The Routledge Handbook of Philosophy of Responsibility [1 ed.] 9781032252391, 9781003282242

The philosophical inquiry of responsibility is a major and fast-growing field. It not only features questions around fre

124 29 60MB

English Pages 548 [549] Year 2023

Report DMCA / Copyright

DOWNLOAD PDF FILE

Table of contents :
Cover
Half Title
Series
Title
Copyright
Contents
Notes on contributors
Preface
Introduction
Part 1 The history of responsibility
Section 1: Responsibility and wrongdoing
1 Plato on vice
2 Hegel on guilt
Section 2: Responsibility and determinism
3 The Stoics: what kind of responsibility is compatible with divine providence?
4 Hobbes against Bramhall: moral responsibility, free will, and mechanistic determination
5 Hume on free will and moral responsibility
6 Sidgwick on free will and ethics
Section 3: The scope of responsibility
7 Aristotle on legal and moral responsibility: interpretation and reform
8 Kant on absolute responsibility and transcendental freedom
Section 4: Individuals and society
9 Responsibility in Confucian thought
10 Aquinas on holding others to blame
Part 2 The theory of responsibility
Section 5: The concepts of responsibility
11 Responsibility and agency
12 Responsibility and causation
13 Responsibility and the deep self
14 Responsibility and the emotions
15 Varieties of answerability
Section 6: The conditions and challenges of responsibility
16 The consequences of incompatibilism
17 Free will and the case for compatibilism
18 Deliberation and the possibility of skepticism
19 Responsibility and manipulation
20 Responsibility and coercion
21 The epistemic condition of moral responsibility
22 Moral competence and mental disorder
23 Excuse, capacity, and convention
Part 3 The practice of responsibility
Section 7: Being and holding responsible
24 Blaming
25 Communicating praise
26 The standing to blame
27 Apology and forgiveness: transaction, conversation, or joint narrative?
28 Taking responsibility
29 Responsibility without blame
30 Holding responsible in the African tradition: reconciliation applied to punishment, compensation, and trials
Section 8: The ethics and politics of responsibility
31 Artificial intelligence and the imperative of responsibility: reconceiving AI governance as social care
32 Moral responsibility for historical injustice
33 Corporate digital responsibility
34 Reckless complicity: international banks and future climate
35 Responsibility and gender
Section 9: Responsibility in the law
36 Legal and moral responsibility
37 The voluntary act requirement
38 Strict liability and strict responsibility
39 Responsibility and pretrial detention
40 Responsibility for others
41 Legitimate divergence between moral and criminal blame
Index
Recommend Papers

The Routledge Handbook of Philosophy of Responsibility [1 ed.]
 9781032252391, 9781003282242

  • 0 0 0
  • Like this paper and download? You can publish your own PDF file online for free in a few minutes! Sign Up
File loading please wait...
Citation preview

THE ROUTLEDGE HANDBOOK OF PHILOSOPHY OF RESPONSIBILITY

The philosophical inquiry of responsibility is a major and fast-growing field. It not only features questions around free will and moral agency but also addresses various challenges in the social, institutional, and legal contexts in which people are being held responsible. The Routledge Handbook of Philosophy of Responsibility is an outstanding survey and exploration of these issues. Comprised of forty-one chapters by an international team of contributors, the Handbook is divided into three clear parts – on the history, the theory, and the practice of responsibility – within which the following key topics are examined: • • • • • • • • •

responsibility and wrongdoing responsibility and determinism the scope of responsibility the responsibility of individuals within society the concepts of responsibility the conditions and challenges of responsibility the practices of being and holding responsible the ethics and politics of responsibility responsibility in the law.

Including suggestions for further reading at the end of each chapter, The Routledge Handbook of Philosophy of Responsibility provides an extremely useful guide to the topic. It will be valuable reading for students and researchers in philosophy and applied ethics, as well as for those in related fields such as politics, law, and policymaking. Maximilian Kiener is a Junior Professor of Philosophy and Ethics in Technology at Hamburg University of Technology, Germany, and an Associate Member of the Faculty of Philosophy at the University of Oxford, UK. He specialises in moral and legal philosophy, with a particular focus on consent, responsibility, and artificial intelligence. His book Voluntary Consent: Theory and Practice is also published by Routledge.

ROUTLEDGE HANDBOOKS IN PHILOSOPHY

Routledge Handbooks in Philosophy are state-of-the-art surveys of emerging, newly refreshed, and important fields in philosophy, providing accessible yet thorough assessments of key problems, themes, thinkers, and recent developments in research. All chapters for each volume are specially commissioned, and written by leading scholars in the field. Carefully edited and organized, Routledge Handbooks in Philosophy provide indispensable reference tools for students and researchers seeking a comprehensive overview of new and exciting topics in philosophy. They are also valuable teaching resources as accompaniments to textbooks, anthologies, and research-orientated publications. Also available: THE ROUTLEDGE HANDBOOK OF AUTONOMY Edited by Ben Colburn THE ROUTLEDGE HANDBOOK OF THE PHILOSOPHY AND PSYCHOLOGY OF FORGIVENESS Edited by Glen Pettigrove and Robert Enright THE ROUTLEDGE HANDBOOK OF PHILOSOPHY OF IMPLICIT COGNITION Edited by J. Robert Thompson THE ROUTLEDGE HANDBOOK OF WOMEN AND EARLY MODERN PHILOSOPHY Edited by Karen Detlefsen and Lisa Shapiro THE ROUTLEDGE HANDBOOK OF PHILOSOPHY OF RESPONSIBILITY Edited by Maximilian Kiener For more information about this series, please visit: www.routledge.com/Routledge-Hand books-in-Philosophy/book-series/RHP

THE ROUTLEDGE HANDBOOK OF PHILOSOPHY OF RESPONSIBILITY

Edited by Maximilian Kiener

Cover image: © Getty Images First published 2024 by Routledge 4 Park Square, Milton Park, Abingdon, Oxon OX14 4RN and by Routledge 605 Third Avenue, New York, NY 10158 Routledge is an imprint of the Taylor & Francis Group, an informa business © 2024 selection and editorial matter Maximilian Kiener; individual chapters, the contributors The right of Maximilian Kiener to be identified as the author of the editorial material, and of the authors for their individual chapters, has been asserted in accordance with sections 77 and 78 of the Copyright, Designs and Patents Act 1988. All rights reserved. No part of this book may be reprinted or reproduced or utilised in any form or by any electronic, mechanical, or other means, now known or hereafter invented, including photocopying and recording, or in any information storage or retrieval system, without permission in writing from the publishers. Trademark notice: Product or corporate names may be trademarks or registered trademarks, and are used only for identification and explanation without intent to infringe. British Library Cataloguing-in-Publication Data A catalogue record for this book is available from the British Library Library of Congress Cataloging-in-Publication Data Names: Kiener, Maximilian, editor. Title: The Routledge handbook of philosophy of responsibility / edited by Maximilian Kiener. Description: Abingdon, Oxon : Routledge, 2024. | Includes bibliographical references and index. Identifiers: LCCN 2023023356 (print) | LCCN 2023023357 (ebook) | ISBN 9781032252391 (hardback) | ISBN 9781032252438 (paperback) | ISBN 9781003282242 (ebook) Subjects: LCSH: Responsibility—Philosophy. Classification: LCC BJ1451 .R68 2024 (print) | LCC BJ1451 (ebook) | DDC 170—dc23/eng/20230828 LC record available at https://lccn.loc.gov/2023023356 LC ebook record available at https://lccn.loc.gov/2023023357 ISBN: 978-1-032-25239-1 (hbk) ISBN: 978-1-032-25243-8 (pbk) ISBN: 978-1-003-28224-2 (ebk) DOI: 10.4324/9781003282242 Typeset in Sabon by Apex CoVantage, LLC

CONTENTS

Notes on contributors x Prefacexvii Introduction Maximilian Kiener

1

PART 1

The history of responsibility

15

Section 1: Responsibility and wrongdoing

17

  1 Plato on vice Marcel van Ackeren

19

  2 Hegel on guilt Mark Alznauer

31

Section 2: Responsibility and determinism

43

  3 The Stoics: what kind of responsibility is compatible with divine providence?45 Rachana Kamtekar   4 Hobbes against Bramhall: moral responsibility, free will, and mechanistic determination Thomas Pink

v

57

Contents

  5 Hume on free will and moral responsibility Peter Millican

68

  6 Sidgwick on free will and ethics Anthony Skelton

82

Section 3: The scope of responsibility

95

  7 Aristotle on legal and moral responsibility: interpretation and reform Terence Irwin

97

  8 Kant on absolute responsibility and transcendental freedom David Sussman

110

Section 4: Individuals and society

123

  9 Responsibility in Confucian thought David B. Wong

125

10 Aquinas on holding others to blame Jeffrey Hause

137

PART 2

The theory of responsibility

149

Section 5: The concepts of responsibility

151

11 Responsibility and agency Maria Alvarez

153

12 Responsibility and causation Alex Kaiserman

164

13 Responsibility and the deep self Monika Betzler

177

14 Responsibility and the emotions Andreas Brekke Carlsson

190

15 Varieties of answerability Maximilian Kiener

204

vi

Contents

Section 6: The conditions and challenges of responsibility

217

16 The consequences of incompatibilism Patrick Todd

219

17 Free will and the case for compatibilism Carolina Sartorio

230

18 Deliberation and the possibility of skepticism Simon-Pierre Chevarie-Cossette

239

19 Responsibility and manipulation Massimo Renzo

250

20 Responsibility and coercion Carla Bagnoli

261

21 The epistemic condition of moral responsibility Daniel J. Miller

274

22 Moral competence and mental disorder Lubomira Radoilska

287

23 Excuse, capacity, and convention David Owens

299

PART 3

The practice of responsibility

311

Section 7: Being and holding responsible

313

24 Blaming Leonhard Menges

315

25 Communicating praise Daniel Telech

326

26 The standing to blame Matt King

337

27 Apology and forgiveness: transaction, conversation, or joint narrative? Andrea Westlund

346

vii

Contents

28 Taking responsibility Elinor Mason

357

29 Responsibility without blame Bruce Waller

368

30 Holding responsible in the African tradition: reconciliation applied to punishment, compensation, and trials Thaddeus Metz

380

Section 8: The ethics and politics of responsibility

393

31 Artificial intelligence and the imperative of responsibility: reconceiving AI governance as social care Shannon Vallor and Bhargavi Ganesh

