​www.theatlantic.com/​magazine/​archive/​2017/​06/​when-​your-​child-​is-​a-​psychopath/​ 524502/​

Atrocity, Evil, and Responsibility    705 knocking someone down and dressing inappropriately comes to the same thing—​a violation of social norms.20 Some commentators doubt that the literature robustly supports the hypothesis that psychopaths are unable to make moral judgments (e.g., Borg and Sinnott-​Armstrong 2013). But a case for psychopaths’ moral non-​responsibility does not require that they not be able to make moral judgments; they might be able to make moral judgments, but not to care about them (Cima, Tonnaer, and Hauser 2010). If they are not able to care about moral judgments, then they may not be morally responsible for failing to act in accord with them. There is mounting evidence that psychopaths’ emotional deficits are associated with neurological deficits, notably in the amygdala and other elements of the paralimbic circuitry (Blair, Mitchell, and Blair 2005: 110–​140; Raine 2013: 97; Ermer, Cope, Nyalakanti, Calhoun, and Kiehl 2012, 2013; Koenigs and Newman 2013: 100–​ 102). More generally, there is likely an “anatomy of violence,” where physiological markers correlate with antisocial behavior. For example, poor prefrontal neurological functioning has been repeatedly implicated in psychopathy and antisocial behavior (Koenigs 2012; Raine 2013: 68; Yang and Raine: 2009), and low resting heart rate, which is associated with fearlessness and low arousal, is correlated with antisocial behavior (Raine 2013: 312).21 It’s likely that people “come from the shop” predisposed to criminality; psychopaths are “abnormal from birth,” often exhibiting a disturbing lack of affect from childhood (Kiehl 2014: 142, 235), and psychopathic traits have been characterized as “moderately to strongly heritable” (Viding, Fontaine, and Larsson 2013: 176). Yet it’s hard not to believe that what happens after birth is also heavily implicated in criminality. Here’s the life history of rapist and murderer Donta Page: Teenage pregnancy. Potential birth complications combined with uncaring, callous mother. Total absence of father. Impoverished neighborhood. Vigorous shaking as an infant that likely resulted in a disconnection between the frontal cortex in the limbic system. Sustained and severe physical and sexual abuse, including rape resulting in scaring and rectal bleeding. Total neglect. Early head injuries and multiple visits to the emergency hospital in the first two years of life. Neurotoxic lead exposure. Poor nutrition. A complete lack of supervision. Learning impairment. A family history of mental illness and signs of depression, ADHD, and conduct disorder as early as elementary school. Impaired executive functioning and memory. Low physiological arousal. Poor functioning of the orbitofrontal and medial prefrontal cortex in addition to reduced temporal-​pole functioning. (Raine 2013: 313–​314)

20 

This finding has been widely enough questioned (e.g., Aharoni, Sinnott-​Armstrong, and Kiehl 2012; Brink 2013a; Vargas & Nichols 2007) that it should not be made the center of a case for non-​responsibility. 21  The relationship is as strong as the relationship between SAT and college GPA; low resting heart rate during stressors may explain 12% of variation in antisocial behavior (Raine 2013: 104).

706    Dominic Murphy and John M. Doris Serious hesitation about holding Page responsible seems mandatory to us. We’re not suggesting that appallingly deprived upbringings are, in themselves, exculpating (though see Buss 1997). In the philosophical argot, we are currentists rather than historicists about the attribution of responsibility: Only the actor’s current states, and not her past history, are directly relevant to the attribution of morally responsible agency (Doris 2015a: 30–​32; with Vargas 2013; pace Fischer 2006: 236–​239). The history is epistemically relevant, of course, for assessing the current state. And in Page’s case, both history and contemporary assessment give us every reason to think that he suffered impairments. Even where upbringings are less egregious than Page’s, our position is that psychopaths are not morally responsible for their moral transgressions (cf. Morse 2008); they seem to suffer quite a general inability for successful self-​direction, which is very naturally construed as a hallmark of morally responsible agency (Doris 2015a: ch. 2). We acknowledge, again, that courts do not consider a diagnosis of psychopathy an “affirmative defense” (where the criminal conduct is acknowledged, but responsibility for it is denied), because psychopaths are asserted to have the relevant rational capacity, meaning they (1) know what they are doing and (2) know it is wrong (Raine 2013: 325). The prima facie considerations favoring this are obvious: Often enough, psychopaths are aware they are breaking the law, and, like Shock Ritchie, try to conceal their crimes.22 Even if one concurs with the law, moral and criminal responsibility look importantly different: While criminal responsibility might be a “thinner” notion, requiring only the relatively minimal capacity to detect when something is in violation of the law—​a capacity psychopaths typically have—​moral responsibility looks to require a genuine, “thick” appreciation of moral demands—​a capacity psychopaths are plausibly thought to lack (cf. Brink 2013a: 27–​32). Yet many commentators acknowledge that psychopathy involves morally relevant deficits while continuing to argue that afflicted individuals are subject to attributions of moral responsibility (e.g., Pillsbury 2013; Maibom 2008; Talbert 2008). Talbert (2008), for example, observes that psychopaths suffer a kind of “moral blindness,” but contends they nevertheless engage in reasoning that exhibits viciousness or ill will (e.g., “that my action will hurt you provides me no reason to refrain”), which justifies holding them morally responsible. Some philosophers, such as Shoemaker (2011a,b; 2015), propose that progress can be made by disentangling multiple senses of moral responsibility, and observe that psychopaths may be responsible in some of these senses and not others. One prominent response along these lines is due to Watson, who distinguishes what is often called attributability, or assigning a wrongdoer some moral fault, and accountability, or subjecting the wrongdoer to adverse responses, like blame (1996; cf. Nelkin 2015). Psychopaths, on this line, manifest moral defects, and so are attributability responsible, but because they are incapable of moral engagement with others, they are not properly 22  Interestingly, psychopaths’ attempts at concealment often seem ineffectual in ways that would amuse in other contexts; the body Richie attempted to bury was discovered in short order. Perhaps this speaks to deficits in executive function.

Atrocity, Evil, and Responsibility    707 subject to moral demands, and therefore not accountability responsible (Watson 2011). This expedient has considerable intuitive appeal, not least because it gives voice to a marked ambivalence many seem to experience in thinking about individuals who are at once inarguably nasty and obviously impaired. We’re nevertheless inclined, with Nelkin (2015: 386–​387), to hold that psychopaths are responsible in neither sense. Our central focus is in the vicinity of attributability: We think that the extent of psychopaths’ morally relevant impairments counsel against attributing evil to them, in the form of dispositions to evil, or evil traits. We’re not sure the issue of accountability is entirely distinct from attributability, in as much as the abilities and attributes that people have do appropriately inform how we respond to them. However, we’ll for the moment focus on attributability, and defer issues associated with accountability until later, when we turn to the question of responding to extreme wrongdoing. For attributability, one might start by noting that, given the robust genetic influences on psychopathy, the unpleasantness psychopaths display is most likely no—​or at least little—​fault of their own. But this argument threatens sweeping skepticism about moral responsibility, because there are robust genetic influences on most human attributes (Nichols and Vargas: 2007). Our sense is that matters are tricky here, and that a “genetic argument” might seem to more readily tell against the responsibility of those suffering other, less unsettling impairments: if you’re inclined to assign the psychopath attributability responsibility for the deficits associated with their disease, are you willing to do the same for those suffering major depressive disorder, post-​partem depression, bipolar disorder, schizophrenia, and Tourette’s syndrome? (We ought to be mindful: The very fact that psychopaths are such unsympathetic sorts may unduly influence theorizing about them.) This looks to us a plausible way to approach the question: Unless there is a compelling argument that someone is responsible for an impairment, they are not responsible for the behaviors caused by that impairment. In this regard, it is worth noting that the psychopath’s deficits may be extremely difficult to ameliorate, even with treatment: To date, clinical outcomes for psychopaths (adult psychopaths, anyway) “have not been encouraging” (Caldwell 2013: 202), and psychopaths continue to have highly elevated rates of recidivism (Kiehl 2014: 118). There may be legitimate cases of “culpable self-​ impairment,” but we think such cases are likely to be few. The most likely candidates probably involve substance abuse, but in cases involving addiction, anyway, we think reservations are required. In cases not involving addiction—​say, an ordinarily sober person ties one on at a party in order to annoy their former spouse—​we may have culpable self-​impairment, but such cases look not to involve persisting impairment, and are therefore not relevant to psychopathy. We should say a bit more about what exactly we take the morally relevant impairments to be. Psychopaths’ poor impulse control and difficulty with realistic planning indicate that they suffer a volitional impairment, but it is less clear that this impairment raises to the level of an excuse (Brink 2013a: 33–​34). Most probably, psychopaths generally pass the “policeman at the elbow” test; the risk of (immediate) disclosure and punishment

708    Dominic Murphy and John M. Doris may dissuade them from offending, suggesting that they have the capacity to resist impulses to wrongdoing. Generally speaking, volitional excuses are vexed; as the American Psychiatric Association (1983: 685; cf. Morse 2002) noted, the boundary between “an irresistible impulse and an impulse not resisted is probably no sharper than that between twilight and dusk.” More promising, perhaps, are arguments based on cognitive impairment:23 psychopaths’ “moral blindness,” or inability to genuinely appreciate the reasons, like avoiding causing suffering to others, for refraining from wrongdoing. This inability does not, as we’ve suggested, prevent psychopaths from identifying conduct regarded as illegal, or, we’re inclined to think, immoral. But what this impairment does involve, as we’ve said, an inability to care about moral considerations, except in so far as it has implications for the psychopath’s perceived self-​interest. One thing this intimates is that attempts to sharply distinguish volitional and cognitive excuses, at least in this context, are likely doomed. Normal moral cognition is, when situated in actual social interaction, relatively “hot” cognition, suffused with emotion (D’Arms and Jacobson forthcoming; Greene 2014; Nichols 2004; Prinz and Nichols 2010). To see something in moral terms is also to feel something about it, and these feelings are in the normal case related, however imperfectly, to doing. Because of their emotional deficits, psychopaths are largely incapable, so far as we can see, of this kind of engaged moral cognition, and are therefore entitled to excuse for moral wrongdoing.24 There is an important caveat here: Psychopathy is an individual difference variable, meaning, obviously enough, that it is a dimension on which people differ. While the “snakes in suits” narrative may be oversold, some psychopaths enjoy better outcomes than others, leading researchers to distinguish between “successful” and “unsuccessful” psychopaths (Raine 2013: 128). Different individuals fall on different points on the PCL-​R, and an individual scoring 20, while still quite different from the norm, might present differently than someone meeting the conventional cutoff of 30, who in turn might be distinguishable from a high-​scorer with of 34 or better—​individuals who impress researchers as possessed of “something qualitatively very different” compared with individuals scoring only a few points less at 30 (Borg and Sinnott-​Armstrong 2013: 108). Different individuals with psychopathic tendencies, even those who score similarly, might present with differing endowments of capacities relevant to the attribution of moral responsibly, and it is possible that high-​function “successful psychopaths” are sometimes morally responsible—​at least to a greater extent than the low-​functioning, high-​scoring recidivist.25 23 

Our assessment of the relative plausibility of volitional and cognitive excuses is the converse of Brink’s (2013a: 34). 24  As Wonderly (2020) puts it, “emotional deficiencies might shield certain reasons from view and/​or inhibit one’s ability to act on such reasons.” 25  As is so often the case, advances in research reveal that there are more things rather than fewer: It’s likely there are multiple types or subtypes of psychopathy (e.g., Mokros, Hare, Neumann, Santtila, Habermeyer, and Nitschke 2015), or alternatively, that psychopathy should not be thought of as a single disorder (Maibom 2018: 72).

Atrocity, Evil, and Responsibility    709 In the end, as always, a particular attribution of responsibility is decided by attention to the details of each case, and we acknowledge that some psychopaths, on some occasions, may be morally responsible. But again, we are arguing for a general presumption, not an exceptionless verdict. As we’ve said, we call this presumption the Presumption of Morally Relevant Impairment: The (defeasible) assumption, in cases of extreme wrongdoing, is that impairment of moral functioning is implicated in the misconduct. The extent might range from the fully exempting or excusing to the mildly mitigating. But this of course, is what must be determined. So the challenge is this: For evil to be justifiably attributed, we must rule out impairment implicated in the wrongdoing to an extent that approaches or meets standards for excusing conditions. It might be objected that this challenge, like demands for ruling out defeaters in the case of epistemic skepticism, overstates the likelihood of defeaters (for discussion, see Doris 2015a: 64–​7 1). Impairments like psychopathy are pretty rare, one might argue, so the presumption should be just the opposite: The absence of morally relevant impairment should be assumed, at least for offenders who are not psychotic or suffering other florid psychiatric impairment. But extreme wrongdoing, though not nearly so rare as we’d wish, is pretty rare too, which makes it plausible to suspect that impairment and deed often travel together. The overall rate of psychopathy has been estimated at between .5 to 1% for males and .1% for females (Kiehl 2014: 272 n. 1), but 15 to 35% of inmates worldwide meet criteria for psychopathy—​with more psychopaths in prisons with higher security ratings (Kiehl 2014: 35)—​and it is conjectured that as many as 77% of psychopaths in the United States are imprisoned (Kiehl: 275 n. 1). Moreover, psychopaths may commit 30 to 40% of the violent crime in the United States (Kiehl and Sinnott-​Armstrong 2013: 1). It’s probably more dramatic still for the abhorrent cases that are the oft-​mentioned exemplars of evil: In one study, 87% of 145 serial killers were judged to be psychopaths (Stone 2009: 209). Granting that psychopathy is an impairment that typically involves excusing impairments, as we’ve argued ought to be granted, and noting the incidence of psychopathy among serious offenders, the Presumption of Morally Relevant Impairment actually seems a pretty modest requirement. It seems more modest still when one realizes that many perpetrators of extreme wrongdoing, like Donta Page, may suffer morally relevant impairments that are not directly associated with psychopathy.