395

32 Moral responsibility for historical injustice Michael Schefczyk

407

33 Corporate digital responsibility Alexander Filipović

419

34 Reckless complicity: international banks and future climate Henry Shue

431

35 Responsibility and gender Paula Casal

442

Section 9: Responsibility in the law

455

36 Legal and moral responsibility Peter Cane

457

37 The voluntary act requirement John Hyman

468

38 Strict liability and strict responsibility R A Duff

480

39 Responsibility and pretrial detention Kimberly Kessler Ferzan

490

viii

Contents

40 Responsibility for others Jenny Steele

501

41 Legitimate divergence between moral and criminal blame Alexander Sarch

514

Index526

ix

NOTES ON CONTRIBUTORS

Maria Alvarez is a Professor of Philosophy at King’s College London. She works and has published on the philosophy of action, reasons, normativity, and moral responsibility. Her book Kinds of Reasons. An Essay in the Philosophy of Action was published by Oxford University Press in 2010 and she’s currently working on a monograph on agency, choice, and moral responsibility. Mark Alznauer is an Associate Professor in the Department of Philosophy at Northwestern University. He has a PhD from the John U. Nef Committee on Social Thought at the University of Chicago (2008). He is the author of Hegel’s Theory of Responsibility (Cambridge University Press, 2015) and the editor of Hegel on Tragedy and Comedy: New Essays (SUNY Press, 2021). He is a former vice president of the Hegel Society of America. Carla Bagnoli has been a Professor of Philosophy at the University of Modena and Reggio Emilia since 2010. Previously, she was a tenured Full Professor at the University of Wisconsin, where she had taught since 1998. Her work focuses on practical reason, moral epistemology, and the philosophy of agency. She authored five monographs, including one on moral responsibility, and edited Morality and the Emotions (Oxford University Press 2011), Constructivism in Ethics (Cambridge University Press 2013), and Time in Action (Routledge 2022). She is currently working on Kantian constitutivism and co-editing a volume under contract with Oxford University Press. Monika Betzler holds the Chair for Practical Philosophy and Ethics at Ludwig-MaximiliansUniversity (LMU) in Munich, Germany. She works and has published on themes in moral psychology, normative ethics, and normativity. She completed a book manuscript on “The Normative Significance of Personal Projects” and is currently working on a monograph on the ethics of personal relationships. She is also co-editing (with Jörg Löschke) a volume on “The Ethics of Personal Relationships” (under contract with Oxford University Press). Peter Cane spent the first half of his career teaching law at Oxford, reaching the rank of Professor. From 1997–2016 he was a professor of law at the Australian National University, x

Notes on contributors

where he is now an emeritus distinguished professor. From 2017–2022 he was a senior research fellow at Christ’s College Cambridge. He has been thinking and writing about legal and philosophical understandings of responsibility for more than twenty years. Peter has been a corresponding fellow of the British Academy since 2007. Andreas Brekke Carlsson is an Associate Professor at Inland Norway University of Applied Sciences. He writes on moral responsibility and blame, including issues concerning guilt, shame and blameworthiness across time. He has recently edited the volume Self-Blame and Moral Responsibility with Cambridge University Press. Paula Casal is a Professor at ICREA, and at Pompeu Fabra University’s Law Department, having held positions at Reading University (2004–2008) and Keele University (1996–2004). She was also Fellow in Ethics at Harvard University (1999–2000), a Keele Junior Research Fellow, also at Harvard (2000–2001), Hoover Fellow at Université  Catholique  de Louvain (2001–2002), Leverhulme Research Fellow at the University of Oxford  (2002–2004),  and Christopher Family Fellow at Stanford University (2018). Her work has appeared in journals like  Ethics,  Economics and Philosophy,  Journal of Medical Ethics, Journal of Moral Philosophy, Journal of Political Philosophy, Hypatia, Political Studies, and Utilitas. She has co-authored Un reparto más justo del planeta with T. Pogge and H. Steiner and Los derechos de los simios with Peter Singer. She is Associate Editor of Politics, Philosophy & Economics, Co-director of the UPF Center for Animal Ethics, and President of the Great Ape Project-Spain. Simon-Pierre Chevarie-Cossette is an Assistant Professor of Practical Philosophy at the University of Neuchâtel, Switzerland. His DPhil (PhD) thesis at the University of Oxford asked whether free will skepticism is a livable doctrine. He now works at the junction of ethics, action, and epistemology. His new project (2024–2027, SNF-funded), The DefenceFirst Approach to Responsibility, targets questions of responsibility negatively, by asking whether troubling factors like determinism or manipulation give a defence (an excuse, an exemption, or a justification). R A Duff is a Professor Emeritus in Philosophy at the University of Stirling. He has published extensively in the philosophy of criminal law – on criminal punishment, on the structures of criminal liability, on the criminal process, and on criminalization. Kimberly Kessler Ferzan is the Earle Hepburn Professor of Law and Professor of Philosophy at the University of Pennsylvania (USA). She has co-authored two monographs, coedited three volumes, and authored over one hundred book chapters, articles, and reviews. She is Co-Editor-in-Chief of Law and Philosophy. Alexander Filipović is a Professor of Social Ethics at the Faculty of Catholic Theology of the University of Vienna. Previous academic positions were in Münster and Munich. His research interests include media ethics, digital ethics, applied ethics, and political ethics. Bhargavi Ganesh is a PhD student at the University of Edinburgh, working on interdisciplinary and mixed-method approaches for designing a responsibility framework for the governance of AI and autonomous systems. She is a member of the School of Informatics’ Artificial Intelligence Applications Institute and the Edinburgh Futures Institute’s  Centre xi

Notes on contributors

for Technomoral Futures, and is affiliated with the UKRI EPSRC grant on Governance of Trustworthy Autonomous Systems. Prior to her PhD work, Bhargavi’s research focused on the impacts of consumer finance policies on marginalized groups. Jeffrey Hause is a Professor of Philosophy and Classics at Creighton University in Omaha, NE, USA. He is a co-editor (with Robert Pasnau) of the Hackett Aquinas Series and specialises in medieval philosophy and history of ethics. John Hyman is a Grote Professor of the Philosophy of Mind and Logic at University College London and an Emeritus Fellow of The Queen’s College, Oxford. His publications include The Objective Eye: Color, Form, and Reality in the Theory of Art (2006), Action, Knowledge, and Will (2015), and A Companion to Wittgenstein (ed. with H.-J. Glock, 2016). Terence Irwin is an Emeritus Professor of Ancient Philosophy in the University of Oxford. Previously he taught at Cornell University. He is the author of Plato’s Gorgias (translation and notes), Clarendon Plato Series, Oxford UP, 1979; Aristotle’s Nicomachean Ethics (translation and notes), Hackett Publishing Co., 3rd edn., 2019; Aristotle’s First Principles, Oxford UP, 1988; Classical Thought, Oxford UP, 1989; Plato’s Ethics. Oxford UP, 1995; The Development of Ethics, 3 vols. (Oxford UP, 2007–9); and Ethics Through History. Oxford UP, 2020. Alex Kaiserman is an Associate Professor of Philosophy at the University of Oxford. He has published on causation, persistence, time travel, free will, responsibility, tort and criminal liability, and the ethics of punishment. He is particularly interested in the concept of partial responsibility, and is currently writing a book on the topic. Rachana Kamtekar is a Professor of Philosophy and Classics at Cornell University. She is the author of many articles on ancient Greek philosophy and of Plato’s Moral Psychology: Socratic Intellectualism, the Divided Soul and the Desire for Good (OUP 2017). She is currently writing a book about anti-determinist arguments in antiquity, provisionally entitled Cause and Human Agency from Aristotle to Alexander. Maximilian Kiener is a Junior Professor of Philosophy and Ethics in Technology at Hamburg University of Technology. He specialises in moral and legal philosophy, with a particular focus on consent, responsibility, and artificial intelligence. Maximilian is also an Associate Member of the Faculty of Philosophy at the University of Oxford, at which he was based between 2015 and 2022, and an Associate Research Fellow at The Institute for Ethics in AI at Oxford, on the ERC-project Roots of Responsibility, based at UCL, and at the Oxford Uehiro Centre for Practical Ethics. His work has been published in, amongst others, Philosophical Studies, The Journal of Moral Philosophy, and Ethical Theory and Moral Practice. He is the author of Voluntary Consent: Theory and Practice (Routledge, 2023) and is currently writing another monograph on responsibility and artificial intelligence. Matt King is a Professor of Philosophy at the University of Alabama at Birmingham, USA, where he has taught since 2014. He works on topics in ethics, the philosophy of law, and on matters of agency and responsibility. Ironically, he does most of his philosophising sitting down.

xii

Notes on contributors

Elinor Mason is a Professor of Philosophy at the University of California, Santa Barbara. She has interests in feminist philosophy and normative ethics as well as agency and responsibility. Leonhard Menges is an Associate Professor of Philosophy at the Department of Philosophy (GW) at the University of Salzburg in Austria. He works on blame, responsibility, free will, and the right to privacy. Thaddeus Metz is a Professor of Philosophy at the University of Pretoria in South Africa. Metz is well known for having analytically drawn on ethical ideas prominent in philosophies of the Global South, particularly Africa, to address a variety of moral, political, and legal controversies. His more than 300 publications include articles in Mind, The Monist, Ethics, and Philosophy & Public Affairs. More than fifty of Metz’s works have been reprinted, sometimes into one of a dozen languages. His latest book is A Relational Moral Theory, which is the first work of African philosophy to have appeared with Oxford University Press in twenty-five years and is being translated into Chinese. Daniel J.  Miller  is an Assistant Professor of Philosophy at West Virginia University. His research consists of work on blameworthiness and ignorance, the ethics of blame and forgiveness, and moral responsibility and control. More recently he has applied his research to several questions arising in health care ethics. Peter Millican is Gilbert Ryle Fellow and Professor of Philosophy at Hertford College, University of Oxford, and Visiting Professor of Philosophy at the National University of Singapore. His philosophical interests are broad, ranging from epistemology and philosophy of language to moral philosophy and philosophy of religion. Over fifty of his published papers have been on the interpretation, analysis, and assessment of classic arguments, deriving especially from the work of David Hume, but also Anselm, Locke, Russell, and Turing. He has also worked extensively on the boundaries between AI, computing, and philosophy, as exemplified by the websites www.davidhume.org and www.philocomp.net. David Owens is a Professor of Philosophy at Kings College London. He has held visiting appointments at All Souls College, Oxford, Yale University, London University, Sydney University, New York University, and the University of Lublin. He is the author of four books:  Bound by Convention  (2022), Shaping the Normative Landscape  (2012), Reason Without Freedom  (2000), and  Causes and Coincidences  (1992). A  collection of his papers Normativity and Control was published in 2017. Thomas Pink is a Professor of Philosophy at King’s College London. He has published books and articles on ethics, political and legal philosophy, and on the metaphysics of agency, free will, and normativity. He is the author of Self-Determination (Oxford: University Press, 2017) and is currently preparing an edition of The Questions Concerning Liberty, Necessity and Chance for the Clarendon works of Thomas Hobbes. Lubomira Radoilska is a Professor of Philosophy at the University of Kent, UK. She works on issues at the intersection of philosophy of action, ethics, and epistemology. Radoilska is the author of Addiction and Weakness of Will (OUP, 2013) and editor of Autonomy and Mental Disorder (OUP, 2012).