3.  The Presumption of Extenuating Circumstances: Situational Explanations of Evil We now turn to forms of extreme wrongdoing perpetrated not by lone individuals like serial killers, but by people acting in institutional settings. Our example will be atrocities

710    Dominic Murphy and John M. Doris on the battlefield. One of us is on record attempting to dampen enthusiasm for individual dispositional explanations of behavior (Doris 2002; 2021), and we have jointly argued that many instances of atrocity in wartime are not attributable to individual pathology but to situational, institutional, and systemic factors (Doris and Murphy: 2007). There are two reasons for thinking individual pathology is not a sufficient explanation for the bulk of wartime atrocity. The first is some intuitive math; on a plausible guestimate, individuals suffering the relevant pathologies are not numerous enough to account for all of the atrocities observed, even if they are substantially overrepresented in such conduct. The second, and more telling, reason is that assessments of perpetrators often indicate that they are tolerably “normal”; it doesn’t always take a monster to commit monstrous deeds. We must therefore consider situational as well as dispositional explanations of extreme wrongdoing, and this section will reprise arguments from our previous paper. The My Lai massacre was carried out by US troops on March 16, 1968. It was not the only atrocity perpetrated in Vietnam, but it remains the best known and most intensely studied. Approximately 500 Vietnamese civilians were killed in what was nominally an attack against enemy insurgents, but only three weapons were recovered. The villagers were rounded up and killed in groups. Then the troops broke for lunch. The commander of the Platoon responsible for the worst excesses, Lt. William Calley, was convicted of murder, but although some of his subordinates were charged, none were ever tried. Some senior officers also suffered disciplinary consequences such as censure or demotion.26 The basic narrative of the day is well-​understood (Bilton and Sim 1992; Allison 2012) but controversy still attends the extent to which the killers believed themselves to be following orders. Many of the My Lai killers insisted they were, and the law of armed conflict requires soldiers to obey orders unless they are clearly illegal: According to Osiel (1999: 5) the “law is now generally understood to require that soldiers resolve all doubts about the legality of a superior’s orders in favor of obedience. It therefore excuses compliance with an illegal order unless the illegality—​as with flagrant atrocities—​would be immediately obvious to anyone on its face.” Even the proviso that one should refuse to obey orders that would obviously be deemed illegal by everyone else requires soldiers to make a judgment, and to do so in institutions and environments that almost everyone else is never in. This judgment 26 

Technically, My Lai may not have been a war crime. The dead were South Vietnamese, and hence allied with the United States, and the incident took place on non-​hostile territory. The Fourth Geneva Convention, which concerns protection of civilians, explicitly exempts the population of allied states from protection as long as “the State of which they are nationals has normal diplomatic representation in the State in whose hands they are” (Geneva Convention Relative to the Protection of Civilian Persons in Time of War, Aug. 12, 1949, 6 U.S.T. 3516). Calley, whose platoon carried out the worst acts that day, was convicted of murder, but not of violating the laws of war. However, the Hague Convention does extend the customary protection of international law to civilians and My Lai counts as a crime against humanity or war crime in common-​sense thought; it is routinely referred to as a war crime in the sources (on the protection of civilians, see Baik 2001).

Atrocity, Evil, and Responsibility    711 must be made in circumstances of extreme stress—​perhaps in life-​threatening circumstances on a battlefield. Determining the legality of conduct prescribed in orders is, we’ve argued, a “cognitive feat,” requiring the (relatively) unimpaired exercise of normative competence. Ingierd and Syse (2005: 96), for example, note that soldiers and officers must routinely make important decisions reflectively, but must also routinely trust their superiors without constantly second-​guessing orders. This involves two levels of cognitive competence—​a first-​order judgment and a judgment about whether a judgment is necessary. Even if one is unconvinced by skepticism regarding the efficacy of reflection (e.g., Kornblith 2012; Doris 2015a); the exercise of “switching” independent thought on and off is, especially in difficult circumstances, likely to be highly cognitively demanding. As we understand it, normative competence references the capacity or capacities that enable people to appreciate normative considerations, understand factual information relevant to these considerations, and effectively integrate the two for the regulation of behavior (cf. Doris 2002: 136; Watson 1993: 126–​127; Wolf 1990: 121–​129). We contend that soldiers are often cognitively degraded in ways that leave them unable to exercise this competence. We conclude that as a consequence, soldiers in combat typically occupy excusing conditions. There is a wealth of evidence to support the view that the cognitive capacities necessary for moral decision-​making can be degraded by context (for review, see Doris 2002; 2021; Doris and Murphy 2007; Olin and Doris 2014). And combat conditions, we contend, are such cognitively degrading circumstances. Determining whether an order is lawful is a cognitive achievement. Those in and around combat conditions—​adverse, often chaotic conditions in which they are frightened, stressed, and perceptually confused—​cannot be reasonably expected to deliver such achievements. They should therefore be presumed, barring decisive evidence to the contrary, to occupy excusing conditions. Even if one thinks general claims about “the power of the situation” have been oversold, there is compelling evidence that the situational pressures of warfare in particular are powerful influences on cognition and behavior (especially in view of the fact that the affected individuals have already been inculcated into military cultures and organizations and are thereby living by distinctive moral standards). While the empirical record suggests that situational pressures can impair normative competence, the extreme and often prolonged situational factors present in warfare can induce severe impairment of normative competence. Combat is, we hardly need to say, unspeakably terrifying; the physiology of fear is a powerful distorter of normal responses, as one shakes or soils oneself, and the surrounding environment is dangerous, loud, and confusing, as engines, weapons, and screams contribute to a disorienting cacophony.27 It is also commonplace among

27 

At My Lai, the perpetrators did not come under fire, but it is reasonably supposed that they experienced at least some degree of fear associated with undertaking military action.

712    Dominic Murphy and John M. Doris scholars that casualties among one’s comrades add to combat stress if they degrade the “primary group” of one’s close compatriots, and this degradation can lead to moral coarsening. Bartov (2001, ch. 2), for example, argues that destruction of the primary groups contributed to the misconduct of the Germans in Russia. At My Lai, Charlie Company was still reeling from recent minefield casualties. In addition to recognizing the role of situational stressors and pressures, it is important to emphasize that the required exercises of normative competence must be effected under a strong expectation of obedience; the training of service men and women is substantially aimed at turning them into obedient subordinates. Armies aren’t democracies, and the issuing of orders is not an invitation to legal or moral debate; indeed, refusal may be met with severe sanction. In short, we concur with McMahan’s (2009: 127) observation that “the stresses of battlefield conditions are frequently so severe as to impair a combatant’s capacity for rationality, creating a state of genuinely diminished responsibility.” Though agreeing with McMahan about the impaired rationality of many combatants who commit atrocities, we accept the conclusion that he thinks is universally recoiled from, namely that perpetrators of atrocities on the battlefield are typically in excusing conditions. To further explicate our argument, we observe that two sorts of pressures cause cognitive degradation and tell against the attribution of responsibility for atrocities and other war crimes.28 First, there are proximal situational pressures that act on soldiers in combat; second, there are distal pressures, “beginning at considerable spatial and temporal distance from the point at which atrocity occurs” (Doris and Murphy 2007: 39). Combinations of proximal and distal forces, we conclude, degrade combatants’ normative competence to a point which impairs their capacity to recognize or understand the moral impermissibility of certain acts that violate the laws of war and/​or ordinary moral norms. Situational and distal pressures work in different ways. US psychiatrists reflecting on World War II concluded that “every man had his breaking point; they estimated this breaking point to occur anywhere between 100 days and 1 year of active combat duty” (Pols and Oak 2007: 2135). That is, the situational pressures of combat are so distressing that all participants will eventually become impaired as a result. American psychiatry went into World War II thinking of mental illness as the result of a relatively stable disposition in an individual, engendered by congenital factors and childhood experiences. Observing the traumatizing effects of combat persuaded them otherwise (Pols 2011); warfare can destabilize even the psychologically healthiest individual. Frightening and disorienting events act on the psychology of men (and increasingly, women) who are already living in stressful and dangerous environments, marked by filth, hunger, and danger. And of course, many soldiers are called upon to inflict and suffer violence that would be unimaginable in most walks of life. Killing and enduring 28  We are grateful to Talbert and Wolfendale (2018: chs. 2 and 4) for critical and clarificatory discussion of Doris and Murphy (2007), especially for pressing us to be clearer about the relationship between normative competence and cognitive degradation.

Atrocity, Evil, and Responsibility    713 the threat of death may lead to a blurring of the ordinary lines between licit and illicit violence, so that “war is by its very nature conducive to criminal behavior” (Frésard 2004: 32). For many combatants, there is psychological degradation in which the ability to appreciate and act on customary moral considerations is no longer fully functional before the onset of full-​blown “shell shock” or “combat fatigue.” Distal pressures arise from the nature of armies and the interest taken in them by the state; no institution has been as central to state formation (Tilly 1990) or as heavily monitored and regulated by governments. Soldiers are the product of a deliberate process of training aimed at fostering obedience and rendering salient considerations favoring violence—​considerations that states normally seek to extinguish in citizens. Similarly, although military training does not aim at making one commit illegal acts, it does aim at transforming the individual into someone for whom lethal violence is a normal option, and it does so with the extensive resources available to a state-​run institution: Ricks (1997) presents basic training in the US Marine Corps as an intensive course of indoctrination and refashioning of the self, designed to uproot old values and replace them with those of military culture. The aim of military culture and socialization, including racial socialization, is to reconfigure the cognitive, evaluative, and motivational states of affected individuals. For instance, the My Lai incident took place in the context of a racialized war featuring dehumanizing characterizations of the enemy. Smith (2012) argues that dehumanization involves representing other human beings as nonhuman animals; people can be brought to see and act toward others in ways that depend on psychological structures that have evolved to enable cognition about nonhuman animals. If this is correct, then behavior can be entrained by systems that make perpetrators treat others as nonhuman animals—​even where the operation of their higher-​level cognition would not make the associated explicit inferences. The killers at My Lai needn’t have been literally confused about the species membership of their victims, but cognitive structures that process information about nonhuman animals and inhibit normal prosocial responses may well have played a role in their acts. Given the general attitudes of contemporary American society and the culture of the US army at the time, it is very likely that such dehumanizing cognition was manifest among the killers at My Lai. And there is reason to think such processing is implicated in atrocity; Dower (1986) has documented the role of racial indoctrination in the allied conduct of World War II in the Pacific, where their conduct was far worse than it was in the much less racialized European theater. These distal pressures, we contend, may cause cognitive degradation. We do not think this cognitive degradation is of the same type as that caused by the situational pressures of the battlefield, but we do think it merits the label. Together, distal and situational pressures motivate a strong presumption that war crimes, at least those in or near combat, are typically carried out by agents in excusing conditions; given their cognitive degradation, they cannot be reasonably expected to produce appropriate moral judgments or behavior. Thus, the perpetrators are not typically morally responsible for what they do. Therefore, they are not appropriately considered evil.

714    Dominic Murphy and John M. Doris Skepticism about Evil maintains that perpetrators of extreme wrongdoing frequently occupy excusing conditions, by virtue of dispositional factors, situational factors, or the combination of both; these circumstances, where they obtain, block the attribution of responsibility, and so the attribution of evil. Our argument, we should emphasize, is an argument for a presumption rather than a conclusion, a presumption that must be made good, or not, on a case-​by-​case basis. To our previously stated Presumption of Morally Relevant Impairments, we now add, for wrongdoing in extreme situations like warfare, The Presumption of Extenuating Circumstances. Together, we contend, these presumptions provide support for a Presumption of Non-​Responsibility—​from which, we contend, Skepticism about Evil follows. For different transgressions, different considerations may be salient. The kind of situational influences so important to understanding wartime atrocities matter less for the individual criminal violence of psychopaths, and the sort of dispositional factors central to thinking about psychopathy will often be inapplicable to the transgression of the ordinary soldier. Of course, these factors are not always readily distinguishable, since behavior, as everyone agrees (Doris 2021; Mehl, Bollich, Doris, and Vazire 2015: 630), is the result of interactions between persons and situations. Which means that very often, both of our presumptions must be evaluated, as both dispositional and situational defeaters of responsibility may be in play. But we do not, we hasten to underscore, say that all such factors are always active. Nor do we deny that in some cases none of them are in force, meaning that we are in such cases confronted with legitimate instances of evil. What we deny is that there are many such cases. Talbert and Wolfendale (2018) adopt the contrary position. Like other philosophers responding to situationism (e.g., Miller 2014; Russell 2009; Snow 2009), Talbert and Wolfendale think of behavior as emerging from an interplay of personality and situation in the tradition of Mischel’s Cognitive-​Affective Personality System (CAPS) (Mischel 1973; Mischel and Shoda 1995). The CAPS models personality as a system of situationally evoked traits that are consistent across situations seen from the perspective of the individual; in situations one construes to be of type X, one consistently acts in what one construes as an X-​appropriate way. There are reasons for thinking that this is not a satisfactory philosophical response to situationist worries about responsibility and character (see Alfano 2013, 78–​79; Doris 2002: 76–​85; Doris 2021; Papish 2017)—​not least of which is the question of whether consistency is in the eye of the construer—​but we will not rehearse them here. For the Presumption of Extenuating Circumstances is sustainable even if Talbert and Wolfendale’s account of character formation holds up. Talbert and Wolfendale (2018: 50) argue that war crimes can be explained by perpetrators’ character traits, understood as “behavior-​informing dispositions that are relatively stable over time, and that are sustained by an agent’s beliefs, desires, and values.” For the purposes of argument, we’re for the moment prepared to grant as much, and we’re further prepared to grant that the personality structures—​the CAPS—​shaped by the distal forces in military socialization are not “external or alien to individual military personnel” (Talbert and Wolfendale 2018: 50). We can also grant Talbert and

Atrocity, Evil, and Responsibility    715 Wolfendale’s (2018: 68) assertion that military personnel play an “active role” in this process, which justifies thinking of their personality system, and its attendant construals, as properly the personnel’s own. We deny however, that this account of personality and personality development entails that perpetrators are generally morally responsible for war crimes. One might object to Talbert and Wolfendale’s characterization by questioning whether the dispositions in question are appropriately thought of as “internal” to the actor in the sense that they require, since the distal forces at issue, inasmuch as they are institutionally imposed on subjects, are very naturally thought of as “external.” But there’s little profit in attempting to identify the elusive external/​internal divide, and we won’t attempt it here. Alternatively, one might object that dispositions with origins in pervasive autocratic socialization, such as that depicted by Ricks (1995) in his account of Marine Corps boot camp, make unlikely targets for attributions of individual responsibility, but as we’ve said, we’re currentists about responsibility, and are therefore bound to eschew such historical argument. Worth noting is that many perpetrators of atrocity are manifestly conflicted (Doris 2002: ch. 3), which casts empirical doubt on claims of authenticity or wholeheartedness, but neither is this our central reservation. Let that be granted: The crucial point is the Presumption of Extenuating Circumstances does not insist that perpetrators have an alienated or inauthentic personality (a disintegrated CAPS) at the moment of action, but only that their cognitive structures are often degraded at that moment. It’s tempting to argue that where cognitive degradation is sufficiently extreme, responsibility grounding identification with one’s conduct is unlikely to be achievable (cf. Doris 2015a: 31–​32), but even supposing it may obtain, if substantial cognitive degradation also obtains, the agent is plausibly supposed to be in excusing conditions, and responsibility does not obtain. Our disagreement with Talbert and Wolfendale may come down to the question of whether the distal processes are cognitively degrading in a way that indicates excuse. There’s more than one reasonable answer to this difficult question, of course, and Talbert and Wolfendale’s extended treatment deserves more discussion than we have time for here. But it does not convince us that there are not numerous and varied considerations suggesting excusing conditions for those perpetrating extreme wrongdoing. To rule out the existence of such conditions, one needs to do extensive ruling out, given the kind of extreme sanctions appropriate to evil. The result of this exercise often will, when done carefully, result in the identification of excusing, or at least mitigating, considerations; as a result, talk of evil contributes relatively little to our moral understanding of extreme wrongdoing. As we said at the outset, this conclusion apparently counsels repugnant lenience: Those who are not morally responsible are not deserving of punishment, so Skepticism about Evil would seem to entail that many people who do the most horrific of deeds are not liable to sanction. They might, of course, be subject to involuntary treatment or preventative custody, as might be appropriate for psychopaths, and the skeptic can readily allow this. But it seems that skeptics must frequently be lenient regarding questions of desert of punishment for extreme wrongdoing. In fact, we think this implication,

716    Dominic Murphy and John M. Doris however unpalatable it may be, should often be accepted. We will close, however, with some considerations that might be thought to mute this result.