xiii

Notes on contributors

Massimo Renzo holds the Yeoh Chair of Politics, Philosophy & Law at King’s College London, where he also directs the YTL Centre for Politics, Philosophy & Law. He has written on political authority, just war, human rights, and philosophy of the criminal law. He is currently interested in moral agency, with a focus on the notions of autonomy, responsibility and meaning in life. Alexander Sarch is a Professor at the University of Surrey School of Law. His work tackles theoretical and doctrinal questions about the criminal law, with a particular emphasis on culpable mental states and the fundamental limits on the reach of criminal liability. He has published widely on criminal culpability, willful ignorance, risk taking, wellbeing, and blame, and his current projects focus on the limits of criminalization, especially in corporate and white-collar contexts. His monograph Criminally Ignorant: Why the Law Pretends We Know What We Don’t came out with Oxford University Press in 2019, and his work has appeared in journals such as Legal Theory, Law & Philosophy, University of California Davis Law Review, Criminal Law  & Philosophy, and Philosophical Studies. Carolina Sartorio is a Professor of Philosophy at Rutgers University, New Brunswick. She is the author of Causation and Free Will (Oxford University Press, 2016), Causalism: Unifying Action and Free Action (Oxford University Press, forthcoming), and co-author (with Robert Kane) of Do We Have Free Will? A Debate (Routledge, 2021). Michael Schefczyk  is a Professor of Philosophy at KIT (Karlsruhe Institute of Technology). He holds a doctoral degree in economics and is author of a monograph on responsibility for historical injustice (Verantwortung für historisches Unrecht, 2012). Michael’s favorite philosophers are John Stuart and Harriet Taylor Mill. Before coming to Karlsruhe, Michael was at Leuphana University Lüneburg and an assistant professor at the University of Zürich, where he submitted his Habilitation in 2007. Henry Shue is a Senior Research Fellow at the Centre for International Studies, Department of Politics and International Relations, University of Oxford (2007–present). He graduated from Davidson College in 1961; studied at Merton College, Oxford, as a Rhodes Scholar; and received a PhD from Princeton University. After teaching at a number of US universities, including the University of Maryland and Cornell, he returned to Merton in 2002 as Senior Research Fellow and Professor of International Relations. His best known work is Basic Rights (Princeton University Press, 1980; 40th anniversary edition, 2020). He has been studying the ethical and political issues about climate change for thirty years; and his first two decades of writings on climate change were published as Climate Justice: Vulnerability and Protection (Oxford University Press, 2014). His most recent book is The Pivotal Generation: Why We Have a Moral Responsibility to Slow Climate Change Right Now (Princeton, 2022). Anthony Skelton is an Associate Professor of Philosophy at the University of Western Ontario, Canada. He specialises in normative ethics and the history of ethics. His work in these areas has appeared in the journals Ethics, Canadian Journal of Philosophy, Journal of Ethics and Social Philosophy, Journal of the History of Philosophy, and Utilitas. He is the author of Sidgwick’s Ethics (Cambridge University Press, forthcoming).

xiv

Notes on contributors

Jenny Steele is a Professor of Law at the University of York. She holds an LLM from the University of Bristol and studied Jurisprudence at Jesus College, Oxford. Steele’s research features, among other aspects, the law of obligations, especially in tort law, the relationship between law and risk, and the role of insurance in private law. Steele is the author of Tort Law: Text, Cases, and Materials, published by Oxford University Press and currently in its fifth edition. David Sussman is the Oscar R. Ewing Associate Professor of Philosophy at Indiana University, having taught previously at the University of Illinois at Urbana-Champaign and Princeton University. His research interests include ethics, Kant, and the philosophy of action. Daniel Telech is a Senior Research Fellow at Lund University, as part of the Lund-Gothenburg Responsibility Project. His research focuses on questions concerning the nature and norms of blame and (especially) praise, moral luck, and relational normativity. Patrick Todd is a Senior Lecturer at the University of Edinburgh, where he has been based since 2013. For the 2022/2023 and 2023/2024 academic years, he is based in Lund, Sweden, as Senior Research Fellow at the Lund/Gothenburg Responsibility Project. He writes on various aspects of the problems of free will and moral responsibility, including issues regarding incompatibilism and compatibilism, and the standing to blame. He has recently published a book with Oxford University Press in metaphysics, The Open Future: Why Future Contingents are All False. Shannon Vallor is Baillie Gifford Professor of Ethics of Data and Artificial Intelligence at the University of Edinburgh, where she is appointed in Philosophy and serves as Director of the Centre for Technomoral Futures in the Edinburgh Futures Institute. She is Turing Fellow and co-Director of the BRAID (Bridging Responsible AI Divides) program of the UKRI’s Arts and Humanities Research Council. She is also Principal Investigator on Responsibility and co-Investigator on the Governance and Regulation node of the UKRI’s Trustworthy Autonomous Systems programme. She is the author of Technology and the Virtues (Oxford University Press, 2016) and editor of the Oxford Handbook of Philosophy of Technology (Oxford University Press, 2022). Marcel van Ackeren is a Senior Lecturer for Ethics at the Faculty of Chemistry and Pharmacy at Wuerzburg University. He worked as a Henkel Fellow and Research Associate at the Oxford Uehiro Centre for Practical Ethics and held Fellowships and Positions at the Centre for Advanced Studies (Wissenschaftskolleg, Greifswald) and the Centre for Advanced Studies in Bioethics (Muenster). He is also an Associate Member of the Faculty of Philosophy in Oxford, where he started to work as a philosopher conducting doctoral research supervised by B. Williams and M. Frede. Bruce Waller (1946–2023) was a Professor of Philosophy at Youngstown State University, where he taught for twenty-nine years. Waller’s research includes several articles and books on free will and moral responsibility, such as Freedom without Responsibility (1990), The Natural Selection of Autonomy (1998), Against Moral Responsibility (2011), Restorative Free Will: Back to the Biological Base (2015), The Stubborn System of Moral Responsibility (2015), The

xv

Notes on contributors

Injustice of Punishment (2017), Free Will, Moral Responsibility, and The Desire to be a God (2021), and The Deep Roots of American Neoliberalism (2022). Bruce Waller was also the co-editor of The Routledge Handbook of the Philosophy and Science of Punishment (2020) and the author of Critical Thinking: Consider the Verdict, now in its seventh edition. Andrea Westlund is a Professor of Philosophy at Florida State University. She works mainly on autonomy and related topics, including shared deliberation, narrative self-understanding, blame and forgiveness, and the moral emotions involved therein. David B. Wong is the Susan Fox Beischer and George D. Beischer Professor of Philosophy at Duke University. His research is principally in ethical theory (moral relativism, pluralism, and disagreement) and Chinese/Western comparative philosophy (classical Confucianism and Daoism).

xvi

PREFACE

Writing this preface fills me with joy and gratitude. It represents the final milestone of a truly unique and exceptionally stimulating intellectual journey, which has allowed me to connect with 41 contributors across the world and to embark on a shared mission to explore the philosophy of responsibility. I was thoroughly impressed by the commitment and excellence of my authors, and I am deeply grateful for everything I learned from them. Although many warned me about editing a Handbook, let alone being its sole editor, I harbor no regret. Editing the Handbook provided me with new ways to deepen and broaden my knowledge of the philosophy of responsibility, and I hope that its readers will accrue similar benefits. But, as with any project of such scale, this Handbook required genuine teamwork. I thank Routledge, especially Tony Bruce and Adam Johnson, for their outstanding support throughout all stages of the process. I particularly thank the Leverhulme Trust for supporting me with an Early Career Fellowship and giving me the necessary freedom to take on such a large project. I also thank University College, Oxford, and the Faculty of Philosophy at Oxford for hosting the seven workshops I organized in connection with this Handbook. During these workshops, all contributors could present their draft chapters and receive feedback. It was during these meetings that work on the Handbook was at its most interactive and collaborative. I am also very grateful to everyone who participated in these workshops and helped the contributors by asking questions and offering comments. In addition, I am extremely grateful to Ben Colbourn, Roger Crisp, Thomas Douglas, John Hyman, Lisa Forsberg, Leonhard Menges, Peter Schaber, Tracey Sowerby, Victor Tadros, Carissa Véliz, and Dominic Wilkinson for their invaluable advice during the preparation of the book proposal or editorial stages, as well as to the various anonymous referees who provided constructive criticism of earlier versions of the chapters. Finally, I wish to express my greatest appreciation for Bruce Waller, who wrote an outstanding chapter for this Handbook. But before he could see it in print, he died on 08 February 2023. I have been an admirer of Bruce’s work for a long time, and I was hugely saddened when I learned of his death. It is a special privilege and honor to have one of

xvii

Preface

Bruce’s last pieces of work in this Handbook, and I am sure it will continue to inspire readers, just as Bruce’s work always has. The philosophy of responsibility is a large and constantly evolving field, so the Handbook has not had an easy task. Nevertheless, I hope it will provide both orientation and innovation across this fascinating area of philosophy. Maximilian Kiener, Oxford, May 2023

xviii

INTRODUCTION Maximilian Kiener1

To be morally responsible means to be eligible for some form of moral assessment. It is normally an assessment of whether one has met specific moral standards or norms, and it presupposes that the conduct or action in question was the result of exercising certain agential capacities. What is more, to be morally responsible also means that others may hold one responsible, namely by means of certain responses or reactions, such as requests for a justification or explanation, the expression of praise or blame, or the demand for an apology. Yet, philosophical controversy is everywhere, including over what the metaphysical presuppositions of responsibility are, what types of responsibility exist, including a distinction between retrospective responsibility for past conduct and prospective responsibility for future conduct, what conditions responsibility has, and what ought to guide the social, institutional, or legal settings in which we hold each other responsible. So understood, the topic of moral responsibility is both extremely old and distinctively new. It is extremely old because the various concepts closely related to moral responsibility, such as human agency, freedom, and the practices of blaming and praising, have been discussed by philosophers since antiquity. For instance, the Stoics explicitly asked what it takes for our actions to be ‘up to us’ or ‘in our power’ (eph’ hêmin), and whether genuinely free action was possible if Chrysippus was right and fate determines everything by antecedent causes (‘omnia fato fieri causis antecedentibus’) (reported in Cicero 1991: De Fato, 41.6–7).2 Aristotle developed a theory of voluntariness (hekon) and claimed that only voluntary conduct is worthy of praise or blame. He then defined voluntariness as requiring the absence of both force (bia) and ignorance (agnoia), and thereby anticipated a framework that resembles what we would now call the control condition of responsibility, challenged by determinism, force, and coercion, and the epistemic condition, challenged by ignorance, lack of awareness, and deception (Aristotle 2014, edited by Barnes and Kenny).3 And Aquinas provided an elaborate account of distinctive responsibility practices. One of these practices, Aquinas says, is connected to the emotion of ‘vengeful anger’ and motivated by a concern for justice (Aquinas 2010: ST, I-II 47.2, 75–76, 80), while another is connected to the aim of ‘fraternal correction’ and motivated by the virtue of mercy (Aquinas 2010: ST, Q33, Disputed Question on Fraternal Correction).4