4.  Collective Responsibility That the individuals who directly participate in an atrocity are not morally responsible does not entail that nobody—​or nothing—​is responsible; responsibility might appropriately affix to other associated individuals, or institutions. When it comes to excusing conditions, for example, one could reason like this: Excusing conditions have causes, so maybe the responsible parties are those who act so as to bring about the excusing conditions that are implicated in the moral failure of direct perpetrators. If we want someone to blame, it should be the individuals or institutions who exert such malign influence. In the military case, an obvious option is to ascribe responsibility to the senior officers whose conduct helps cause enormities like My Lai, either by direct causal influence such as orders, or by contributing to institutional structures that distort the cognitive and motivational structures of perpetrators. This approach would shift responsibility from the perpetrators to their superiors, as we do when we forego punishing those following orders, but hold responsible those who gave the orders. In this first case, of displaced responsibility, we identify specific men or women as guilty other than those who carried out the crime: For example, in 1947 the Yugoslav government executed Alexander Loehr, who had commanded the Luftwaffe units that bombed Belgrade in 1941. In a second approach, collective responsibility, we hold an institution responsible; as when it is argued that the US army, or US government, or even American society as a whole bears the responsibility for My Lai. Collective responsibility, especially that of institutions, often looks attractive and intuitive, but it is notoriously difficult to establish the right sort of causal relation. If Cain slays Abel, we can attribute the responsibility to Cain without further ado, but if Abel dies in an airstrike by a foreign air force, it is unclear how to operationalize responsibility in a way that lets us blame the relevant actor. (Airstrikes require flight crews, ground crews, mission planners, weapons designers, and more.) Cain is a bad actor, but in what sense is an institution or collective an actor, bad or otherwise? On some accounts of collectives, we suggest, problems concerning individual responsibility of the sort we’ve raised remain in force. Yet, attempts at formulating distinctively collective responsibility face daunting obstacles—​daunting enough, we think, that prudence dictates looking elsewhere for solutions to our difficulty. If we are to regard armed forces, or any other institutions, as responsible qua collectives, we need an account of collectives that makes these attributions of responsibility plausible. While philosophers disagree about the correct understanding of the nature and influence of social groups, there are various accounts on offer. Of special interest here are those that, like our understanding of evil, involve some variant of

Atrocity, Evil, and Responsibility    717 psychologism. These accounts maintain that responsibility-​apt, broadly mental states may be attributed to groups; if so, groups may properly be attributed evil, on the framework we have been assuming here. Tollefsen (2015: 126–​135) argues that the hallmark of holding agents responsible is the practice of engaging in dialogue with them; we attempt to reason with or persuade group agents (by writing letters of complaint to corporations, for instance) and our doing so is bound up with the reactive attitudes. Although we deal with specific individuals, it is the company on whose behalf they speak that we are trying to reach. Tollefsen thinks that the simplest way to save these phenomena is to acknowledge that corporations and other collectives or institutions are appropriate targets of reactive attitudes, and are therefore actually moral agents (although she is prepared to accept a weaker position on which they are held accountable, rather than fully responsible). One could hold (as perhaps Strawson did) that our practice of targeting the reactive attitudes as we do settles questions of responsibility, but the view favored here is that the reactive attitudes track facts about agents that bear on questions of moral responsibility, and their use ought to be disciplined by theoretical and empirical considerations. With respect to collectives, the question is what properties of collectives correspond to the relevant psychological properties that make humans responsible. Huebner (2014: 12–​13) argues that a plausible account of group mentality should be modeled on well-​supported claims from the cognitive sciences about the nature of individual cognition. On his view, members of groups can pass information and exercise control within the group, and the responsible agents are those individuals who exer­ cise control at key moments. Although Huebner is interested in group minds, his analysis treats collectives as structured systems, like individual minds. This lets us assign responsibility to key elements in the structure, which may be sub-​collectives or even individuals. It therefore suggesting displaced responsibility rather than group responsibility. Accordingly, Huebner (2014: 218–​219) argues that attempts to hold corporate bodies or other collectives responsible are often errors, because the relevant decision makers, such as CEOs, can be identified. If this is analysis is correct, the most natural way to think of military groups, given how readily commanding officers are identified as “key elements” in a command structure, is in terms of displaced responsibility, where officers may be attributed responsibility for misconduct occurring under their command. A similar conclusion appears to follow from List and Pettit’s (2011) analysis of groups. List and Pettit (2011: 33) argue that the usual way a group is formed is through the joint intention of its members to form a group around a shared goal. According to them, an agent “has representational states, motivational states, and a capacity to process them and to act on their basis,” so a group agent must possess the right representational and motivational resources (2011: 20). To form a group agent, rather than a mere collective, individuals must share more than an intention to act together; each member must intend, and all must know that the other members intend, to act together “to form and enact a single system of belief and desire” (List and Pettit 2011: 34). It must therefore be possible to combine the individual attitudes of the members of the group via

718    Dominic Murphy and John M. Doris “an aggregation function” (List and Pettit 2011: 48) into a rationally coherent set. List and Pettit (2011: 49) insist that one condition on an aggregation function must be “anonymity” in which all individuals’ attitudes receive equal weight (49). This condition is violated by any dictatorial function in which one agent gets to determine the attitudes of others. But it appears that military groups may very often exhibit this dictatorial function, as superiors’ orders (and other influences on group culture) substantially determine the attitudes, and especially the intentions, of subordinates. But if this is the case, it looks like we can, once more, assign responsibility to individuals, and needn’t brave the uncertainties of collective responsibility. There is good reason to think, then, that in the military and other strongly hierarchical contexts the most appropriate analysis is displaced responsibility, rather than collective responsibility, even if collective responsibility is appropriate for groups with less sharply delineated authority structures. But if so, this “upward displacement” does not necessarily ameliorate the uncomfortable sensation of repugnant lenience. For suppose we hold commanding officers responsible, and punish them appropriately. It remains the case, if our account is correct, that the subordinate combatants who actually did the horrific deeds are not responsible for what they did. If their non-​responsibility mandates non-​punishment, we still have an awkward forbearance, especially if we concurrently punish officers who, whatever else they did, did not themselves perform the horrific deeds.29 Circumstances are still more uncomfortable if the superior officer did not explicitly order the atrocity. Finally, even with commanding officers at a distance from the conduct conditions, we may often be left with issues of exculpating dispositional and institutional factors of the sort we have already noted. But suppose we are wrong in our suspicions, and there exists an account of collective responsibility applicable to military contexts. It is not clear that this circumstance eliminates the appearance of repugnant lenience either. First, note the complexity: If punishment is supposed to be proportionate to the gravity of the offense, and not all members of the collective contributed equally to the offense, we need a way of apportioning responsibility across the guilty group, which is likely to be a daunting proposition. And if this is done, some of the responsibility will be shifted away from the direct perpetrators, and in many such cases, their burden of punishment may seem offensively light. We do not say that this difficulty cannot be solved, in particular cases. But we think the difficulty is severe enough that if we wish to avoid the appearance of repugnant lenience, we should seriously consider an expedient that is itself unattractive: severing the connection between moral responsibility and liability to punishment.

29  Some of this awkwardness may be dispelled if we grant responsibility comes in degrees (see Nelkin 2016; Tierney 2019). But even if we attribute reduced responsibility for superior officers “indirectly” involved in an atrocity, an uncomfortable appearance of misproportion remains, if in the same instance no responsibility is attributed subordinates who were “direct” participants.

Atrocity, Evil, and Responsibility    719

5.  Strict Liability We once tentatively suggested (Doris and Murphy 2007: 51–​53) that one response to atrocity in warfare might be the application of strict liability, which we understood as “rejecting a mens rea requirement for criminal liability.” The issues in legal theory are complex,30 but our motivation was straightforward: Strict liability avoids “repugnant lenience.” On this approach, perpetrators of atrocity could be liable to punishment despite the fact that, by virtue of occupying excusing conditions, they do not meet the psychological conditions for moral responsibility attribution. Thus, war criminals might be punished for deeds they were not morally responsible for committing. Additionally, we wondered if this expedient might be extended to moral responses other than punishment, in a sort of strict liability standard for some instances of reactive attitudes like anger and outrage. In closing, we’ll—​ again with a certain ambivalence—​ expand a bit on these suggestions, turning to questions associated with proper accountability for extreme wrongdoing. We will limit discussion to situational influences on wartime behavior, and eschew discussion of dispositional influences such as psychopathy. We will again note, however, that the diagnosis of psychopathy is generally not treated by courts as a defense, a practice which, if we are right about psychopathy constituting impairment, is in serious need of a compelling rationale. It may be that some of our remarks on strict liability might inform such a rationale, but for reasons of space, we will not explore the issue here. As we admitted, our proposal that strict liability might apply to war crimes is “not . . . entirely appealing,” for “prosecuting individuals who occupy excusing conditions looks more like scapegoating than justice” (Doris and Murphy: 51). It appears that we are recommending punishment for those who do not deserve punishment, in violation of very plausible tenets of both morality and law. Do we escape repugnant lenience at the cost of reprehensible unfairness? Brink (2013b, 2019) says as much: John Doris and Dominic Murphy appeal to situationist psychology to claim that we should offer a wide-​ranging excuse for wartime wrongdoing. They try to avoid the unwelcome consequences of this kind of promiscuity about excuse by endorsing a form of strict liability that would punish despite the existence of an excuse. But this compounds one mistake—​an insufficiently discriminating conception of excuse—​with another—​the failure to recognize that excuse is a true defense that justifies acquittal. We can easily avoid the second mistake by not making the first one. (Brink 2019) [31] 30 

For some discussion, see Duff (2006) and Larkin (2014). Brink (2019) also contends that our understanding of strict liability is absent from actual criminal law. We get the contrary impression from Moore (2018: 527–​528). But in any event, while we defer to 31 

720    Dominic Murphy and John M. Doris We are, according to Brink, applying a misbegotten solution to a chimerical problem. Regarding the first issue, our promiscuity with excuses, we haven’t much to add to our earlier discussion than we’ve already added before. It is worth stressing, however, that most important to our arguments are not the experiments of “situationist psychology,” which it might be tempting to dismiss (wrongly, we think32) as unreplicable or “cute.” Instead, at the core of our argument is the hard facts of warfare and combat. There’s nothing cute about these, and they’ve replicated, over and over, across the whole bloody expanse of human history. In many cases, to call such horrific pressures “extreme” risks a gross understatement, and the effects can be described, quite without hyperbole, as cognitive and emotional catastrophe. It’s not “promiscuous” to identify excusing conditions in such circumstances; on the contrary, it seems to us, such findings are an altogether expectable, and defensible, outcomes of a sober look at the facts. Careful inquiry into particular cases may fail to identify excusing conditions, and the likelihood of this result increases, as Brink (2013b: 144) would maintain, where circumstances diverge from the apocalyptic chaos of “hot” combat. But as we’ve said, we’re arguing for a presumption, rather than a conclusion, and the Presumption of Non-​Responsibility, it still seems to us, is reasonably applied to many conditions of war. One reason for resisting this presumption is its normative implications, so let’s turn to what Brink thinks is our second mistake, the invocation of strict liability. Is excuse a “true defense” that always militates against punishment? There might, of course, be consequentialist, forward-​looking justifications for a strict liability approach: Perhaps a zero-​tolerance policy mandating punishment, even for those perpetrators of atrocity who occupied excusing conditions, is justified by its deterrent value. One could object to this on purely ethical grounds: Divorcing punishment from desert in this way—​making it “telishment” in Rawls’s (1955: 11) memorable term—​might be morally objectionable, whatever the consequences, particularly for those favoring a retributivist approach to punishment (e.g., Morris 1968; Murphy 1971). Our own hesitation was more pragmatic: We’re not confident such a policy would have the desired deterrent effect, especially for soldiers in combat (Doris and Murphy 2007: 52).33 Our argument, roughly, is that the conditions of combat destabilize morally responsible agency, and why think people with their agency destabilized will respond in prudentially appropriate ways to the incentives of deterrence? Another possible rationale for the application of strict liability in this context may be found in communicative theories of punishment, which justify punishment by virtue of the message it conveys to the criminal or other affected parties. Communicative theories of punishment are often associated with retributivism, but they needn’t be (Duff 2003: 388), and there may be good consequentialist reasons for punishments understood Brink’s expertise in the state of the law, our concern is primarily with the moral implications of our notion of strict liability, rather than the extent to which is manifest in current legal practice. 32  33 

For discussion, see Doris 2005; 2021. For criticism of the deterrent potential of strict liability, see Robinson (2018).

Atrocity, Evil, and Responsibility    721 as communicative acts. For war crimes, a nation or other legal entity might punish combatants who perpetrated atrocity while occupying excusing conditions in order to communicate to victims and other members of the international community that the outrage is recognized and will not be tolerated. This might have desirable consequences, such helping to repair relations between formerly warring nations. But it might be morally important independent of consequences. For example, perhaps it is sometimes morally obligatory to express outrage irrespective of consequences; someone might be morally criticizable for failing to express outrage at the Holocaust even if that expression is of negligible consequence. For atrocity, a sufficient expression of outrage might require substantial punishment. One can see grotesquerie here. (Much too often, calls to “make an example of ” someone bear an unseemly stench.) We see it too (and we’re sure Brink does as well). But the alternative to the morally grotesque expedient of punishing the excused is the morally grotesque expedient of not punishing those who perpetrate horrors on their fellow human beings. There is no good alternative here: Only alternatives that are, at best, a little less shitty than others. If we, in responding to atrocity on the part of our armed services, walk away feeling as though we responded well, there’s a very good chance we walk away delusionally. And that, in the end, is the take home from our ambivalent ruminations about strict liability. Increasingly, clinicians working on trauma in military populations have turned to the construct of moral injury to understand the damage done to combatants, damage done not only by witnessing or suffering harm, but also damage done by inflicting it (e.g., Flipse Vargas, Hanson, Kraus, Drescher, and Foy 2013; Griffin, Purcell, Burkman, Litz, Bryan, Schmitz . . . and Maguen 2019; Shay 2014; Sherman 2015). This injury will inevitably be inflicted on many of our fellow citizens when we send them to war. If nothing else, we hope that our discussion serves to underscore that fact.

6.  Conclusion Our argument for Skepticism about Evil is based on the observation that warranted attributions of evil require warranted attributions of moral responsibility, together with the Presumption of Non-​Responsibility, which maintains that for typical putative instances of evil, attributions of moral responsibility are not warranted. We developed this Presumption of Non-​Responsibility by defending two supporting presumptions, The Presumption of Morally Relevant Impairment, and The Presumption of Extenuating Circumstances. This argument does not entail that there are no instances of genuine evil, but it does suggest that such occurrences are unlikely enough that the discourse of evil, so pervasive in responses to extreme wrongdoing, is likely to be badly misleading. As to what responses would be more appropriate, our assessment is, unfortunately, pessimistic. There is unlikely to be a morally satisfying response to moral horror—​yet another

722    Dominic Murphy and John M. Doris reason, in the unlikely event more reasons are needed, to work harder at preventing the horrors in which we are too often complicit.

References Aharoni, E., Sinnott-​Armstrong, W., and Kiehl, K. A. (2012). Can psychopathic offenders discern moral wrongs? A new look at the moral/​conventional distinction. Journal of Abnormal Psychology, 121(2), 484–​497. Alfano, M. (2013). Character as moral fiction. Cambridge: Cambridge University Press. Allison, W. T. (2012). My Lai: An American atrocity in the Vietnam War. Baltimore: Johns Hopkins University Press. American Psychiatric Association. (1983). Statement on the insanity defense. American Journal of Psychiatry, 140, 681–​688. Baik, T-​U. (2001). A war crime against an ally’s civilians: The No Gun Ri massacre, Notre Dame Journal of Law, Ethics & Public Policy, 15, 455–​505. Barnes, E. (2016). The minority body: A theory of disability. Oxford: Oxford University Press. Bartov, O. (2001). Hitler’s army: Soldiers, Nazis, and war in the Third Reich. Oxford: Oxford University Press. Bilton, M., and Sim, K. (1993). Four hours in My Lai. Penguin. Blair, J., Jones, L., Clark, F., and Smith, M. (1997). The psychopathic individual: A lack of responsiveness to distress cues? Psychophysiology, 34(2), 192–​198. Blair, J., Mitchell, D., and Blair, K. (2005). The psychopath: Emotion and the brain. Blackwell Publishing. Blair, R. J. R. (1995). A cognitive developmental approach to morality: Investigating the psychopath. Cognition, 57(1), 1–​29. Blair, R. J. R., Jones, L., Clark, F., and Smith, M. (1995). Is the psychopath ‘morally insane’? Personality and Individual Differences, 19(5), 741–​752. Borg, J. S., and Sinnott-​Armstrong, W. (2013). Do psychopaths make moral judgments? In K. A. Kiehl, & W. P. Sinnott-​Armstrong (Eds.), Handbook on Psychopathy and Law, 107–​128. Oxford: Oxford University Press. Brink, D. O. (2013a). Responsibility, incompetence, and psychopathy. The Lindley Lectures, 53, 13–​129. Brink, D. O. (2013b). Situationism, responsibility, and fair opportunity. Social Philosophy and Policy, 30(1–​2), 121–​149. Brink, D. O. (2019). The nature and significance of culpability. Criminal Law and Philosophy, 13(2), 347–​373. Buss, S. (1997). Justified wrongdoing. Noûs, 31(3), 337–​369. Caldwell, M. (2013). Treatment of adolescents with psychopathic features. In K. A. Kiehl and W. P. Sinnott-​Armstrong (Eds.), Handbook on Psychopathy and Law, 201–​228. Oxford: Oxford University Press. Cima, M., Tonnaer, F., and Hauser, M. D. (2010). Psychopaths know right from wrong but don’t care. Social Cognitive and Affective Neuroscience, 5(1), 59–​67. Cleckley, H. (1964). The mask of sanity: An attempt to clarify some issues about the so-​called psychopathic personality. (4th ed.) CV Mosby. Originally published 1941. D’Arms, J., and Jacobson, D. Forthcoming. Rational sentimentalism. Oxford: Oxford University Press. De Wijze, S. (2002). Defining evil: Insights from the problem of “dirty hands.” The Monist, 85(2), 210–​238.

Atrocity, Evil, and Responsibility    723 Doris, J. M. (2002). Lack of character: Personality and moral behavior. New York: Cambridge University Press Doris, J. M. (2005). Replies: Evidence and sensibility. Philosophy and Phenomenological Research, 72, 656–​677. Doris, J. M. (2015a). Talking to our selves: Reflection, ignorance, and agency. Oxford: Oxford University Press. Doris, J. M. (2015b). Doing without (arguing about) desert. Philosophical Studies, 172(10), 2625–​2634. Doris, J. M. (2018). Ironic deliberations: A (regrettably incomplete) response to Fischer, Nelkin, and Vargas. Social Theory and Practice, 44(2), 279–​296. Doris, J. M. 2021. Character trouble: Undisciplined essays on moral personality should be agency and personality. Oxford: Oxford University Press Doris, J. M., and Murphy, D. (2007). From My Lai to Abu Ghraib: The moral psychology of atrocity. Midwest Studies in Philosophy, 31(1):25–​55. Dower, J. W. (1986). War without mercy: Race and power in the pacific war. WW Norton. Duff, A. (2003). Punishment, communication, and community. In D. Matravers and J. Pike (Eds.), Debates in Contemporary Political Philosophy: An Anthology, 387–​407. London and New York: Routledge. Duff, A. (2009). Legal and moral responsibility. Philosophy Compass, 4(6), 978–​986. Ermer, E., Cope, L. M., Nyalakanti, P. K., Calhoun, V. D., and Kiehl, K. A. (2012). Aberrant paralimbic gray matter in criminal psychopathy. Journal of Abnormal Psychology, 121(3), 649–​658. Ermer, E., Cope, L. M., Nyalakanti, P. K., Calhoun, V. D., and Kiehl, K. A. (2013). Aberrant paralimbic gray matter in incarcerated male adolescents with psychopathic traits. Journal of the American Academy of Child & Adolescent Psychiatry, 52(1), 94–​103. Fischer, J. M. 2006. My way: Essays on moral responsibility. New York: Oxford University Press. Fischer, J.M., and Ravizza, M. (1998). Responsibility and Control. Cambridge: Cambridge University Press. Flipse Vargas, A., Hanson, T., Kraus, D., Drescher, K., and Foy, D. (2013). Moral injury themes in combat veterans’ narrative responses from the National Vietnam Veterans’ Readjustment Study. Traumatology, 19(3), 243–​250. Forth, A. E., Bo, S., and Kongerslev, M. (2013). Assessment of psychopathy: The Hare Psychopathy Checklist measures. In K. A. Kiehl and W. P. Sinnott-​Armstrong (Eds.), Handbook on Psychopathy and Law, 5–​33. Oxford: Oxford University Press. Frésard, J-​J. (2004). The roots of behavior in war: A survey of the literature. Geneva: International Committee of the Red Cross. Garcia, E. V. (2002). A Kantian theory of evil. The Monist, 85(2), 194–​209. Gibbard, A. (1990). Wise choices, apt feelings: A theory of normative judgment. Cambridge: Cambridge University Press. Greene, J. (2014). Moral tribes: Emotion, reason, and the gap between us and them. Penguin. Griffin, B. J., Purcell, N., Burkman, K., Litz, B. T., Bryan, C. J., Schmitz, M., . . . and Maguen, S. (2019). Moral injury: An integrative review. Journal of Traumatic Stress. Hare, R. D. (1993). Without conscience: The disturbing world of the psychopaths among us. Guilford Press. Hare, R.D. (2003). The Hare Psychopathy Checklist-​Revised (2nd ed.). Toronto, ON: Multi-​ Health Systems. Haybron, D. M. (2002). Moral monsters and saints. The Monist, 85(2), 260–​284. Huebner, B. (2013). Macrocognition. Oxford: Oxford University Press.

724    Dominic Murphy and John M. Doris Ingierd, H and Syse, H. (2005). Responsibility and culpability in war. Journal of Military Ethics, 4, 85–​99. Jalava, J., and Griffiths, S. (2017). Philosophers on psychopaths: A cautionary tale in interdisciplinarity. Philosophy, Psychiatry, & Psychology, 24(1), 1–​12. Jalava, J., Griffiths, S., and Maraun, M. (2015). The myth of the born criminal: Psychopathy, neurobiology, and the creation of the modern degenerate. Toronto, ON: University of Toronto Press. Kiehl, K. A. (2014). The psychopath whisperer: The science of those without conscience. Broadway Books. Kiehl, K. A., and Sinnott-​Armstrong, W. P. (2013). Introduction. In K. A. Kiehl and W. P. Sinnott-​ Armstrong (Eds.), Handbook on Psychopathy and Law, 1–​2. Oxford: Oxford University Press. Koenigs, M., 2012. The role of prefrontal cortex in psychopathy. Reviews in the Neurosciences, 23, 253–​262. Koenigs, M., and Newman, J. P. (2013). The decision-​making impairment in psychopathy: Psychological and neurobiological mechanisms. In K. A. Kiehl and W. P. Sinnott-​Armstrong (Eds.), Handbook on Psychopathy and Law, 93–​106. Oxford: Oxford University Press. Kornblith, H. (2012). On reflection. Oxford: Oxford University Press. Larkin, P. J. (2014). Strict liability offenses, incarceration, and the cruel and unusual punishments clause. Harvard Journal of Law and Public Policy, 37(3), 165–​1121. List, C and Pettit, P. (2011) Group agency: The possibility, design, and status of corporate agents. Oxford: Oxford University Press. Litton, P. (2007). Responsibility status of the psychopath: On moral reasoning and rational self-​governance. Rutgers Law Journal, 39, 349–​392. Luna, E. (2013). Psychopathy and sentencing. In K. A. Kiehl and W. P. Sinnott-​Armstrong, (Eds.), Handbook on Psychopathy and Law, 358–​369. Oxford: Oxford University Press. Maibom, H. L. (2008). The mad, the bad, and the psychopath. Neuroethics, 1(3), 167–​184. Maibom, H. (2018). What can philosophers learn from psychopathy? European Journal of Analytic Philosophy, 14(1), 63–​78. McGinn, C. (1997). Ethics, evil, and fiction. Oxford: Clarendon Press. McKenna, M., and Van Schoelandt, C. (2015). Crossing a mesh theory with a reasons-​ responsive theory: Unholy spawn of an impending apocalypse or love child of a new dawn?” In A. Buckareff, C. Moya, and S. Rosell (Eds.), Agency, Freedom, and Moral Responsibility, 44–​64. New York: Springer. McMahan, J. (2009). Killing in war. Oxford: Oxford University Press. Mehl, M. R., Bollich, K. L., Doris, J. M., and Vazire, S. (2015). Character and coherence: Testing the stability of naturalistically observed daily moral behavior.” In C. Miller, M. R. Furr, A. Knobel, and W. Fleeson (Eds.), Character: New Directions from Philosophy, Psychology, and Theology, 630–​651. Oxford: Oxford University Press. Miller, C. B. (2014). Character and moral psychology. Oxford: Oxford University Press. Mischel, W. (1973). Toward a cognitive social learning reconceptualization of personality. Psychological Review, 80, 252–​283. Mischel, W. and Shoda, Y. (1995) A cognitive-​ affective system theory of personality: Reconceptualizing situations, dispositions, dynamics, and invariance in personality structure. Psychological Review, 102(2), 246–​268. Mokros, A., Hare, R. D., Neumann, C. S., Santtila, P., Habermeyer, E., and Nitschke, J. (2015). Variants of psychopathy in adult male offenders: A latent profile analysis. Journal of Abnormal Psychology, 124(2), 372. Moore, M. S. (2018). The strictness of strict liability. Criminal Law and Philosophy, 12(3), 513–​529.

Atrocity, Evil, and Responsibility    725 Morris, H. (1968). Persons and punishment. The Monist, 52(4), 475–​501. Morse, S. J. (2002). Uncontrollable urges and irrational people. Virginial Law Review, 88, 1025–​1078. Morse, S. J. (2008). Psychopathy and criminal responsibility. Neuroethics, 1, 205–​212. Morse, S. J. (2013). Preventive detention of psychopaths and dangerous offenders. In K. A. Kiehl and W. P. Sinnott-​Armstrong (Eds.), Handbook on Psychopathy and Law, 321–​345. Oxford: Oxford University Press. Murphy, D. (2006). Psychiatry in the scientific image. Cambridge, MA: MIT Press. Murphy, J. G. (1971). Three mistakes about retributivism. Analysis, 31(5), 166–​169. Nichols, S. (2004). Sentimental rules: On the natural foundations of moral judgment. Oxford: Oxford University Press. Nichols, S., and Vargas, M. (2007). How to be fair to psychopaths. Philosophy, Psychiatry, & Psychology, 14(2), 153–​155. Nelkin, D. K. (2011). Making sense of freedom and responsibility. Oxford: Oxford University Press. Nelkin, D. K. (2015). Psychopaths, incorrigible racists, and the faces of responsibility. Ethics, 125(2), 357–​390. Nelkin, D. K. (2016). Difficulty and degrees of moral praiseworthiness and blameworthiness. Noûs, 50(2), 356–​378. Olin, L., and Doris, J. M. (2014). Vicious minds: Virtue epistemology, cognition, and skepticism. Philosophical Studies, 168(3), 665–​692. Osiel, M. (1999). Obeying orders: Atrocity, military discipline and the law of war. Transaction Publishers. Papish, L. (2017). CAPS psychology and the empirical adequacy of Aristotelian virtue ethics. Ethical Theory and Moral Practice, 20(3), 537–​549. Pereboom, D. (2001). Living without free will. Cambridge: Cambridge University Press. Pereboom, D. (2014). Free will, agency, and meaning in life. Oxford: Oxford University Press. Perrett, R. W. (2002). Evil and human nature. The Monist, 85(2), 304–​319. Pillsbury, S. H. (2013). Why psychopaths are responsible. In K. A. Kiehl and W. P. Sinnott-​Armstrong (Eds.), Handbook on Psychopathy and Law, 297–​318. Oxford: Oxford University Press. Pols, H. (2011) The Tunisian campaign, war neuroses, and the reorientation of American psychiatry during World War II. Harvard Review of Psychiatry, 19, 313–​320. Pols, H., and Oak, S. (2007). War & military mental health: The US psychiatric response in the 20th century. American Journal of Public Health, 97(12), 2132–​2142. Prinz, J. J., and Nichols, S. 2010. Moral emotions. In J. M. Doris and The Moral Psychology Research Group (Eds.), The Moral Psychology Handbook, 111–​ 146. Oxford: Oxford University Press. Raine, A. (2013). The anatomy of violence: The biological roots of crime. Vintage. Rawls, J. (1955). Two concepts of rules. The Philosophical Review, 64, 3–​32. Ricks, T (1997) Making the corps. New York: Scribner Rice, M. E., and Harris, G. T. (2013). Psychopathy and violent recidivism. In K. A. Kiehl and W. P. Sinnott-​Armstrong (Eds.), Handbook on Psychopathy and Law, 231–​249. Oxford: Oxford University Press. Robinson, P. H. (2018). Strict liability’s criminogenic effect. Criminal Law and Philosophy 12(3), 411–​426. Russell, D. C. (2009). Practical intelligence and the virtues. Oxford: Oxford University Press. Shay, J. (2014). Moral injury. Psychoanalytic Psychology, 31(2), 182–​181. Sher, G. (1987). Desert. Princeton, NJ: Princeton University Press.