1 DOI: 10.4324/9781003282242-1

Maximilian Kiener

Yet, despite such pertinent scholarship, of which my examples are only the tip of the iceberg, moral responsibility is also a distinctively new, or at least genuinely modern, topic. To say that it is new or modern means that central paradigms in the debate on responsibility have shifted since early modernity and thereby reframed the scholarship on responsibility up to this day. Three developments are especially noteworthy. First, the Scientific Revolution in the 16th and 17th centuries, including the groundbreaking work of Nicolaus Copernicus and Isaac Newton, fundamentally changed our understanding of the natural world and the place of humans within it (cf. Principe 2011). Of particular importance was the rise of mechanistic determinism, the view that every event is necessitated by a cause and that all causation happens through the contact of one body with another (Martinich 2016). Thomas Hobbes was a paradigmatic proponent of this view and emphasized that mechanistic determination not only applied to the natural world but also to human conduct, including to the human will.5 Mechanistic determinism was a paradigm shift in at least two major respects: first, mechanistic determinism, i.e., determinism by natural causes, was different from Stoic providential determinism, i.e., determinism by (divine) fate, and thus explained the determination of human conduct in very different terms. Second, mechanistic determinism opposed scholastic and ancient views about the world being in a state of teleology where everything has its own inner telos, or purpose, which it strives to fulfil (Aristotle 1961, 2014). Rather, mechanistic determinism considered the natural world free of any genuine intrinsic, natural values or purposes, so that whatever shaped human conduct could no longer be understood as the fulfilment of some inner telos (Cf. Mill 1874: “On Nature”), but more likely as the result of external force. Today, further scientific advances build on this legacy and follow the two major paradigm shifts. Modern physics (Hodgson 2011), neuroscience, including the famous experiment by Benjamin Libet6 (Bayne 2011), as well as recent advances in genetics (Willmott 2016) and behavioral psychology (Gilovich et  al. 2002), have further shaped our understanding of the driving forces behind human action. And while some of the 20th century’s key texts on moral responsibility hold a wider significance, it is against this distinctive background of modern scientific progress that they are normally read. For instance, Peter Strawson’s Freedom and Resentment formulated a new approach to adjudicate between opposing views on whether determinism precludes genuine moral responsibility by focusing on what he called reactive attitudes, such as resentment and indignation, and the significance of our interpersonal interactions (Strawson 1963). Harry Frankfurt focused on ‘the ability to do otherwise’, i.e., the ability to do or not do a certain act, which is seen to be precluded by determinism, and claimed that it is not needed for moral responsibility (Frankfurt 1969). Moreover, Thomas Nagel’s Moral Luck examined in closer detail how a lack of control in various ways, including over the outcomes of one’s actions, the circumstances of one’s actions, or even one’s character, might affect our moral responsibility (Nagel 1979). Second, from early modernity onwards, philosophers have developed and relied on a distinction that was not present in Ancient or scholastic thought, namely between ethics and morality. As Simon Robertson explains, ‘[o]n a wide construal, ethical enquiry [i.e., ethics] encompasses the whole domain of normative thinking as it bears on how people should live and what a good life involves’ (Robertson 2020: 18). This includes the Socratic question “How should one live?” and the pursuit of happiness (eudaimonia) that was so characteristic of Ancient philosophy. From early modernity, however, philosophers began to focus on a more restricted domain, which is often termed ‘morality’. Bernard Williams explains morality in this narrower sense as ‘any scheme for regulating the relations between people 2

Introduction

that works through informal sanctions and internalised dispositions’ (Williams 1995: 241), and he emphasizes that morality, so understood, centers on the concept of obligations owed to others and the liability to blame (Williams 1985, ch. 10). Examples of this modern approach to morality include contractualist theories, as pioneered by Hobbes, Rousseau, and Locke, which aim to guide the resolution of interpersonal conflict rather than the pursuit of a good life, deontological theories, as championed by Kant, who dethroned happiness as the supreme principle of normative thought and put the notion of obligation at the center of his view, and even consequentialist theories, within which – for instance – J S Mill argued that [w]e do not call anything wrong, unless we mean to imply that a person ought to be punished in some way or another for doing it [either] by law . . . by the opinion of his fellow creatures . . . [or at least] by the reproaches of his own conscience. (Mill 2015, Utilitarianism: ch. 5)7 Today, the prevalence and further development of these moral theories build on this legacy and make it the case that debates about responsibility are mostly fostered in this narrower field of morality. In fact, the central themes of moral responsibility are also central themes of morality in the narrow sense. To begin with, obligations have become a central aspect in debates around responsibility. For instance, Antony Duff claims that the scope of responsibility, i.e., what we are responsible for, can be determined by our (prospective) obligations (Duff 2009).8 Moreover, liability to blame, sanctions, or redress has been one of the most, if not the most, discussed topic in current debates on responsibility. Philosophers often characterize being responsible either in terms of being a fitting target to reactive attitudes, including blame, and therefore already liable to blame, or at least in terms of being eligible for moral assessment that can eventually lead to blame (Pereboom 2001; Smith 2005). Furthermore, it is most noteworthy that contemporary philosophers also reflect something close to the morality–ethics distinction when they separate in their accounts of responsibility a deontic dimension, concerned with rights and obligations and related to ‘morality’ in the narrow sense, from an aretaic dimension, concerned with virtue and character and related to ‘ethics’ more generally. An example of one such philosopher is Gary Watson, who famously introduced ‘two faces’ of responsibility, namely attributability, which is concerned with a virtue-related assessment of a person’s conduct (close to an aretaic dimension), and accountability, which is concerned with blaming practices and the violations of rights and other interpersonal moral norms (close to the deontic dimension) (Watson 1996). It is in these instances where the focus on ‘morality’ as opposed to ‘ethics’ is loosened, but still never lost as the central anchor of the debate. Finally, the third development, which makes the debate around moral responsibility distinctively modern, concerns the rise of liberal politics and modern legal systems. In Ancient philosophy, it was unthinkable that a human could live outside a political community. Aristotle said that man is by nature a social animal and he ‘who is so self-sufficing that he has no need to . . . [be] part of a state . . . must be either a lower animal or a god’ (Aristotle 1959: Politics, 1253a27). From early modern political philosophy, however, philosophers disagreed. Hobbes, Locke, and Rousseau considered living in a state of nature, one which pre-existed any political association, as a real (albeit unattractive) alternative, and asked how to justify a state’s coercive power at the center of their philosophical thought. But with the requirement to justify state power also (gradually) came the accountability of those 3

Maximilian Kiener

exercising the power towards those affected by it. Today’s democracies explicitly feature state accountability as a core principle (Przeworski et al. 1999) and many current social and political movements, such as #MeToo, Yellow Vests, or Fridays For Future, emphasize various aspects of accountability and responsibility too. But modern liberal political systems have not only featured state responsibility but also individual responsibility. The focus on the individual and their consent or agreement to being governed has played such a significant role in political and moral theories since Hobbes that claims about individual responsibility could be supported on this basis too, even outside academic discourse. For instance, various politicians across the globe have stressed the significance of individual responsibility. Ronald Reagan famously said that ‘We must reject the idea that every [time] a law’s broken, society is guilty rather than the lawbreaker. It is time to restore the American precept that each individual is accountable for his actions’ (quoted from Barker 2009: 66). And Tony Blair, in notable agreement with Margaret Thatcher, emphasized the ‘responsibility on the individual – to take the chance, to make something of their lives and use their ability and potential to the full’ (quoted from Baldock et  al. 2007: 702) Some authors, such as Yascha Mounk, therefore even argue that 20th-century politics created an age of individual responsibility (Mounk 2017),9 where an individual’s responsibility – including retrospective responsibility for one’s past conduct as well as the prospective responsibility or obligation to take care of oneself – is considered more important than structural or societal problems more generally. Contemporary philosophical scholarship may not fully align with these political trends all the time. Yet, philosophers have also focused on individual responsibility in political contexts, for instance when discussing how, or to what extent, individual responsibility for some of one’s disadvantages precludes one’s claim to public assistance (Knight & Stemplowska 2014; Scanlon 2000). Closely related to the rise of modern liberal politics is also the development of modern legal systems. Questions about responsibility have featured prominently in the criminal law (Duff & Green 2013), for example when deciding what type of misconduct qualifies as a serious or ‘public’ wrong, as well as in discussions about the relationship between morality and law. Other examples come from tort law (Oberdiek 2018), where various forms of liability and responsibility are developed, public law (Cane 2002), where the specific relation between the state and its citizens shapes ascriptions of responsibility, and international law (Besson 2022), where increasing globalization creates new power as well as challenges from collective action. In all these areas, distinctive questions of responsibility arise and connect to our modern political and legal systems. But it is also noteworthy that these developments in politics and law have not only called attention to institutionalized contexts of responsibility, but also to a greater variety of responsibility practices in social settings that pertain more to our personal and private lives (Hutchison et al. 2018; Mason 2019). Taken together, scientific advances, the idea of ‘morality’, and the changes in our politics, laws, and societies have rendered moral responsibility a distinctively modern topic. Today’s philosophical inquiry into moral responsibility is a very well-developed, yet still rapidly growing field. The debates not only reinvigorate traditional questions about free will and the metaphysics of responsibility, but also touch on new terrain by analyzing responsibility in its social, institutional, and legal contexts. With emerging technologies, further scientific advances, climate change, social inequality, and issues of global and intergenerational justice, the interest in moral responsibility is also likely to grow even further. But above all, it is most noteworthy that the topic of moral responsibility has become a focal point where 4

Introduction

significant developments in science, politics, and society meet foundational philosophical research on free will, agency, and morality. It is this vast debate that this Handbook ambitiously aims to capture. Its objective is to provide a comprehensive, yet by no means exhaustive, companion. It seeks to capture the old and the new, the history of the philosophy of responsibility as well as current and future debates surrounding it. And it looks to bring different perspectives into dialogue with each other. In so doing, this Handbook strives to be innovative by not only reflecting the current state of the art, but also by allowing its authors to make an original contribution to the field. The selection of topics was influenced by other related publications too. I  hope the Handbook will prove useful alongside the complementary Oxford Handbook on Moral Responsibility, edited by Dana Nelkin and Derek Pereboom, who ‘provide a comprehensive survey of scholarship on this topic since 1960, with a focus on the past three decades’. Unlike that particular work, the current Handbook has taken a wider perspective both historically and systematically, and places an additional emphasis on the ethics, law, and politics of responsibility in Part 3. Additionally, this Handbook can complement the Routledge Handbook of Collective Responsibility, edited by Saba Bazargan-Forward and Deborah Tollefsen, which focuses on collective responsibility, while the current Handbook mostly features individual responsibility. The present Handbook aims to be an easily accessible and comprehensive resource for advanced undergraduates, those wishing to learn about the state of the art of debates on responsibility, established researchers in the field, and practitioners of law and policymakers who wish to deepen their understanding of the philosophy of responsibility.