726    Dominic Murphy and John M. Doris Sherman, N. (2015). Afterwar: Healing the moral wounds of our soldiers. Oxford: Oxford University Press. Shoemaker, D. (2011a). Attributability, answerability, and accountability: Toward a wider theory of moral responsibility. Ethics, 121(3), 602–​632. Shoemaker, D. W. (2011b). Psychopathy, responsibility, and the moral/​conventional distinction. The Southern Journal of Philosophy, 49(s1), 99–​124. Shoemaker, D. (2015). Responsibility from the margins. Oxford: Oxford University Press. Smith, D.L. (2012). Less than human. Griffin. Snow, N. E. (2010). Virtue as social intelligence: An empirically grounded theory. Routledge. Steiner, H. (2002). Calibrating evil. The Monist, 85(2), 183–​193. Stone, M. H. (2009). The anatomy of evil. Prometheus Books. Swinburne, R. (1978). Natural evil. American Philosophical Quarterly 15(4), 295–​301. Talbert, M. (2008). Blame and responsiveness to moral reasons: Are psychopaths blameworthy? Pacific Philosophical Quarterly, 89(4), 516–​535. Talbert, M. and Wolfendale, J. (2018). War crimes: Causes, excuses, and blame. Oxford: Oxford University Press. Tierney, H. (2019). Quality of reasons and degrees of responsibility. Australasian Journal of Philosophy, 97(4), 661–​672. Tilly, C. (1990). Coercion, capital, and European states, AD 990–​1990. Blackwell. Thompson, P. (2002). The evolutionary biology of evil. The Monist, 85(2), 239–​259. Tollefsen, D. (2015). Groups as agents. Cambridge: Polity Press. Tremain, S. (2017). Foucault and feminist philosophy of disability. Ann Arbor: University of Michigan Press. Vargas, M. R. (2013). Building better beings: A theory of moral responsibility. Oxford: Oxford University Press. Vargas, M. R. (2015). Desert, responsibility, and justification: A reply to Doris, McGeer, and Robinson. Philosophical Studies, 172(10), 2659–​2678. Viding, E., Fontaine, N. M., and Larsson, H. (2013). Quantitative genetic studies of psychopathic traits in minors: Review and implications for the law. In K. A. Kiehl and W. P. Sinnott-​Armstrong, (Eds.), Handbook on Psychopathy and Law, 93–​106. Oxford: Oxford University Press. Wallace, R. J. (1994). Responsibility and the moral sentiments. Cambridge, MA: Harvard University Press. Watson, G. (1993). Responsibility and the limits of evil: Variations on a Strawsonian theme. In J. M. Fischer and M. Ravizza (Eds.), Perspectives on Moral Responsibility, 119–​149. Ithaca, NY: Cornell University Press. Wolf, S. (1990). Freedom within reason. Oxford: Oxford University Press. Wonderly, M. (2020). Psychopathy, agency, and practical reason. In R. Chang and K. Sylvan (Eds.), Routledge Handbook of Practical Reason. New York: Routledge. Woolfolk, R. L. (1999). Malfunction and mental illness. The Monist, 82(4), 658–​670. Yang, Y. and Raine, A. (2009). Prefrontal structural and functional brain imaging findings in antisocial, violent, and psychopathic individuals: a meta-​analysis. Psychiatry Research: Neuroimaging, 174(2), 81–​88.

Index

Due to the use of para id indexing, indexed terms that span two pages (e.g., 52–​53) may, on occasion, appear on only one of those pages.

A

B

Adams, M.M., 593, 599, 602–​3 Adams, R.M., 53n.2, 316 Addenbrooke, M., 650–​51 Aharoni, E., 399–​400, 705n.20 Alcoff, L., 680n.34 Alexander, J., 494n.1 Alexander, L., 236 Alexander, M., 380–​81 Alfano, M., 471n.8, 714 Allais, L., 368n.1, 592, 594, 595, 596, 598, 605, 618n.50 Allison, W.T., 710 Alloway, R., 458 Alloway, T., 458 Allport, G.W., 669n.3 Alston, W., 628n.5 Alvarez, M., 136n.13 Amodio, D.M., 672n.8 Amstutz, M., 604n.19 Anderson, E., 414, 416–​17, 418, 688 Andow, J., 502 Anglin, W.S., 550n.10 Annas, J., 471n.8 Antony, L.M., 683 Arneson, R., xxix, 6, 8n.5, 10–​11, 13, 14–​15, 19, 20, 22, 23, 414, 419, 556n.24 Aronson, J., 679–​80 Arpaly, N., 33n.10, 53n.2, 120, 120n.19, 120n.20, 121–​22, 122n.24, 195, 197, 317, 337n.15, 337n.17, 338n.18, 338n.19, 341n.23, 425, 426, 455, 648, 652n.5, 657n.9, 673–​74 Audi, R., 251–​52 Austin, J.L., 39n.20, 98 Awh, E., 458 Ayer, A.J., 180, 181–​82

Baars, B.J., 448–​49 Bach, K., 93n.4 Baddeley, A., 458 Baik, T., 710n.26 Baker, L.R., 189 Balaguer, M., 49, 508n.17, 508n.18, 508n.21 Balicki, A., 252 Ballantyne, N., 469n.1 Banaji, M., 451, 460, 669n.3, 671–​72 Bargh, J., 451n.7 Barkan, E., 618n.49 Barnard, M., 647 Barnes, E., 697n.2 Barnes, S., 400 Baron, M., 475 Baron, R., 470 Baron-​Cohen, S., 391n.6 Bartov, O., 711–​12 Bash, A., 593 Bassetti, C., 457 Batson, C., 469–​70 Baumeister, R.F., 216, 444 Bechara, A., 657n.9 Beebee, H., 354 Bell, M., 278n.10, 279n.11, 560n.37, 562, 562n.43, 593, 595 Benbaji, Y., 416 Bennett, J., 10, 275n.7, 315, 338n.18, 567, 568n.3, 599, 601, 602–​3, 618n.47 Berlin, I., 205 Berman, M., 225–​26, 379n.11, 382, 599 Berofsky, B., 437–​38 Berridge, K.C., 652–​53 Bhargava, R., 612n.13

728   Index Bidshahri, R., 82 Bilton, M., 710 Bingeman, E., 559n.34 Björnsson, G., xxxi, 141n.25, 188, 495n.3, 498–​ 99, 501n.8, 503n.11, 503–​4, 504n.13, 507n.16, 509n.23, 511n.27, 593 Blackstone, W., 368, 368n.2 Blackwelder, J., 476 Blair, J., 389, 391n.6, 399–​400, 401, 402–​4, 704–​5 Block, N., 150, 150n.47, 448–​49 Blum, L., 680n.32 Blustein, J., 595, 616n.34 Bodenhausen, G., 460–​61 Boehm, C., 252 Bohman, J., 680n.34 Bollard, M., 391n.6 Borg, J.S., 705, 708 Bourget, D., 511 Boxill, B., 578n.11 Boyd, R., 251–​52, 257, 258n.20 Bradford, G., 129n.36 Bradley, F.H., 205 Braham, M., 131n.2, 145n.34 Braithwaite, J., 261n.23, 612n.12, 620n.56 Brandon, T., 651 Bratman, M.E., 138n.16, 140n.22 Brennan, S., 678–​79 Brink, D., xxviii–​xxix, 21, 21n.17, 31n.8, 48, 111–​12n.3, 251–​52, 270n.2, 366, 373, 374n.5, 375, 381n.12, 383–​84, 471n.9, 474–​75, 477n.17, 478, 478n.18, 479n.19, 482, 483, 483n.27, 485, 486, 546n.5, 552n.16, 555n.21, 701, 705n.20, 706, 707–​8, 708n.23, 719–​20, 719–​20n.31 Bronson, J., 647 Brooks, R.L., 618n.49 Brown, E., 558n.31 Brown, J.D., 216 Brownstein, M., 668, 671–​72, 671n.6, 672n.7, 675–​76, 677, 686, 689–​90 Brudner, A., 397–​98 Brunning, L., 600n.16, 605 Buckwalter, W., 494n.1, 502 Buell, S.W., 131n.2 Buss, S., 327–​28n.1, 706 Butler, J., 592, 611n.9 Byrd, J., 97n.15, 172n.2

C Caldwell, M., 707 Calhoun, C., 592, 593–​94, 602, 668, 670–​7 1, 685, 687–​91, 687n.52 Camp, V., 242–​43 Campbell, C., 437–​38 Campbell, J.K., 31n.8, 35n.15, 37–​38 Cann, A., 476 Canon, L., 470 Capes, J., 161, 162n.1, 168–​69, 187, 352n.3 Carey, J.M., 234, 235n.5 Carlson, B.E., 650–​51 Carruthers, P., xxx, 448, 458n.13, 458n.15, 460–​61 Caruso, G., xx–​xxi, xxv, 4, 5, 9n.6, 20, 21n.17, 204, 222–​25, 222n.1, 226–​29, 238, 242, 370–​ 71, 398, 451n.7, 528n.11, 528n.13, 528–​29n.14, 535–​36 Ceci, S.J., 682n.43 Chalmers, D.J., 149n.42, 511 Chan, H., 495n.3, 498, 505n.14 Chandler, D., 616n.32 Chandler, J., 495n.2 Chaplin, R., 559n.35 Chasteen, A.L., 679–​80 Chisholm, R., 628 Choi, S., 44–​45n.27 Christie, N., 261n.23, 620n.56 Cima, M., 705 Ciurria, M., 471n.9 Clark, C.J., 501n.9 Clark, M., 312 Clark, T.W., 673–​74 Clarke, R., xxi, 28n.3, 31n.8, 39, 39n.20, 44–​ 45n.27, 45n.30, 47–​48n.36, 92n.2, 92n.3, 93n.5, 94n.6, 94n.7, 95n.12, 97n.15, 98n.18, 102n.29, 105n.36, 171–​72, 172n.2, 173–​74, 317, 351, 631n.9 Cleckley, H., 703–​4 Clermont, K., 396–​97 Coates, J.C., xxi–​xxii, 112n.6, 116n.12, 116n.14, 128, 270n.1, 272n.4, 275n.7, 279, 291n.4, 292n.5, 299n.13, 299–​300n.15, 479n.19, 545n.4, 551n.13, 556n.26, 560n.38, 572n.8, 581n.12, 597n.7 Cogley, Z., 273–​74n.5 Coguen, S., 680–​81

Index   729 Cohen, G.A., 281, 414, 415, 417 Collins, R.L., 661 Collins, S., 131n.2 Colvin, E., 133n.6, 133n.7 Connelly, K., 561n.41 Cook, B.L., 656 Copp, D., 131n.1 Corbetta, M., 458 Cornell, N., 596, 599 Corntassel, J., 613n.17 Corrado, M., 222–​23 Correll, J., 681–​82 Cova, F., 495n.2, 501n.9, 502 Crick, F., 150n.45, 234–​35 Crocker, D.A., 612n.11 Cushman, F., 403–​4, 497n.6, 511n.27 Cuypers, S.E., 536n.26 Cyr, T., 47, 47–​48n.36

D Dadds, M.R., 402 Daly, M., 263 Damasio, A., 374n.7 Daniels, N., 5 Darke, S., 651 Darley, J., 469–​70 D’Arms, J., 311–​12, 314, 708 Darwall, S.L., 16–​17, 17n.13, 275, 277, 309, 314, 558n.33 Dasgupta, N., 655–​56 Davidson, D., 31n.7, 92n.2 Davies, W., 661n.14 Dawel, A., 402 De Brigard, F., 498 Deery, O., 190, 191, 192, 193n.5, 193, 498, 503n.11, 507n.16, 509n.23 Dehaene, S., 448–​49, 458 Deigh, J., 391–​93, 405 Della Rocca, M., 160 Demetriou, K., 189–​90 Dennett, D., xix, 3, 20, 33n.10, 139n.19, 477n.17, 524–​28, 527n.8, 530–​31, 537n.27 Dennis, M.L., 651 Deutschländer, R., 439 Devine, P.G., 671–​72, 673, 675–​76 De Wijze, S., 699 Dietrich, F., 510n.24

Dill, B., 652–​53, 653n.6, 654 Dillon, R.S., 604n.19 Dineen, K., 391n.6 Doris, J., xxxiv, 13, 16n.12, 18–​19, 21–​22, 111–​ 12n.3, 251–​52, 251n.10, 469n.2, 471, 471n.7, 471n.9, 477n.15, 487n.30, 699n.5, 700, 701n.10, 706, 709–​11, 712, 712n.28, 714, 715, 719, 720n.32, 720 Double, R., 5, 213, 215–​16n.11, 224, 507n.16 Dowe, P., 95n.12, 96n.14, 107n.39, 107n.40, 354 Dower, J.W., 713 Downie, R.S., 593 Drayton, L.A., 401–​2 Driver, J., 593 Duff, A., 281, 368n.1, 384n.14, 393, 396, 406, 719n.30, 720–​21 Duggan, T., 477n.17 Duijf, H., 131n.2, 135n.10 Dworkin, G., 7, 10, 419 Dyzenhaus, D., 614n.24, 615n.27, 616n.36