Overview The Handbook is divided into three major parts. Part 1, THE HISTORY OF RESPONSIBILITY, will first feature important figures in the history of philosophical ideas, laying out the rich resources that existed before responsibility became an explicit topic in contemporary scholarship. In this part, the Handbook will show how key figures in the history of ideas have shaped our thinking about responsibility, but also highlight how their approaches differ from how we think about responsibility today. This section starts with two distinctive perspectives on wrongdoing, namely with chapters on Plato on vice and Hegel on guilt. Marcel van Ackeren discusses Plato on vice, and in particular Plato’s view that vice is ignorance and that wrongdoing is involuntary. Although these two key claims are pillars of Plato’s moral philosophy throughout his entire work, van Ackeren shows how their underpinnings change significantly throughout Plato’s work and allow us to derive important Platonic resources for the debate on responsibility today. Mark Alznauer continues with a chapter on Hegel on guilt. Alznauer argues that the single most striking feature of Hegel’s treatment of Schuld (responsibility or guilt) is the way Hegel situates it within his social theory, namely by developing a social, rather than formal or individualistic, conception of agency. This approach, Alznauer explains, develops from Hegel’s treatment of the inner-outer problem in logic and decreases the significance of an agent’s immediate consciousness when it comes to determining her intentions or motives. Alznauer’s discussion of Hegel starts a theme that will runs through the Handbook, namely one in which responsibility is considered from a genuinely social, rather than metaphysical perspective (see also the contribution from Vallor and Ganesh). 5

Maximilian Kiener

The Handbook then continues with four distinctive approaches to responsibility and determinism, namely with chapters on the Stoics on responsibility in a world of (divine) providence and fate, on Thomas Hobbes on responsibility in a world of mechanistic determination, on David Hume’s account of causation, determinism, and intentional action, as well as on Henry Sidgwick’s view that the resolution of the debate on determinism and free will is of little significance to ethics. Rachana Kamtekar focuses on the claim of Ancient Greek and Roman Stoics that the fatedness of all things is compatible with actions being attributable to agents. Kamtekar outlines various criticisms against Stoicism in this regard and elaborates on how and why the Stoics could explicitly reject attitudes of blame and anger for wrongdoing, while still adhering to the view that the practices of blame, such as reproach and punishment, can be justified for other reasons, including forward-looking reasons. Thomas Pink focuses on Hobbes and mechanistic determination. Pink explains how  Hobbes responded to a previous debate about moral responsibility between Protestant and Catholic theologians conducted within Aristotelian scholasticism. The theologians assumed that responsibility depended on non-mechanical forms of power – on our responsiveness to a distinctive power of reason and on the exercise by us of an equally distinctive power of self-determination. They mainly debated how far the exercise of selfdetermination need involve contingency. Pink shows how Hobbes rejected these powers of reason and of self-determination, and held that all human action is necessitated by prior causes from within material nature. Peter Millican introduces the reader to David Hume, and in particular to Hume’s views on causation, determinism, intentional action, and responsibility. Though very famous, Millican argues, the key discussions ‘of liberty and necessity’ from the 1739  Treatise of Human Nature and the 1748 Enquiry concerning Human Understanding appear to conflict in various respects, and Hume’s position is commonly misunderstood. Millican explains how these interpretive difficulties can be reconciled, revealing a consistent compatibilist position which makes space for moral responsibility within a sentimentalist virtue ethics. Finally, Anthony Skelton focuses on Henry Sidgwick’s claim that resolving the free will controversy is of “limited” significance to ethics. Skelton clarifies the premisses behind Sidgwick’s argument, addresses various concerns about Sidgwick’s position, and argues Sidgwick’s position is best understood as an intellectual forerunner to views attempting to establish which aspects of moral argument (if any) survive the discovery that the world is deterministic. The next group of articles focuses on the scope of responsibility, that is, the range of things people are responsible for. Terrence Irwin starts from Aristotle’s view that only voluntary conduct deserves praise or blame. Drawing on Aristotle’s three different ethical treatises, Irwin focuses especially on ignorance and forethought, i.e., aspects that we would now consider to be part of an epistemic condition of responsibility. Irwin’s contribution stands out, among other reasons, by connecting the philosophical discussion of Aristotle’s views with challenges from questions arising from Athenian law and actual Ancient Athenian legal cases. David Sussman then presents Kant’s conception of moral responsibility, arguing that it is perhaps the most radical view in the history of philosophy. According to Sussman, Kant insists ‘that a person is fully responsible not just for her actions as an adult but for her entire life and character, despite all the natural and historical contingencies that condition her existence.’ Sussman argues that Kant’s idea of an “intelligible” choice, as developed in 6

Introduction

Religion within the Limits of Reason Alone, is key in this regard and connected to Kant’s retributivist and ‘quasi-legal’ approach to responsibility. The last cluster of chapters in this first part addresses responsibility beyond the merely individual sphere and takes into account the relevance of groups, communities, and society (thereby continuing a theme that, as I mentioned earlier, started with Alznauer’s discussion of Hegel). David B. Wong provides perspectives from Chinese Philosophy. Wong emphasizes that Confucian views consider relationships as integral to a fully realized human life and adds that these roles are often structured by social roles, both public and private, as in cases of family and friendship, and political or societal, as in cases of one’s relation to a government. Moreover, this relational perspective also informs key philosophical concepts in Ancient Chinese thought, including ze 責 duty and ren 任 responsibility (the latter especially related to holding a position or office). Jeffrey Hause then focuses on two distinctive blaming practices in Aquinas, as mentioned earlier in this introduction. One of these practices is connected to vengeful anger and the virtue of justice, while the other is connected to fraternal correction and the virtue of charity. Hause describes how Aquinas attends closely to the place of individuals within their respective communities or societies, as well as why Aquinas’ view must appear to contemporary readers as an especially distinctive approach to blaming. Part 2, THE THEORY OF RESPONSIBILITY, is subdivided into two main sections. It starts with various concepts around responsibility, where the authors explain how responsibility connects to important other concepts in philosophy, such as agency, causation, the self, emotion, judgment, obligation, virtue, and autonomy. Taken together, this section could be understood in terms of what Peter Strawson would call a ‘connective conceptual analysis’, i.e., a conceptual map that describes the various interdependent connections between responsibility and other concepts, rather than just breaking them down individually. In so doing, the section enables readers to appreciate the place of responsibility in philosophical thinking more widely, and to gain an understanding that the specialized research contained within monographs and articles cannot provide. Maria Alvarez discusses responsibility and agency. Alvarez defends the claim that responsibility depends on agency, while rejecting the view that we are responsible only for our actions and their consequences. Rather, we are responsible for things over which we have some control – control that, she argues, depends on the possession of distinctive agential capacities. To spell out this idea, Alvarez focuses on so-called ‘two-way’ powers or abilities. She claims, further, that while control is a condition for responsibility, questions about degrees of blame or merit for the things we are responsible for depend on a variety of factors, including whether the agent had ‘eligible’ alternatives. Alex Kaiserman discusses the view that responsibility requires causation. Kaiserman concedes that such a view faces serious challenges in the context of collective action. Yet, he nevertheless argues that it can still be defended on a more nuanced understanding of the metaphysics of causation, including a proper understanding of difference-making, (collective) omissions, degrees of causal contribution, and an arguably neglected distinction between partially causing X and causing a part of X. Monika Betzler discusses responsibility and the self, and in particular the view that we can only be responsible for things that, in some suitable sense, express who we are. Betzler outlines the theoretical benefits of such, often called Deep Self Views, and addresses various challenges, including the criticism that the reflection of a deep self in someone’s conduct is either unnecessary or insufficient for their responsibility. In response, Betzler motivates a 7

Maximilian Kiener

modified version of the Deep Self View, but also flags persisting challenges for this approach to moral responsibility. Andreas Brekke Carlsson focuses on responsibility and emotion. Carlsson shows how Peter Strawson’s seminal publication Freedom and Resentment has put moral emotions at the center of the debate on moral responsibility. More particularly, Carlsson outlines how being responsible, as well as being held responsible, are intimately connected to various emotions, including guilt and remorse, or resentment and indignation, as well as to the justification of these emotions. This raises the question of whether, and if so, how, an account of moral emotions can help us understand the conditions of moral responsibility. Finally, Maximilian Kiener shows how responsibility connects to judgment, obligation, virtue, and autonomy, by elaborating on three different approaches to moral answerability. Kiener explains how different views of answerability start from a common basis but differ when explaining answerability either in terms of a reflection of judgment, a social or legal practice, or a disposition for an inner dialogue. Kiener argues that the three accounts offer unique insights and they are all needed for a comprehensive understanding of moral responsibility. The next three chapters address selected themes from the debates on determinism and free will. Carolina Sartorio makes a case for compatibilism of determinism and free will, considering free will in terms of the kind of freedom or control required for moral responsibility. After outlining the general motivations for the compatibilist approach, Sartorio discusses two different types of compatibilism, namely ‘leeway’ compatibilism – the view that free will requires options that one could have chosen from (“alternative possibilities” of action, or the ability to do otherwise) – versus ‘source’ compatibilism – the view that free will depends on the actual (rather than alternative or counterfactual) sources, motivations, or causes of one’s behavior. Patrick Todd focuses on incompatibilism and, in particular, argues that incompatibilism is often misrepresented as the view that we would have to give up the reactive attitudes if we found out that determinism is true. Todd suggests that incompatibilists can reject this view in two ways. The flip-flopping incompatibilist rejects the (implied) conditional ‘If we found out determinism is true, we would find out that no one deserves the reactive attitudes’, while the innocent incompatibilist rejects the (also implied) conditional ‘If we found out that no one deserves the reactive attitudes, we would have to give up the reactive attitudes’. Lastly, Simon-Pierre Chevarie-Cossette shows how skepticism about moral responsibility faces a serious challenge. Chevarie-Cossette claims that whenever we rationally deliberate about what to do, we presuppose that we have at least more than one option available to us. Yet, the skeptical position, which portrays free will as an illusion, denies exactly that and therefore threatens to undermine the rationality of our deliberations, rendering skepticism a doctrine that is impossible to live by. Chevarie-Cossette discusses what the actual presumptions in our rational deliberations are and critically assesses a popular attempt by skeptics (and others) to show that good deliberation only requires a belief in its efficacy. Finally, the remaining chapters focus on the conditions of responsibility, especially on what responsibility requires beyond the aspects discussed in the context of free will and determinism. The authors direct their attention to conditions of control (and how they may be undermined by coercion or manipulation), to conditions of knowledge (and how they may be undermined by ignorance), and to conditions of capacity (and how they may be undermined by mental illness and injury). An additional contribution on conventions 8