E Eggleston, B., 416 Ekstrom, L., 550n.10 Elster, J., 216 Elzein, N., 555n.21 Erksine, T., 131n.1 Ermer, E., 705 Estlund, D.M., 510n.24 Eyal, N., 420, 421, 422

F Faden, R., 228–​30 Fara, M., 35n.15, 37–​38, 38n.18, 44–​45n.27, 171 Faraci, D., 317 Faucher, L., 672–​73 Fehr, R., 594n.4 Feinberg, J., 131n.1, 179, 236, 260–​61, 271–​72, 351, 352n.2 Feldman, F., 699–​700 Feltz, A., 187–​88, 214n.10, 444n.10, 501, 501n.9 Ferzan, K., 236 Fine, A., 140n.21, 391–​92, 393, 397–​98, 399, 405n.14 Fingarette, H., 650–​51

730   Index Fischer, J.M., xvii, 32n.9, 33n.10, 35–​36, 36n.16, 40n.21, 40–​41, 42–​43, 47, 49, 77, 80n.3, 85, 96n.13, 97n.15, 97n.16, 98n.20, 126n.32, 127, 127n.33, 158–​61, 163, 165–​66, 168, 172, 172n.2, 173, 181–​82, 189, 248n.3, 249n.5, 270n.2, 306, 315, 318, 348n.1, 360, 373, 393, 437–​38, 453, 477–​ 78, 477n.17, 479n.19, 479n.20, 482n.25, 488, 504n.13, 552n.17, 634n.14, 701n.9, 701n.10, 706 FitzPatrick, W., 330n.5, 345n.31 Flanagan, O., 296n.10, 469n.2, 471n.7, 471n.8, 651 Flaspohler, P., 619n.52 Flint, T., 635 Focquaert, F., 391–​92 Foddy, B., 650–​51 Forscher, P.S., 669n.2, 671n.6, 681n.35, 683n.45, 683n.46, 689 Forth, A.E., 702 Fox, A.R., 389 Fox, K., 459 Frankfurt, H., xvii, 32n.9, 34–​35, 39, 48–​49, 64n.16, 94n.6, 95n.9, 96–​97, 97n.15, 136n.13, 157, 158–​59, 160, 172n.2, 180, 181–​82, 249, 249n.5, 249n.6, 355–​56, 360n.7, 453, 502, 507, 575 Frankish, K., 672n.7 Franklin, B., 368n.2 Franklin, C., 47, 47–​48n.36, 49n.38, 160, 161, 275n.7 Freeman, M., 614n.24 Freeman, S., 342n.25 Freeman, V., 613n.17 French, P.A., 131n.1, 137n.14, 142n.28, 621n.58 Frésard, J., 712–​13 Frick, J., 281n.12 Fricke, 618n.51 Fricker, M., 5, 11–​12, 275–​77, 283, 602, 677n.23, 680–​81 Fridja, N., 314n.11, 320 Fried, I., 433n.1, 433, 433n.2, 434–​35 Friedman, M., 280 Funk, F., 23n.20

G Galinsky, 671–​72 Gamlund, E., 593n.3 Garcia, E.V., 592n.2, 597–​98, 602, 699 Gardner, J., 205

Gardoni, P., 229 Garrard, E., 592, 618n.51 Gawronski, B., 460–​61 Gazzaley, A., 458 Gellner, E., 253 Gelman, S., 461 Gert, B., 477n.17 Gibbard, A., 704 Gibson, J.L., 617n.40 Gilbert, M., 140n.22, 140n.24 Gill, M.B., 388n.1, 392–​93 Ginet, G., 28n.3, 31n.8, 34n.12, 36, 49, 97n.15, 159, 160, 161, 163, 169, 172n.2 Glannon, W., 406 Glenn, A., 391–​92, 406 Glick, D., 193n.5 Glover, J., 271–​72 Godman, M., 400–​1 Goldin, C., 682 Goldman, D., 573 Goleman, D., 216 Gomperz, H., 247, 247n.*, 260–​61 Goodale, M., 448 Goodwin, R.E., 496n.4, 510n.24 Govier, T., 610n.4, 610n.5, 613n.20, 616n.31, 616n.32, 616n.33, 619n.53, 619n.54 Graham, J., 310–​11 Graham, P., 273–​74n.5, 352n.3, 552n.17 Green, J.R., 139n.17 Greene, J., 708 Greenwald, A., 451, 460, 460n.17, 681–​82 Griffin, B.J., 721 Griffiths, S., 389, 395, 400, 406–​7, 702–​3, 702n.14 Griswold, C., 592, 594, 595, 598n.9, 611n.8 Guerrero, A., 124–​25, 124n.28, 125n.29, 126, 129n.36, 330n.5, 334n.12, 345n.30, 346n.32 Gutmann, A., 613n.15, 613n.18

H Haas, D., 196n.8 Haber, J.G., 599–​600 Hahn, A., 676 Haidt, J., 310–​11 Haji, I., 27–​28, 32n.9, 33n.10, 43n.24, 47, 181–​82, 186–​87, 204n.1, 224–​25, 271–​72, 352n.3, 391–​ 92, 396n.11, 477n.17, 553n.18, 553, 555n.20, 555n.22, 555n.23

Index   731 Hall, N., 359n.5 Hamber, B., 615n.27, 617n.38 Hamilton, V.L., 254, 258–​59 Hampton, J., 592, 595, 598n.8, 602, 611n.8, 613n.20, 614n.21, 617n.43 Hankins, K., 306, 306n.2, 306n.5 Hardin, C., 671–​72 Hare, R.D., 388–​89, 394–​95, 394n.10, 403–​4, 702, 703–​4 Harman, E., xxvii–​xxviii, 334n.12, 344n.27, 346n.34, 425, 427, 593 Harman, G., 469n.2 Harris, G.T., 703n.16 Harris, S., 527n.7, 528n.11, 528–​29n.14, 529n.15, 530–​31 Hart, C., 650–​51 Hart, H.L.A., 14, 18, 102n.29, 205, 235–​36, 373, 382, 531–​32n.21 Hartman, R.J., 116n.12, 538n.28 Haslanger, S., 688 Haybron, D.M., 699 Haynes, J.D., 434, 439 Heath, J., 143n.30 Heddon, B., 346n.34 Heetderks, K., 617n.38 Held, V., 131n.1, 134n.8, 144, 144n.31 Helm, P., 635n.15 Henrich, J., 248n.4, 258n.20 Herdova, M., 451n.7, 471n.9, 477n.15, 479n.19, 479n.20, 481n.22, 482, 483n.27, 488, 488n.32, 488n.33 Hermanson, S., 682n.41, 682n.42, 682n.43 Hess, K., 141n.25 Heyman, G., 650–​51, 661 Hick, J., 639, 639n.21, 639n.22, 639n.23, 639n.24 Hieronymi, P., xix–​xx, 53–​54, 54n.3, 57, 57n.8, 58, 59–​60, 65n.17, 66–​67, 99n.21, 270–​7 1, 277, 299n.13, 299–​300n.15, 301n.19, 302n.21, 344n.29, 578n.11, 592, 593, 594, 595, 597n.7, 598, 615n.28, 673–​74 Himmelreich, J., 142n.27 Hobbes, T., 180, 181–​82 Hodgson, D., 453 Hofmann, W., 460–​61 Hogan, T., 327–​28n.1 Holder, C., 613n.17

Holmgren, M.R., 604n.19, 611n.8, 617n.44, 618n.51 Holroyd, J., 451n.7, 671–​73, 672n.7, 673n.12, 675–​77, 675n.17 Holton, R., 652–​53, 653n.6, 654 Honderich, T., 213, 532–​34, 533n.24 Honore, T., 209–​10 Hook, S., 7 Horgan, T., 31n.8, 35n.15 Horsbrugh, H.T.N., 593 Hörstkotter, D., 674 Hoskins, Z., 396 Howard-​Snyder, F., 552n.16, 641n.30 Howse, R., 616n.35, 617n.41 Huebner, B., 683, 717 Hughes, P.M., 590n.1, 615n.26 Hume, D., 180, 181–​82, 304–​6, 306n.5, 392n.8 Hunt, D., xxiii, 32n.9, 160, 161, 163, 165, 167 Hurley, S.L., 16n.12, 631n.9 Husak, D., 225–​26, 384 Hutchison, K., 283, 563n.44 Hyde, J.S., 680n.29

I Ingierd, H., 710–​11 Irwin, T., 28n.5 Isen, A., 470

J Jackson, F., 5, 332n.9 Jacobson, D., 311–​12, 314, 708 Jalava, J., 389, 395, 400, 406–​7, 702–​3, 702n.14 James, I., 681n.38 James, W., 216–​17 Jankélévitch, V., 616n.31 Januszko, P., 457 Jaworska, A., 149n.40 Jefferson, A., 3, 4, 5, 11, 12, 13, 16, 400–​1 Jeppson, S., xx–​xxi, 86, 194, 498–​99 Johnston, M., 304 Johnstone, G., 618n.48 Jost, J.T., 669n.3 Jurjako, M., 389, 400, 403–​4, 403–​4n.13, 406–​7 Jusyte, A., 402

K Kadish, S.H., 216–​17 Kagan, S., 116n.14, 423

732   Index Kahn, J., 394–​95, 403–​4 Kamtekar, R., 471n.8 Kane, R., xvii, 28n.3, 31n.8, 49, 77, 85–​86, 159, 163, 165–​66, 171–​72, 205–​6, 437–​38, 453, 454, 543–​44, 543n.2, 550 Kang, J., 672 Kant, I., 207–​8, 231, 232, 239, 382, 405, 405n.14, 519, 519n.1 Karremans, J.C., 594n.4 Kauppinen, A., 304, 388n.1, 391–​92, 494n.1 Kearns, S., 186, 471n.9, 477n.15, 479n.19, 479n.20, 481n.22, 482, 483n.27, 488n.32, 488, 488n.33 Keller, S., 343n.26 Kelley, A., 616n.31 Kelly, D., 672, 672n.7, 674, 677, 682n.40 Kelly, E., 85–​86, 380–​81, 397–​98, 399, 400–​1 Kendler, K.S., 657n.9 Kennett, J., 391–​92, 391n.6, 393, 397–​98, 399, 405n.14, 651, 656–​57, 657n.9, 664–​65 Khoury, A., 187, 598 Kiehl, K.A., 701–​9, 701n.11, 703n.15, 705n.20 King, M., xxx, 282n.13, 448, 449n.2, 450n.3, 450n.4, 452n.11, 462–​63, 658 Kiss, E., 617n.41 Kitayama, S., 258–​59 Kneer, M., 497n.6 Knobe, J., 214n.10, 352–​53, 353n.4, 494n.1, 495n.2, 497, 498, 501n.9 Koch, C., 150n.45 Koenigs, M., 704, 705 Kohn, A., 546–​47n.6 Kolber, A., 382 Kolnai, A., 592, 593, 594, 595, 615n.28 Kolodny, N., 417 Konstan, D., 605 Kornblith, H., 710–​11 Kornhauser, L.A., 145n.35 Koss, M.P., 261n.23 Kripke, S., 508n.20 Kritz, N.J., 616n.37 Krog, A., 611n.6 Kumar, R., 292n.5, 342n.25 Kumar, V., 507n.16, 511n.27 Kundera, M., 259 Kupperman, J.J., 471n.8

Kurzban, R., 458 Kymlicka, W., 149, 149n.41

L Lacey, N., 255n.16, 258 Lackey, J., 343n.26 Lane, K., 672 Lapp, W.M., 661 Larkin, P.J., 719n.30 Larsen, R., 395 Latané, B., 470 Lee, C.K., 682n.43 Lehrer, K., 31n.8 Lemos, J., 208–​9, 222–​23, 231 Lenman, J., 115n.11 Lesky, A., 255 Leslie, S.J., 461 Levin, P., 470 Levy, N., xxiv, 53n.1, 54, 57, 58, 61–​62, 61n.13, 64, 74, 76, 77, 83, 103n.32, 123n.25, 206, 208–​ 9, 224–​25, 391–​92, 399, 400, 451n.6, 454–​56, 460–​61, 527n.7, 594, 647–​48, 650–​51, 653–​77, 686 Lewis, C.S., 230–​31 Lewis, D., 31n.8, 38, 100n.25, 304, 353–​55, 508n.20 Li, P., 675n.18 Libet, B., 217, 433n.1, 433–​35, 436, 440, 450–​51, 451n.6 Lieberman, M., 470n.5 Lilienfeld, S.O., 678n.24, 679n.26, 681n.35, 688n.54 Lippert-​Rasmussen, K., 417–​19 Lippmann, W., 669n.3 List, C., xxii, 131n.1, 132n.3, 133n.5, 135n.11, 136n.13, 137n.14, 138n.16, 139n.20, 141n.26, 142n.27, 142n.29, 144n.32, 145n.35, 146n.37, 147n.38, 149n.42, 150n.44, 150n.46, 424, 496n.5, 717–​18 Litton, P., 388–​89, 704 Llewllyn, J., 616n.35, 617n.41 Lloyd-​Jones, H., 255–​56 Locke, J., 304 Loeb, D., 251–​52, 251n.10, 257–​58 Lombrozo, T., 187–​88 Longman, T., 612–​13n.14

Index   733 Loomis, E., 124n.27 Ludwig, K., 495n.2, 496n.4, 499–​500 Luna, E., 703 Lycan, W.G., 187

M MacAskill, W., 334n.12, 345n.30, 345n.31, 346n.32 Machery, E., 497n.6, 672n.7 Mackenzie, C., 283, 563n.44 Mackie, J.L., 145n.33, 217, 251–​52, 251n.10, 529n.16, 531n.20, 637, 637n.16, 637n.17, 638n.19 MacLachlan, A., 604n.19, 605 Macnamara, C., 270n.2, 277 Macrae, C.N., 671–​72 Magill, K., 567n.1 Maibom, H., 389, 391–​92, 395, 402–​3, 702–​3, 702n.13, 706, 708n.25 Maier, J., 44–​45n.27 Makridakis, S., 216 Malatesti, L., 388–​89, 400, 403–​4, 403–​4n.13, 406–​7 Mallon, R., 495n.2 Mandelbaum, E., 501n.9 Maranges, H.M., 497n.6 Maraun, M., 702–​3 Markovits, J., 337n.15, 337n.17, 338n.18, 338n.19 Markus, H., 258–​59 Martin, C.B., 38 Marx, K., 412–​15, 417, 427–​28 Mas-​Colell, A.M., 139n.17 Mason, E., 283, 549n.9 Mateeva, T., 214n.10 Matheson, B., 189–​90, 598 Matravers, M., 393 Matthews, K., 470 Mavda, A., 451n.7 May, J., 452n.11, 658 May, L., 131n.1 May, S.C., 612–​13n.14 McCallister, W., 262–​63 McCollum, J., 680n.34 McConnell, D., xxxiv, 656–​57, 664–​65 McCormick, K., 5, 10n.10, 12–​13n.11, 16n.12, 507n.16 McGary, H., 598n.9

McGeer, V., xix, 3, 5, 6n.3, 10, 11–​12, 13, 16, 17, 19, 23n.20, 275–​77, 275n.7, 277n.9, 670n.5 McGinn, C., 699 McGrath, S., 354 McGraw, P.A., 312 McHugh, C., 316–​17 McIntyre, A.G., 97n.15, 172n.2 McKenna, M., xix, 4n.1, 8n.5, 9, 11, 12–​13n.11, 20n.15, 27–​28, 29n.6, 32n.9, 33n.10, 35, 36–​37, 43n.24, 46n.34, 47, 48n.37, 48–​49, 57n.8, 58, 62n.14, 65, 77, 84, 95n.9, 115n.10, 159, 161, 163, 168–​69, 170–​72, 180n.1, 180, 181–​83, 189–​90, 194–​96, 194n.6, 196n.8, 197, 207, 212, 249n.7, 270n.2, 275–​77, 310, 313–​14, 316–​18, 356, 360, 373, 423, 424, 469n.3, 471n.9, 471–​72n.10, 477n.15, 477n.16, 477n.17, 478, 479n.20, 482, 482n.23, 482n.25, 483n.27, 484–​85n.28, 485, 490n.35, 522n.2, 531–​32n.21, 578n.11, 596, 701n.10 McKiernan, A., 562n.43 McMahan, J., 327–​28n.1, 422, 712 McMillan, J., 388–​89 McNaughton, D., 592, 618n.51 Mehl, M.R., 714 Mele, A., xxiii, 27–​28, 46n.33, 54, 161, 163, 164, 165, 167, 181–​83, 182n.2, 186, 188, 189, 190, 193, 217, 224–​25, 318, 374n.6, 433n.1, 434, 435–​36, 437n.6, 438, 439–​40, 439n.7, 442, 443, 444, 453, 471n.9, 471–​72n.10, 490n.35, 501n.8 Mendoza, S.A., 672n.8 Menges, L., 273–​74n.5 Menon, V., 458 Menzies, P., 142n.27 Merritt, M., 469n.2 Midgley, M., 405n.14 Milam, P., xxxiii, 211n.8, 545n.3, 546–​47, 547n.7, 559n.35, 573, 586n.13, 586n.14, 595n.5, 596, 600n.16, 604n.19, 605, 612n.10 Miles, J.B., 527n.7, 528n.11, 528–​29n.14 Milgram, S., 469 Millan, M., 501 Miller, C.B., 469n.3, 470n.5, 471n.6, 471n.7, 471n.8, 471n.9, 473n.11, 473–​74n.13, 477n.15, 483n.27, 490n.35, 714 Miller, D.E., 5, 6n.3, 10, 11, 13, 16, 17n.13, 17, 18–​19 Miller, D.J., 105n.35

734   Index Miller, J., 435n.4 Miller, J.G., 258–​59 Miller, W.I., 257n.19 Millikan, R.G., 508n.20, 509n.23 Milner, A., 448 Minow, M., 614n.23 Mischel, W., 714 Miyake, A., 458 Moellendorf, D., 610n.4 Mokros, A., 708n.25 Moleskis, A., 216 Moller, D., 334n.12, 346n.32 Montague, P.R., 443 Monteith, M.J., 671–​72, 675–​76, 675n.19, 676n.20 Montgomery, C., 391n.6 Montmarquet, J., 471n.8 Moody-​Adams, M., 328n.2, 330n.5, 677n.23 Moore, G.E., 7 Moore, M.S., 92n.2, 95n.12, 108n.41, 108n.43, 372n.3, 373, 382, 383, 719–​20n.31 Moore, R.E., 93n.4, 96n.14 Morris, H., 205, 223–​24, 230–​31, 378–​79, 609, 609n.1, 720 Morris, S.G., 9n.6, 20, 21n.17 Morse, S., 236, 373, 374n.5, 391–​92, 393, 663, 704, 706, 707–​8 Moskowitz, G.B., 671–​72, 675n.18 Moss-​Racusin, C.A., 682 Moya, C., 158, 162, 168–​69 Murphy, C., 229, 617n.39 Murphy, D., xxxiv, 111–​12n.3, 471, 471n.9, 477n.15, 487n.30, 703n.16, 709–​10, 711, 712, 712n.28, 719, 720 Murphy, J., 590, 592, 593, 594, 595, 597n.7, 598, 611n.8, 613n.20, 614n.21, 617n.43, 720 Murray, D., 187–​88, 498, 503 Murray, M., 641, 641n.30 Myers, C.K., 678–​79

N Nadelhoffer, T., 214n.10, 502, 502n.10, 503n.11, 528n.11, 530–​31, 536n.25, 690–​91 Nado, J., 508n.19 Nagel, T., 149, 149n.42, 206, 252, 310, 340n.20, 423–​24, 527n.7, 539 Nahmias, E., 190, 191, 192, 193n.5, 193, 471n.7, 471n.9, 495, 497, 498, 503

Naylor, M., 161 Neale, J., 650–​51 Nelkin, D., xix, 20n.16, 21, 31n.8, 33n.10, 35n.15, 37–​38, 39n.20, 42, 46, 48, 62n.14, 82–​83, 84, 87, 103n.30, 103n.33, 105n.36, 106n.37, 112n.6, 121–​22, 121n.21, 128–​29, 129n.35, 129n.36, 158, 171, 181–​82, 196–​97, 270n.2, 273, 273–​74n.5, 275, 277, 300n.16, 306n.4, 342n.25, 351, 366, 373, 374n.5, 375, 391n.6, 394, 469n.3, 471n.9, 473n.11, 473–​74n.13, 477n.15, 477–​78, 477n.17, 479, 480, 481n.21, 485n.29, 495n.3, 498, 503, 543n.1, 551n.14, 552n.16, 554, 555n.21, 556n.26, 557n.27, 594–​95, 599, 602, 603–​4, 659n.11, 700, 701n.10, 706–​7, 718n.29 Nelson, P., 632n.11, 632–​35 Newberry, P.A., 592n.2, 597–​98 Newman, J.P., 704 Nichols, S., 5, 13, 16n.12, 214n.10, 495n.2, 497, 498, 501n.9, 502, 503–​5, 507n.16, 509n.22, 511n.27, 532n.22, 705n.20, 707, 708 Nietzsche, F., 216–​17 Nisbett, R., 469n.2, 471n.7 Norcross, A., 556n.25 Nosek, B.A., 310–​11 Novitz, D., 593 Nowell-​Smith, P., 3, 7–​8, 8n.4, 523n.3 Nozick, R., 5, 377n.10 Nussbaum, M., 228–​29, 228n.2, 255–​56, 260, 296n.10 Nutt, D.J., 654–​55

O Oak, S., 712 O’Connor, T., 437–​38 Oldenquist, A., 230–​31 Olin, L., 312, 711 Olson, M., 140n.23 Oshana, M., 283, 563n.44 Osiel, M., 710 Ostrom, E., 140n.23 Oswald, F.L., 460n.17 Otsuka, M., 158, 161

P Palmer, D., 160, 161 Paolacci, G., 495n.2 Papish, L., 714

Index   735 Parfit, D., 85–​86 Pasternak, A., 131n.1, 141n.25, 148n.39 Paul, L.A., 359n.5 Paulhus, D.L., 234, 235n.5 Payne, B.K., 681–​82 Pearl, J., 190n.4 Pears, D.F., 7 Pennington, C.R., 680n.28 Pensky, M., 615n.26 Pereboom, D., xx–​xxi, xxii, 4, 8n.5, 9n.6, 9n.7, 12–​13n.11, 13, 16n.12, 20–​21, 22n.18, 23n.19, 23, 27–​28, 32n.9, 39n.20, 41–​42, 45n.31, 67, 75, 84–​86, 111n.1, 115, 115n.10, 158, 159–​60, 161–​62, 162n.1, 163, 164, 165, 166, 167, 168–​70, 174, 179, 180n.1, 180, 181–​82, 182n.2, 183–​84, 185–​86, 187, 188–​89, 190, 193, 195–​97, 196n.8, 204, 208–​9, 211n.7, 211n.8, 212, 222–​23, 224–​ 25, 226–​27, 238, 241–​43, 247n.1, 249, 254n.15, 274n.6, 275n.7, 292n.5, 294n.8, 306n.4, 311, 313–​14, 316, 317n.14, 318, 370–​7 1, 396n.11, 398, 423, 424, 452n.9, 471–​72n.10, 474, 475n.14, 495n.3, 498–​99, 503–​4, 523n.4, 527–​ 31, 528n.12, 528n.13, 532n.22, 532n.23, 535–​36, 545, 545n.3, 545n.4, 546, 550, 550n.11, 553n.18, 553, 555n.23, 555–​56, 556n.25, 561n.40, 573, 574, 579, 586n.13, 588, 594, 640n.26, 700n.8 Pernu, T.K., 555n.21 Perrett, R.W., 699 Persson, K., 498, 507n.16, 509n.23, 511n.27 Pettigrove, G., 592, 593, 595, 596, 599–​600, 600n.15, 604n.19, 605, 611n.8, 612n.10, 615n.28 Pettit, P., 11, 12, 131n.1, 132n.3, 133, 133n.5, 133n.6, 133n.7, 135, 135n.11, 135n.12, 137n.14, 138n.16, 139n.20, 141n.26, 142n.27, 142n.29, 144n.32, 145n.35, 146n.37, 147n.38, 673–​74, 676, 686n.50, 717–​18 Pham, P., 612–​13n.14 Philpott, D., 615n.25 Picinali, F., 396–​97 Pickard, H., 273–​74n.5, 311, 647–​48, 648n.3 Pillsbury, S.H., 706 Piovarchy, A., 559n.36 Plakias, A., 251–​52, 251n.10 Plant, E., 671–​72, 673, 675–​76 Plantinga, A., 637–​39, 637n.18 Plazzi, G., 457

Pols, H., 712 Portmore, D.W., 133n.4, 316–​17 Poston, T., 641n.31 Powers, M., 228–​30 Prager, C.A.L., 616n.32 Prinz, J.J., 708 Pronk, T.M., 594n.4 Putnam, H., 508n.20

Q Quine, W.V., 140n.21 Quinton, A., 137–​38, 137n.15

R Radcliffe, E.S., 392n.8 Radzik, L., xxxiii, 282, 559–​60, 562n.42, 562, 614n.22, 617n.42, 617n.45, 619n.55 Raine, A., 704, 705, 705n.21, 706, 708 Ramadan, H.M., 390n.4 Raven, M., 359n.6 Ravizza, M., xvii, 32n.9, 33n.10, 35–​36, 40–​41, 40n.21, 42–​43, 47, 49, 77, 96n.13, 97n.15, 126n.32, 127, 127n.33, 172, 172n.2, 173, 181–​82, 248n.3, 249n.5, 270n.2, 306, 315, 318, 348n.1, 360, 373, 393, 437–​38, 453, 477–​78, 477n.17, 479n.20, 482n.25, 488, 701n.9, 701n.10 Rawls, J., xix, 5, 14, 18, 383n.13, 414, 415, 424, 425, 427–​28, 552n.15, 699n.7, 720 Raz, J., 102n.29 Rees, C.F., 675–​76 Reeve, J., 470 Reiman, J., 396–​97 Rice, M.E., 703n.16 Richards, N., 595, 598n.9 Richerson, P., 258n.20 Richman, K.A., 82 Rickless, S.C., 103n.30, 103n.33, 105n.36, 106n.37, 351 Ricks, T., 713, 715 Ripley, D., 501n.9 Robb, D., xxiii, 163, 164, 165, 167 Robichaud, P., 95n.10, 450n.3 Robins, L., 655 Robinson, P., 240–​41, 379n.11, 720n.33 Robinson, T.E., 652 Robison, J., 561n.41 Rodgers, T., 471n.8

736   Index Rodin, J., 470 Roedder, E., 672, 672n.7, 674, 682n.40 Rose, D., 502, 503–​5 Rosen, G., xxiv, 38, 123–​24, 125, 126, 270n.2, 271, 273–​74n.5, 327–​28n.1, 330n.5, 425, 594 Rosenthal, D., 441–​42 Roskies, A.L., 497, 498 Ross, Ja., 334n.12, 346n.33 Ross, Jo., 651 Ross, L., 469n.2, 471n.7 Ross, W.D., 382 Rotberg, R.I., 613n.15, 615n.25 Rouse, C., 682 Rovane, C., 137n.14 Rowe, M., 678–​79 Rowe, W., 628n.5, 638–​39n.20 Royce, J., 610n.2 Rudy-​Hiller, F., 555n.23 Russell, D., 471n.7, 471n.8, 714 Russell, P., xxxii, 212, 213, 249n.7, 304–​5, 310–​11, 522n.2, 526n.6, 536n.26, 537–​38, 569, 570, 570n.5, 598n.11

S Sabini, J., 471n.8 Sager, L.G., 145n.35 Saloner, B., 656 Sanders, J., 254, 258–​59 Sarch, A., 462–​63 Sarkissian, H., 471n.8, 497, 498 Sartorio, C., xix, 32n.9, 33n.10, 33n.11, 35, 37, 37n.17, 47, 47n.35, 95n.12, 97n.15, 98n.18, 108n.42, 158, 159–​60, 172, 172n.2, 173, 174, 181–​82, 349, 352n.2, 354–​55, 356–​57, 358, 359–​ 60, 360n.7, 361, 471n.9, 472, 477n.15, 477n.17, 480, 482n.25, 488, 489 Sartre, J., 49 Saul, J., 671–​72, 672n.7, 677, 679–​80, 682 Savulescu, J., 650–​51 Scanlon, T.M., xix–​, 10, 21n.17, 53–​54, 54n.5, 57, 59, 59n.10, 60–​61, 62, 64n.16, 64–​66, 67, 67n.18, 78, 78n.2, 99n.21, 100n.22, 179, 211n.7, 269, 270n.2, 272, 274–​75, 277, 280, 283, 292n.5, 294n.7, 294n.8, 299–​300n.15, 316–​17, 377n.9, 415, 557, 557n.28, 559–​60, 613n.20, 648 Scarantino, A., 320