Introduction

then casts a critical perspective on these chapters and asks whether focusing on an agent’s psychology is sufficient to determine responsibility. Massimo Renzo focuses on responsibility and manipulation. Renzo is interested in classic cases in which a manipulator deliberately influences a target to act in certain ways by shaping the target’s perception of their situation. When wrongful, manipulation makes a person fail to respond adequately to the reasons they have and thus interferes with their practical agency. Renzo uses this analysis to explain the impact of manipulation on various types of responsibility, including attributability and accountability (building on the work of Gary Watson), as well as the sort of responsibility that grounds liability to defensive harm. Carla Bagnoli focuses on responsibility and coercion. She distinguishes cases of coercion, in which a person’s responsibility for action is suspended due to severe impediments of that person’s agential powers, and cases of coercive threats, in which the person’s agential powers are exploited or distorted, rather than fully undermined. In the latter cases, the victim may have reasons to take or claim responsibility for submission or resistance. More complex cases of coercion build on oppressive practices and structures, and raise distinctive issues regarding the responsibility to resist. Daniel J. Miller focuses on the epistemic condition, i.e., the condition that responsibility requires some sort of knowledge, awareness, or at least the absence of certain types of ignorance, and he outlines two major approaches: one approach centers on the notion of reasonable expectation about what agents could or could not have been aware of, while the other approach features the notion of an agent’s quality of will. Miller explains how these views differ with respect to moral ignorance and what agents can be directly (as opposed to derivatively) blameworthy for, and he develops a novel hybrid proposal that combines their strengths. Lubomira Radoilska discusses two challenges related to a person’s competence. The first concerns ‘hypo-agency’ and a decreased ability to distinguish right from wrong, while the second concerns ‘hyper-agency’ and an increased such ability, so that the distinction between right and wrong becomes disproportionately salient or irresistible. Radoilska states that neither hypo- nor hyper-agency necessarily thwarts moral competence. In drawing on memoirs of people with bipolar disorder, autism, and schizophrenia, Radoilska argues for a contextual assessment of competence and responsibility. David Owens considers the relation between responsibility and conventions. Owens focuses on cases in which it is admitted that a person did something wrong so that they could only avert, or mitigate, blame by means of an excuse, rather than justification. More specifically, Owens focuses on motivational excuses, drawing on things like fear, anger, or exhaustion, rather than epistemic excuses, which concern various types of ignorance. On this basis, Owens argues that conventions shape what constitutes a legitimate motivational excuse, at least in cases where conventions also determine a person’s obligations. Part 3, THE PRACTICE OF RESPONSIBILITY, then looks at how responsibility shapes our interpersonal interactions in various contexts and how it raises serious ethical, political, and legal questions. This part is also subdivided into three sections: Blame and praise, The ethics and politics of responsibility, and The law of responsibility. In the first section, entitled Blame and praise, the authors address practices of blaming and praising in our interpersonal interactions, describe their value and disvalue, and focus on related questions such as when, if ever, people are in a position or have the standing to blame others (where, for instance, hypocrisy might undermine standing), or what is morally required in the aftermath of justified blame, e.g., apologizing and forgiving. 9

Maximilian Kiener

Leonhard Menges discusses blaming, which is a topic that has not only shaped the debate on moral responsibility but also gained currency in moral psychology more generally. Menges outlines what he calls functionalist, non-functionalist, and mixed theories about the nature of blame and asks which desiderata an adequate theory of blame should meet. Menges then opposes the requirement of a unifying account of our everyday blame practices and suggests that an adequate theory of blame should rather help us address distinct philosophical challenges, including questions about free will. Daniel Telech focuses on praise as an integral part of our responsibility practices. In outlining the basic features of praise, Telech argues that praise is a ‘communicative phenomenon’ as it seeks uptake from its addressee. Specifically, Telech maintains, praise invites its target to accept credit in valuing the praiseworthy action from the praiser’s perspective. Telech further develops this communicative view by defending it against three objections, namely that it is descriptively inadequate regarding our actual practices, that it is morally mistaken when we owe others a response, and that it implies redundant communication. Matt King elaborates on the standing to blame, with a particular focus on hypocritical blame King’s chapter shows how and why blaming can still be inappropriate even if it targets the right people, for the right things, and to the right extent. In so doing, King elucidates important conditions of appropriate blame as they apply to the person blaming, rather than the people or the actions being blamed. King outlines different methodologies and explanations of the standing to blame, addresses potential grounds for skepticism, and closes with a comparison to praise. Andrea Westlund focuses on three models of apologies, which she calls the transactional model, the conversational model, and the narrative model. Westlund contends that each of them helps us understand key features of our moral practices around apology and forgiveness, but she also adds that ‘a form of joint storytelling’ holds greater significance than any of these models has so far recognized. But only once such storytelling is better understood, Westlund says, can we see how apologies are mutually empowering rather than asymmetrically dominating. Elinor Mason focuses on ‘taking responsibility’ in a distinctive philosophical sense, according to which it does not merely mean acknowledging one’s responsibility but rather extending it, even in retrospect, and thereby changing the ‘responsibility facts’. The steep challenge for such a view, Mason adds, is to explain how one’s responsibility can become at least partly independent from the casual history of one’s conduct and instead be based on an act of will or declaration. Mason suggests that this challenge can be met and develops her proposal in response to views proposed by Susan Wolf, Joseph Raz, and David Enoch. Bruce Waller separates the type of moral responsibility associated with desert of blame and praise from other types of responsibility that lack this association. More specifically, Waller objects to the idea that we ever genuinely deserve to be blamed or praised, and instead argues for what he calls ‘take-charge’ responsibility, a distinctive form of responsibility that people claim for their own lives, projects, and choices. Such responsibility, Waller adds, is like role-responsibility, not a desert-based form of moral responsibility. Yet, it holds significant value for our lives and ought to be preferred over the traditional desert-based model. Thaddeus Metz explores the topic of reconciliation in the African tradition(s). In particular, he explains the implications of these traditions of thought on three specific matters, namely on how the punishment of offenders ought to help them reform their moral flaws, on how the compensation of victims ought to improve their lives rather than merely restore some previous situation, and on how an offender’s family members may become 10

Introduction

responsible too. In each case, Metz argues, an idea of reconciliation, as inspired by African approaches, can teach us something important and therefore ought to guide how we hold people responsible for their wrongdoing. The second section, entitled The ethics and politics of responsibility, focuses on selected challenges in applied ethics and politics. Authors address the significance of responsibility in the context of artificial intelligence, historical injustice, and climate change, amongst other topics. Bringing these topics into one place promises to create novel synergy between hitherto separate fields of inquiry. Shannon Vallor and Bhargavi Ganesh focus on responsibility in the context of artificial intelligence (AI). They explain how AI systems have acquired effective agential power and, as a result, conflict with our current frameworks of moral and legal responsibility. Vallor and Ganesh argue that this situation creates serious challenges for social trust and legitimate governance, and they recommend that we therefore reframe the debate about AI and responsibility in terms of ‘a relational practice of social care’, considerate of asymmetries in vulnerability, agency, and power. Michael Schefczyk focuses on historical injustices, understood as large-scale wrongs in the distant past, and related obligations to rectify, compensate, or apologize. Schefczyk discusses what he calls interactional approaches to historical injustice, which focus on the interaction between the potentially numerous perpetrators and victims, as well as structural approaches, which center on institutionalized or socially mediated forms of wrong and injustice. Schefczyk suggests that we make room for both approaches, proposes an account on how to examine remedial responsibility for historical injustices, and also examines the excuse from cultural ignorance. Alexander Filipović focuses on Corporate Digital Responsibility (CDR), understood as a form of responsibility that applies to organizations, companies, NGOs, or institutions specifically in the digital age. Filipović explains how CDR further updates the better-known idea of Corporate Social Responsibility (CSR) in contexts where society, politics, and economies are increasingly shaped by new digital technologies. At the same time, however, CDR is firmly rooted in business ethics and inherits norms on the role responsibilities and mutual accountability of various agents and entities in the business world. Henry Shue outlines the role of banks when they give loans to fossil fuel companies. He argues that banks become guilty of some form of complicity when they help accelerate climate change with their financial support. Shue advances this view by arguing that banks make a direct contribution to climate change, that banks could more easily support morally superior enterprises, including renewable energies, and that banks have the power to shape the future of the next generations (given the decades-long lifetime of infrastructure like pipelines). Paula Casal explores responsibility for harmful forms of behavior that are statistically strongly associated with one sex. She discusses self-harm and harm to others on the assumption that they result from social conditioning and on the assumption that they have evolutionary origins. She also examines behaviors that originate from neglect of, or respect for, our duties to others, including our offspring, and whether or not holding individuals responsible for gendered behaviors has beneficial consequences. The final subsection, entitled Responsibility in the law, addresses legal questions about responsibility and bring together legal and philosophical scholarship on responsibility. This section’s hypothesis is that our understanding of moral responsibility, or responsibility more generally, could be advanced by attending closely to certain aspects about responsibility in 11

Maximilian Kiener

the law. This is because the law has to deal with responsibility in a very specific, social, and institutional setting, rather than merely in metaphysical or abstract terms. Peter Cane criticises views that dismiss the law as generally irrelevant to, or at least as a distorted picture of, morality and thereby also neglect the concepts of legal responsibility in the inquiry of moral responsibility. To oppose such views, Cane argues for an approach that depicts both law and morality in naturalistic and evolutionary terms and he claims that law and morality are ultimately ‘ontologically similar and epistemologically related’. On this basis, Cane concludes that any investigation of moral responsibility must remain incomplete unless it incorporates insights from the law. John Hyman focuses on the voluntary act requirement in the criminal law and argues that its plausibility depends on the empiricist theory that a voluntary act is a movement caused by a volition or act of will. Hyman claims, against this theory, that ‘voluntary action is in fact not a single phenomenon, with a single definition, but a combination of two phenomena that need to be explained separately: action and voluntariness’. Hyman’s support for this view includes general considerations from the philosophy of action as well as an analysis of Libet’s famous neuroscientific study of voluntary action. Antony Duff starts from the common view that criminal responsibility requires not only a conduct element, i.e., the actus reus, but also a mental element, i.e., the mens rea, as expressed by the slogan ‘an act does not make the actor guilty unless the mind is also guilty’. However, as Duff explains, some offenses do not require that the prosecution prove mens rea, and thus they depart from this orthodox view. Duff explains such cases as involving forms of either ‘strict’ liability or ‘strict’ responsibility, and critically discusses what could justify this potentially morally problematic part of the law. Kimberly Kessler Ferzan focuses on responsibility and pretrial detention. Ferzan asks how the state could justify detaining people before their trial, i.e., before their responsibility or guilt is determined. Ferzan discusses the relation between a defendant’s responsibility and the state’s reasons for detention, including detention on the grounds of dangerousness, flight, and obstruction of justice. Ferzan concludes by casting doubt on the moral legitimacy of detentions in cases where people cannot make bail. Jenny Steele focuses on responsibility for others. Steele inquires how legal doctrines about vicarious liability and philosophical debates on vicarious (moral) responsibility can be brought into a fruitful dialogue. Steele primarily focuses on private law and suggests that related legal doctrines about vicarious liability are in fact close to central tenets in our thinking about moral responsibility. Steele’s discussion features the hazy boundaries of individual agency, our moral entanglement with others, and various types of interpersonal relationships. Finally, Alexander Sarch focuses on contexts in which moral and criminal blame diverge. Although Sarch holds that the deviation of the criminal law from morality can be due to merely practical constraints and, for this reason, be also an occasion for regret; he also adds that this need not be the case. In particular, Sarch aims to ‘dispel the idea that the criminal law’s divergences from morality always sacrifice our principles to the messy necessities of practice’ and suggests that ‘criminal law’s divergence from morality can itself be a matter of principle’.

Notes 1 I thank Thomas Douglas, Leonhard Menges, and Marcel van Ackeren for their very helpful comments on an earlier draft of this introduction. 2 See Kamtekar in this volume.

12

Introduction See Irwin in this volume. 3 4 See Hause in this volume. 5 In Hobbes’ words: ‘For the will, being nothing else but the last appetite in deliberating, is as necessarily determined to every particular action as the appetite to which immediately precedes the action is determined’ (Hobbes 2014, Leviathan: Chapter 21). 6 This experiment highlighted that our brains may initiate our voluntary movements and actions before we are in any way aware of having decided to execute them. 7 For discussion, see Skorupski (1999). 8 For different views on the relation between responsibility and obligation, see Shoemaker (2015) and Scanlon (2000). 9 I am indebted to Mounk for the references to Reagan and Blair.

References Aquinas, T. (2010). Summa Theologica: Translated by Fathers of the English Dominican Province. Boston: MobileReference.com. Aristotle. (1959). Politics. London: Cambridge, MA: Heinemann; Harvard University Press. ———. (1961). Aristotle’s Physics (Bison bk. ed.). Lincoln: University of Nebraska Press. ———. (2014). Aristotle’s Ethics: Writings from the Complete Works – Revised Edition. Edited by Jonathan Barnes and Anthony Kenny. Princeton, NJ: Princeton University Press. Baldock, J., Manning, N., & Vickerstaff, S. (2007). Social Policy (3rd ed.). Oxford: Oxford University Press. Barker, V. (2009). The Politics of Imprisonment: How the Democratic Process Shapes the Way America Punishes Offenders. Oxford and New York: Oxford University Press. Bayne, T. (2011). “Libet and the Case for Free Will Scepticism,” in R. Swinburne (ed.), Free Will and Modern Science (pp. 25–46). London: British Academy. Besson, S. (2022). Theories of International Responsibility Law. Cambridge: Cambridge University Press. Cane, P. (2002). Responsibility in Law and Morality. Oxford, UK; Portland, OR: Hart Pub. Cicero, M. T. (1991). De Fato. Los Altos, CA: Packard Humanities Institute. Duff, A. (2009). Answering for Crime: Responsibility and Liability in the Criminal Law. Oxford: Hart. Duff, A., & Green, S. P. (2013). Philosophical Foundations of Criminal Law. Oxford: Oxford University Press. Frankfurt, H. G. (1969). “Alternate Possibilities and Moral Responsibility,” The Journal of Philosophy, 66(23), 829–839. DOI: 10.2307/2023833. Gilovich, T., Griffin, D. W., & Kahneman, D. (2002). Heuristics and Biases: The Psychology of Intuitive Judgement. Cambridge: Cambridge University Press. Hobbes, T. (2014). Leviathan. Oxford: Clarendon Press. Hodgson, D. (2011). “Quantum Physics, Consciousness, and Free Will,” in R. Kane (ed.), The Oxford Handbook of Free Will: Second Edition (pp. 56–83). Oxford: Oxford University Press. Hutchison, K., Mackenzie, C., & Oshana, M. (2018). Social Dimensions of Moral Responsibility. New York: Oxford University Press. Knight, C., & Stemplowska, Z. (2014). Responsibility and Distributive Justice. Oxford: Oxford University Press. Martinich, A. P. (2016). “Thomas Hobbes and John Bramhall,” in K. Timpe, M. Griffith, & N. Levy (eds.), The Routledge Companion to Free Will (pp. 303–312). New York and London: Routledge. Mason, E. (2019). Ways to be Blameworthy: Rightness, Wrongness, and Responsibility. Oxford: Oxford University Press. Mill, J. S. (1874). Nature, the Utility of Religion, and Theism. London: Longmans, Green, Reader, and Dyer. ———. (2015). On Liberty, Utilitarianism, and Other Essays (New ed.). Oxford: Oxford University Press. Mounk, Y. (2017). The Age of Responsibility: Luck, Choice, and the Welfare State. Cambridge, MA: Harvard University Press. Nagel, T. (1979). Mortal Questions. Cambridge: Cambridge University Press.

13

Maximilian Kiener Oberdiek, J. (2018). Philosophical Foundations of the Law of Torts. Oxford: Oxford University Press. Pereboom, D. (2001). Living Without Free Will. Cambridge and New York: Cambridge University Press. Principe, L. M. (2011). The Scientific Revolution: A Very Short Introduction. Oxford: Oxford: Oxford University Press. Przeworski, A., Stokes, S. C., & Manin, B. (1999). Democracy, Accountability, and Representation. Cambridge: Cambridge University Press. Robertson, S. (2020). Nietzsche and Contemporary Ethics. Oxford: Oxford University Press. Scanlon, T. (2000). What We Owe to Each Other. Cambridge, MA and London: Belknap Press of Harvard University Press. Shoemaker, D. (2015). Responsibility from the Margins. Oxford: Oxford University Press. Skorupski, J. (1999). Ethical Explorations. Oxford: Oxford University Press. Smith, A. (2005). “Responsibility for Attitudes: Activity and Passivity in Mental Life” Ethics, 115(2), 236–271. DOI: 10.1086/426957. Strawson, P. F. (1963). Freedom and Resentment. London: The British Academy. Watson, G. (1996). “Two Faces of Responsibility,” Philosophical Topics, 24(2), 227–248. Williams, B. (1985). Ethics and the Limits of Philosophy. Cambridge, MA: Harvard University Press. ———. (1995). Making Sense of Humanity: And Other Philosophical Papers 1982–1993. Cambridge: Cambridge University Press. Willmott, C. (2016). Biological Determinism, Free Will and Moral Responsibility: Insights from Genetics and Neuroscience. Cham, Switzerland: Springer Nature.

14

PART 1

The history of responsibility

Section 1: Responsibility and wrongdoing

1 PLATO ON VICE Marcel van Ackeren

The early Greeks did not have a single (technical) term for what we now call moral responsibility, but they were very interested in blaming and praising people for their actions. Early authors like Homer, Pindar, Hesiod, and the tragedies even foreshadow what we now call the freedom or control condition of responsibility as well as the epistemic condition of responsibility. For instance, Oedipus kills his father and marries his mother not knowing about his parents’ identity and not being free to escape the curse and his fate. Given this background it is hardly surprising that also Plato discussed many questions that are still part of our contemporary debates on responsibility. The first section will discuss the two central theses of Plato’s early dialogues, namely that virtue is knowledge (and that, accordingly, vice is ignorance) and that wrongdoing is involuntary. The second section will elucidate how the middle and late dialogues develop these two claims, which remain central topics, in the light of a new moral psychology and the theory of forms (see Kamtekar 2017). In the third and final section, I will address the question of whether Plato’s claims that vice is caused by ignorance and that it is involuntary imply that vicious people are not responsible for their actions. I will conclude that there two ways to understand Plato and each poses a different challenge to the current debate. Either we take Plato to say that no wrongdoer is responsible because of the ignorance which makes the action involuntary – in this case the challenge lies in very strong relation of the epistemic and control condition in Plato, which amounts to a peculiar form of elimintativism or asymmetry about responsibility – or we take Plato to say that wrongdoers are responsible despite the ignorance and the involuntariness. This would challenge the view that involuntariness and ignorance diminish responsibility.

I  The early dialogues Plato’s views that bear on what we call responsibility are part of an ethical outlook, that is, like many other ancient ethical theories, eudaemonistic. He holds the view that the highest aim of all humans is happiness or well-being (eudaimonia), and his outlook can be considered to be virtue-based insofar as virtue (aretê), that is, the excellence of the soul, is taken to be what we need to achieve this highest aim. Within this general framework, Plato argued for the two central claims of his ethics. 19

DOI: 10.4324/9781003282242-4

Marcel van Ackeren

Claim 1: Virtue is knowledge and vice is ignorance Plato’s first claim that virtue is knowledge and that vice is explained by ignorance is best explained by distinguishing four more specific assumptions by Plato. First, there is knowledge about virtue. Socrates paved the way for what we now label moral cognitivism by claiming that moral judgements can be true or false. In the Apology (20b), Socrates asks: “Who has knowledge (epistêmê) of that kind of excellence (aretê), that of a man and a citizen?” Socrates does not only assume that questions about the things that matter most (ta megista, Ap. 22d), that is, the good life and virtues, can be answered falsely or correctly, but also that the truth-value of these answers is exactly like that of answers regarding other fields of knowledge, such as mathematics or medicine (La. 184e185, Charm. 174–175, Ap. 25d-e). Some scholars have argued that the type of knowledge that Socrates seeks is non-propositional and a form of practical knowing-how (for instance, Gould 1955), but the majority of more recent scholarship sees clear indications that Socrates is looking for propositional knowledge: Socrates’ famous ‘What is X?’-questions are part of his quest for definitions, and in his cross-examination (elenchos), especially in the early dialogues, the views of his interlocutors about virtue are rejected as mistaken or inconsistent.1 Second, virtue itself is knowledge. Socrates proposes a fully-fledged intellectualism by assuming that knowledge is necessary and sufficient for being virtuous (Euthyd. 280a). The relation between knowledge and virtue lies in the assumption that being virtuous and being able to define that virtue, that is, having knowledge about the virtue, cannot be separated (Charm. 189e-190b and La. 189e-190b). To Plato, being virtuous is not only a moral disposition, it is also a certain epistemic state: being able to define a virtue does not only show that one has knowledge about that virtue, it is also indicative of the possession of that virtue. If one is virtuous, one has knowledge about virtue, because only knowledge makes us virtuous. As to the knowledge that makes us virtuous, we need to discuss two features, namely that knowledge (i) leads to a unity of virtue(s) and (ii) what the objects of that knowledge are. As to the unity (i), the cross-examination of the beliefs about one single virtue often makes inconsistencies obvious: an interlocutor makes assumptions about one single virtue at odds with also held views about another virtue. This provides us with two lessons: a single virtue cannot be defined without knowing all virtues and the definition is not a oneline verbal statement, which can be remembered or parroted without understanding or in isolation from the definition of another virtue. What kind of unity does Plato have in mind? One line of interpretation, spearheaded by Vlastos (1972), argues that the unity is a unity of virtues (plural) because each of the virtues is a distinct part of a whole, so that each virtue requires a separate definition. Since Penner (1973), an opposing view holds that the unity is a unity of virtue (singular) because the virtues are based on one overarching knowledge, not separable definitions. Devereux (1982) exemplifies a third line of interpretation, assuming that Socrates in fact argues for both accounts, namely for the unity of virtue in the Protagoras and for a unity of virtues in the Meno and Laches. In all likelihood, Plato argued that the different virtues and their specific definitions of, for example, what is to be feared in the case of courage must be rooted in a more general knowledge regarding what is good (see La. 199c-e). 20