Scarre, G., 603n.18 Schaffer, J., 92n.2, 96n.14, 164, 354–​55 Scheffler, S., 417, 424 Schellenberg, J.L., 640, 641 Schlick, M., 3, 7–​8, 32–​33, 271, 523n.3 Schlosser, M., 451n.7, 673–​74 Schönherr, J., 597–​98, 605 Schroeder, T., 120, 120n.19, 120n.20, 121–​22, 122n.24, 648, 652n.5, 657n.9 Schroeter, F., 508–​9, 508n.21 Schroeter, L., 508–​9, 508n.21 Schütz, A., 216 Schwarz, W., 435n.4 Schwitzgebel, E., 150n.44, 150n.46 Sebastian, C., 391n.6, 391n.7, 401–​2 Segall, S., 418–​19 Sekatskaya, M., 195n.7, 196n.8 Sen, A., 228–​29 Sepielli, A., 334n.12, 346n.33 Setiya, K., 27n.2 Shabo, S., xxxii–​xxxiii, 160, 161, 189, 211n.8, 532n.23, 545, 546, 547, 567n.2, 569n.4, 571n.6, 573, 583, 586n.14 Shah, P., 458 Shamay-​Tsoory, S.G., 401–​2 Sharma, R.N., 252, 253–​54 Sharp, C., 401–​2 Sharrif, A.F., 234, 235, 237 Shaw, E., xxix, 393, 398, 406–​7 Shay, J., 721 Shelby, T., 281 Shepherd, J., 98n.18, 217, 449n.1, 452n.8, 471n.9, 490n.35, 497n.6, 503n.11, 503n.12 Sher, G., xx, 53n.2, 54, 102n.29, 274, 275, 450n.3, 450n.4, 450n.5, 558n.31, 700 Sherman, N., 721 Shih, M., 680n.27 Shoda, Y., 714 Shoemaker, D., xx–​xxi, xxvi–​xxvii, 4n.1, 20n.15, 27–​28, 58, 64, 65, 77, 78–​80, 81, 82–​83, 84, 85–​86, 87, 250n.8, 270n.2, 273–​ 74n.5, 289n.3, 292n.5, 296n.9, 311n.9, 313–​14, 314n.10, 316, 318n.15, 320, 342n.25, 351, 371–​ 72, 399, 400–​1, 452n.9, 498–​99, 548, 548n.8, 549, 567n.1, 706–​7 Shulman, G., 458 Shweder, R.A., 258–​59

Index   737 Sidgwick, H., 212n.9 Sie, M., xxxiv–​xxxv, 673–​74, 680n.31, 681n.36, 683–​84, 687n.52 Silver, K., 148n.39 Silver, M., 471n.8 Sim, K., 710 Sinnott-​Armstrong, W., 400, 647–​48, 705n.20, 705, 708, 709 Skinner, B.F., 523n.3, 527, 527n.8 Slingerland, E., 471n.8 Sliwa, P., 332n.8, 338n.18 Smart, J.J.C., 3, 8–​10, 8n.5, 9n.7, 10n.10, 13, 16n.12, 23, 75, 76, 78, 84, 86, 212n.9, 271 Smilansky, S., xxiv–​xxv, 203–​4, 205, 206, 207, 208–​10, 208n.3, 208n.4, 209n.6, 212, 213–​14, 214n.10, 215, 215–​16n.11, 216–​18, 222–​23, 224, 532, 534–​36 Smiley, M., 131n.1, 677n.23 Smith, Ad., 116n.14, 305–​6 Smith, An., xix–​xx, xxv–​xxvi, 4n.1, 53–​54, 54n.4, 57n.8, 57–​58, 59–​60, 60n.11, 60n.12, 61, 65n.17, 66–​68, 74, 77, 78–​81, 99n.21, 100n.23, 270n.2, 271n.3, 273–​74n.5, 275n.7, 277, 283, 292n.5, 299–​300n.15, 316–​17, 449, 450n.3, 452n.9, 498–​99, 560n.38, 562, 578n.11, 621, 621n.59, 621n.60, 673–​74, 675–​ 76, 686 Smith, D.L., 713 Smith, H., 99n.21, 101n.27, 119–​20, 119–​20n.17, 328n.2, 450n.3 Smith, M., 35n.15, 37–​38, 38n.18, 43n.24, 251–​ 52, 257, 304, 306n.2, 306n.5, 306, 327–​28n.1 Smith, P., 93n.4 Snidal, D., 139n.18 Snoek, A., 656–​57, 664–​65 Snow, N.E., 714 Sommers, T., xxiv, 204, 211n.8, 213, 243, 247–​ 48, 248n.4, 253, 253n.13, 256–​57, 261n.23, 314, 495n.3, 507, 523n.4, 545n.3, 573 Soon, C.S., 433n.1, 433, 434 Speak, D., xxxiii, 640n.26, 643n.33 Spiekermann, K., 496n.4, 510n.24 Spinoza, B., 197, 474 Springer, E., 677n.23 Sreenivasan, G., 471n.8 Sripada, C., 453, 501n.8, 648, 654n.8 Stanford, M., 651

Steele, C.M., 679–​80 Steiner, H., 699 Steinpreis, R.E., 682 Stenning, A., 82n.4 Stern, L., 311 Stich, S.P., 251n.10 Stone, M.H., 699, 709 Stover, E., 612–​13n.14 Strawson, G., xxiv, 4, 9n.6, 23, 49, 111n.1, 204, 213, 216–​17, 218n.12, 224, 307, 528n.12, 573, 631–​32, 632n.10 Strawson, P.F., xvii, 4, 5, 7–​8, 8n.4, 9n.9, 10–​11, 13, 15, 16, 17, 18, 19, 77, 81, 86, 111–​12n.3, 113n.7, 120n.18, 121–​22n.22, 212–​13, 217, 250–​51, 269–​70, 272–​73, 287–​91, 287n.1, 293, 295–​96, 297–​98, 300, 303, 307–​10, 309n.8, 320–​ 21n.20, 369, 371–​73, 452n.10, 507, 520–​24, 527–​28, 537–​38, 539n.30, 542, 544–​45, 563, 566–​68, 569, 570–​7 1, 590, 592, 620n.57, 670 Stroud, S., 343n.26 Struthers, C.W., 594n.4 Stump, E., 32n.9, 160, 163, 640n.27 Sturgeon, N., 256n.18, 257 Sue, D., 678–​79 Sverdlik, S., 102n.29 Sweetman, J., 672n.7 Swenson, P., 47–​48n.36, 112n.6, 127, 127n.33, 128, 173, 479n.19 Swinburne, R., 640n.27, 699n.4 Syse, H., 710–​11 Sysma, J., 494n.1

T Tabandeh, A., 229 Tadros, V., 222–​23, 382 Talbert, M., xix–​xx, 8n.5, 41–​42, 53–​54, 53n.2, 55n.7, 63, 64n.16, 64–​65, 66–​67, 83, 99n.21, 100n.23, 101n.27, 104n.34, 277, 342n.25, 471n.9, 482n.26, 487n.30, 578n.11, 706, 712n.28, 714–​15 Talbot, B., 496n.4, 496n.5, 500n.7, 510n.24, 511n.26 Tamber-​Rosenau, B., 458 Tannenbaum, J., 149n.40 Taylor, C., 680n.33 Taylor, M.C., 497n.6 Taylor, R., 180–​82

738   Index Taylor, S.E., 216 Teitel, R., 614n.23 Temkin, L., 418–​19 Terzaghi, M., 457 Thompson, D., 613n.15, 613n.18, 615n.25 Thompson, J., 618n.49 Thompson, P., 699 Tierney, H., 112n.6, 122n.23, 123n.26, 187, 193n.5, 718n.29 Tillem, S., 389, 403 Tilly, C., 713 Timmons, M., 327–​28n.1 Timpe, K., 27–​28, 640n.26, 643n.33 Todd, P., 48, 186, 187, 224–​25, 281, 306, 315, 560n.39 Tognazzini, N., 48, 80n.3, 85, 270n.1, 272n.4, 273–​74n.5, 275n.7, 279, 289n.3, 291n.4, 292n.5, 296n.9, 299n.13, 299–​300n.15, 560n.38, 597n.7 Tollefsen, D.P., 131n.1, 137n.14, 139n.19, 139n.20, 717 Tononi, G., 150n.45 Tremain, S., 697n.2 Triandis, H., 258–​59 Tucker, C., 471n.8 Tuggy, Dale, 634n.13 Tuomela, R., 137n.14, 140n.22 Tutu, D., 612–​13n.14 Twambley, P., 599

U Uddin, L., 458 Uhlmann, E.L., 676n.21 Unkelbach, C., 681n.37, 681n.39 Unsworth, N., 458 Usher, M., 190

V Vaihinger, H., 216–​17 Valian, V., 669n.3, 679–​80 Van Den Haag, E., 396–​97 Van Hees, M., 131n.2, 145n.34 Van Inwagen, P., 28n.3, 28n.4, 31n.8, 34n.12, 45, 97n.15, 161, 248, 249, 437–​38, 498, 507, 634–​ 35, 634n.14, 640n.27 Van Ness, D.W., 617n.38 Van Schoelandt, C., 48–​49, 701n.10 Van Voorst Vader-​Bours, N.M., 676–​77, 680n.31, 681n.36, 683–​84

Vanwoerden, S., 401–​2 Vargas, A.F., 721 Vargas, M., xix, 3, 4n.1, 5, 6, 9n.8, 10n.10, 10, 11–​12, 12–​13n.11, 13, 14, 16n.12, 16, 17, 19, 20, 33n.10, 43n.24, 115n.11, 128n.34, 212, 250n.9, 270n.2, 275–​76, 277n.9, 318, 318n.15, 320–​ 21n.20, 373, 471n.9, 477n.17, 488, 489, 495n.3, 507n.16, 509n.22, 509n.23, 552n.15, 555n.21, 597n.7, 629n.6, 643, 687n.51, 700, 701n.10, 705n.20, 706, 707 Varzi, A.C., 92n.2 Verdeja, E., 618n.47 Vermazen, B., 92 Verwoerd, W., 619n.53, 619n.54 Vetter, B., 44–​45 Viding, E., 402–​3, 705 Vihvelin, K., 28n.3, 31n.8, 33n.10, 35n.15, 37–​38, 43n.24, 43, 43n.25, 44–​45n.27, 171 Vilhauer, B., 208n.4, 398 Voils, C.I., 676n.20 Vonasch, A., 444, 445n.11, 501n.8 Von Hirsch, A., 366 Vranas, P., 471n.8, 552n.16

W Walen, A., 230–​31, 236, 237, 240–​41 Walgrave, L., 242–​43 Walker, M.U., 275, 595, 604n.19, 605, 612n.10, 613n.20, 615, 615n.29, 615n.30, 617–​18, 617n.42, 617n.45, 617n.46 Wallace, R.J., xx, 21n.17, 32n.9, 45n.29, 58, 62n.14, 67, 181–​82, 206–​7, 211n.8, 249, 250n.8, 269, 270n.2, 270–​7 1, 272–​73, 273–​74n.5, 275n.8, 280, 289n.2, 289n.3, 292n.5, 293n.6, 297n.11, 300n.16, 300n.17, 301n.18, 302n.20, 303n.22, 310, 311, 314, 342n.25, 371–​72, 373, 523n.4, 527n.7, 527n.9, 539n.30, 539n.31, 551n.13, 560n.39, 568n.3, 569, 647–​48, 701 Waller, B., 204, 208–​9, 210, 224–​25, 232, 233–​34, 528n.11, 528–​29n.14, 535–​36 Walters, J., 599 Waltz, K., 139n.18 Ward, L., 311 Warmke, B., xxx–​xxxi, xxxiii, 469n.3, 471n.8, 471n.9, 477n.15, 477n.16, 482, 482n.23, 483n.27, 484–​85n.28, 485, 490n.35, 590n.1, 594, 595, 598n.11, 599–​600, 600n.14, 600n.17, 601, 603–​4, 605 Washington, N., 677

Index   739 Watson, G., xvii, 9n.7, 12, 18n.14, 27–​28, 46n.34, 48–​49, 54, 54n.5, 55, 56–​57, 57n.8, 58, 65, 75–​77, 84, 85, 249, 249n.6, 269, 270, 271–​72, 273, 273–​74n.5, 275, 277, 287n.1, 296n.9, 297n.12, 299n.14, 300n.17, 307, 308, 309, 310, 311, 317n.14, 342n.25, 371–​72, 377n.9, 498–​99, 523–​24, 525n.5, 551n.13, 558, 558n.32, 572, 578n.11, 620n.57, 647–​48, 706–​7, 711 Weatherburn, D., 240 Weatherson, B., 346n.34 Webber, J., 471n.8 Wegner, D.M., 217, 442, 450–​51, 451n.6 Weinberg, J., 495n.2 Weinstein, H.M., 612–​13n.14 Weschler, L., 616n.37 Whinston, M.D., 139n.17 Whittle, A., 171–​72 Widerker, D., 95n.9, 97n.15, 157, 159, 163, 169–​70, 172n.2, 356 Wieland, J.W., 95n.10, 450n.3 Wielenberg, 471n.8 Williams, B., xxxii, 18–​19, 206, 213, 251–​52, 310–​11, 519, 538–​39, 539n.31 Williams, L.E., 451n.7 Williams, W.M., 682n.43 Wilson, M., 263

Wolf, S., xvii, 31n.8, 33n.10, 46, 55–​56, 62n.14, 63, 67, 76, 86, 87, 158, 171–​91, 196, 213, 249, 269, 270–​7 1, 273–​74n.5, 275n.8, 292n.5, 293n.6, 327–​28n.1, 373, 477–​78, 477n.17, 543–​ 44, 549n.9, 555n.21, 557n.29, 711 Wolfendale, J., 471n.9, 487n.30, 712n.28, 714–​15 Wonderly, M., 708n.24 Wood, A., 475, 475n.14 Woodward, J., 190n.4 Woolfolk, R.L., 697n.2 Woolford, A., 613n.16 Wouters, A., 673–​74 Wright, R.W., 145n.33 Wyma, K.D., 161

Y Yaffe, G., 653n.7, 659 Yang, Y., 705

Z Zagzebski, L., 160, 633n.12 Zaibert, L., 593, 598, 599–​600, 602–​4 Zalaquett, J., 616n.37 Zehr, H., 611n.7, 618n.48 Zheng, R., 672–​73, 677, 677n.23, 686, 687–​88 Zimmerman, M., xx, 76n.1, 87, 271–​72, 327–​ 28n.1, 351, 352n.3, 476–​77