Plato on vice

As to the objects of knowledge (ii), Socrates explains in the Euthyphro (6d-e) that his questions are geared towards finding a single Form (eidos) that causes the essential features of a virtuous act (Mn. 70a-71b). The Forms also have a normative and action-guiding aspect, as they are paradigms (paradeigmata, Euthyph. 6e) explaining why it is better to have a certain virtue than not to have it. Now, what is the relation of the unity (i) and the objects of knowledge, the Forms (ii)? The notion of such a Form (eidos, idea) must be part of any definition of virtue, which is why definitions in the form of a list of virtuous actions fails to account for the conceptual unity of the various actions following from that virtue. Third, virtue is a craft. Plato explains virtue in terms of craft or competence (technê, a term also used as a general synonym for knowledge), which supports the view that knowledge and definitions are propositional but require more than a simple one-sentence explanation. The majority of anglophone literature agrees on the importance of craft analogies in Plato’s concept of virtue (see Irwin 1995, ch. 5; for doubts see Vlastos 1991; Roochnik 1996 denies deny the importance of technê as vehemently as some continental interpreters bordering Heideggerian technophobia). Craft analogies are very frequent in Plato, but they are not his invention. Those whom he ridiculed as sophists and also his great local competitor Isocrates used such analogies as well when discussing virtue (compare Isocrates orat. 15, 209 with Ap. 20a-b). However, Plato developed an original and specific concept of technê, which is outlined in the Ion and the Gorgias and is the basis for his intellectualism regarding virtue. Against the sophistic polymaths, he argued for a strict and exclusive correspondence between a craft and an object or task (ergon), so that each object and task require a specific craft and each craft deals with only one object, yet also necessarily with the whole of it and not just a single part. An expert is an expert not only regarding a well-defined field, but necessarily the entire field.2 This has implications for the claim that virtue is knowledge. Expertise about virtue must be knowledge about the entire field of virtue. That is why Socrates demands an expert on virtue to define more than one single virtue and speak consistently about all of them. The definition of a single virtue must be imbedded in knowledge about virtue as such. With this, inconsistencies are supposed to be excluded and the ability to succeed in a Socratic cross-examination of beliefs secured. The knowledge of a craft is also essentially normative, as it implies the knowledge of the evaluative standards for the object in question (see Ion 530c-532c, Theaet. 178b) as well as all appropriate actions in this field (Ion 540b). The knowledge of a craft is further distinct from mere experience (empeiria, Gorg. 463a-c, 464c), as it is also knowledge about the nature (physis) and causal principles (aitia) of that field, it is mathematically precise, involves calculation (diarithmein, Gorg. 500e-5001e) and measurement (metretikê, Prot. 352–358): Plato foreshadows some important aspects of Aristotle’s concepts of virtue and justice by introducing the Pythagorean distinction between arithmetical and geometrical, that is, proportional, equality by Archytas into ethics (Gorg. 508a). Fourth, virtue precludes akrasia (and guarantees control). By claiming that knowledge is necessary and sufficient for virtue, Socrates implies that all other factors, such as majority votes, tradition, pleasure, power, and so on, are irrelevant when it comes to virtuous action (see Crit. 47a-48a). Pleasure is of special importance to Plato, for he is keen to show that desires and feelings cannot influence the relation between knowledge and virtue, respectively ignorance and vice. Plato devotes one of his most elaborate arguments in the early dialogues (Prot. 352–358) to argue against what is  – since book VII of Aristotle’s 21

Marcel van Ackeren

Nicomachean Ethics – known as akrasia, which literally means lacking command or power. Socrates denies that wrongdoing can be explained by assuming that what we believe to be good can be overcome by pleasure or pain and hope or fear. Thus, the argument does not regard the weakness of will; it rather concerns wrongdoing and the alleged weakness of beliefs and aims at proving that ignorance is necessary and sufficient for failure. An influential view by Vlastos (1956, 1972) and Penner and Rowe (1994) holds that the argument is not about denying belief–akrasia, but only knowledge–akrasia. Yet, Socrates is aware of the difference and includes both (see Prot. 357b-c, 358c-d). Even if the argument is about belief–akrasia, it is not a step back from the intellectualistic position regarding knowledge, as the argument concludes that if our beliefs fully determine our course of actions, then the explanation for the difference between virtue and vice must lie in a difference between different cognitions: knowledge and ignorance. As the argument is based on a hedonistic premise, Socrates concludes that the art of measuring the outcome of an action, in terms of pleasure and pain, secures knowledge. Socrates also provides an error theory, according to which ignorance, the wrong judgment about the balance of pleasure and pain, comes from overestimating the amount of pleasure and pain when it is immediate and underestimating it when it is in the more distant future. From the early dialogues on (see La. 192, and especially the discussion of pleasure with Callicles in the Gorgias), Plato might have felt the need to develop a more complex moral psychology, acknowledging the impact of affections on our decisions and actions. In sum, Plato’s first central claim that virtue is knowledge and that vice is ignorance is expressive of a strong moral cognitivism and intellectualisms. Now, let us turn to the second central claim about the involuntariness of wrongdoing.

Claim 2: Wrongdoing is involuntary Plato’s second central claim is the Socratic paradox,3 i.e., the view that no one does wrong willingly. It is discussed in two dialogues, each posing its own exegetical questions, but the main problem is to find a unifying interpretation. In the Gorgias (466a-8e), Socrates argues that if we choose to do something and if this action turns out to be bad, that is, worse than other possible alternatives of action in terms of benefit, it follows that we acted involuntarily and lacked the power (dynamis) to control our actions: we did not do what we wished to do and did only what seemed best to us (see Penner 1991). The underlying assumption is that we all wish (boulesthai) for that which is really good and really contributes to our well-being and that all of our actions are geared towards this end. Plato’s concept of what we wish (boulesthai) might seem strange, as he views it to be non-contingent and invariably fixed on what is really good and distinguishes it from doing what seems best and from acting according to our desire (epithymia). In the Meno (77B-78B), Socrates, again, claims that everyone desires the good and no one desires what is bad. According to the influential interpretation by Santas (1979, ch. 6; McTighe 1984 or Weiss 1985), Socrates’ claim here is that one can indeed want things that are really bad, provided only that one believes them to be good. This reading of the Meno gained further attractiveness, for it is the way in which Aristotle (NE III, 4 and Topics VI 8.146b36–147a11) reconstructs Socrates’ position and it can also be couched in terminology by Anscombe that was picked up by Davidson: we can want things that are really bad provided we want them under the description of a ‘good thing’. 22

Plato on vice

Prima facie, we thus have two ways of understanding the Socratic paradox that no one does wrong willingly, one based on Gorgias, saying that no one wants what is really bad and the other based on the Meno, holding that people are capable of wanting things that are really bad if they believe them to be good. There is, however, a unifying interpretation which sees a fundamental consistency between the two dialogues (Penner & Rowe 1994, based on Bluck 1961: 257): the claim in the Meno that we desire apparent goods can be reduced to the stronger position of the Gorgias, which is only seemingly different and in conflict with it: when we desire things under the description of a “good thing” (Meno), we indeed desire what is really good (Gorgias), for if we knew that these things were be really bad, we would not wish for or pursue them. In order to appreciate this solution, it is important to note that the Gorgias distinguishes between wishing (boulesthai) to do something that is really good and the desire (epithymia) to do something which is in fact bad but only seems to be good to us. This interpretation has advantages, but also serious implications. The advantage is that the interpretation is in alliance with Plato’s account of language (Crat. 387a): when we say that we want to do something, we really want to do that and not something else: for example, when we want to feed ourselves, to eat some food, we really want to eat food and not only appear to eat or eat what only appears to be food. According to Plato, this also holds true for virtue and goodness: “when it comes to the good nobody is content with the possession of the appearance but all men seek the reality, and the semblance satisfies nobody here” (Resp. 505d-e). The implication of this interpretation of the Socratic claim that no one errs willingly leads us to an important aspect of Plato’s moral psychology. The denial of akrasia in the Protagoras was based on a hedonistic premise and psychological eudaimonism, but, as Rachana Kamtekar pointed out (2017, ch. 3), psychological hedonism is not sufficient for the stronger claim of the Gorgias that we all wish to do what is really good and therefore do wrong unwillingly. Kamtekar argues the Socratic claim is based on a conception of what we naturally whish. We all wish what is our natural advantage and therefore really good. This interpretation also fits with what the Symposion has to say about eros, a desire that is ultimately directed to the Forms, which are entities that are really good. This underlines that the Gorgias is a transitional dialogue from the early dialogues towards the middle ones, a transition in which Plato sets forth an ambitious theory of the ontological status of what is good and what our desire for the good is based on. In sum, the early dialogues draw a picture of responsibility in which matters of knowledge and ignorance on the one hand and involuntariness on the other are tied together: the difference between the virtuous and the vicious is not a difference in what those who act virtuously and those who act viciously want; the difference lies only in the intellects, in the epistemic status of these people, rather than their characters or dispositions. Knowing what is good enables people to act virtuously and reach what they want, namely that which is really good. The vicious, on the other hand, have false beliefs about what is really good and hence mistakenly choose vicious activity. Although all humans invariantly desire that which is really good, they need knowledge. Having the wish alone does not imply knowledge about the object of that wish or the skill or competence needed to choose the right means to fulfil it. In section 3 I will discuss in what way this can be read as a challenge to mainstream views of the contemporary debate on responsibility. 23

Marcel van Ackeren

II  The development of the middle and late dialogues With the so-called middle dialogues and especially the Republic, Plato’s account of virtue as knowledge, vice as ignorance and the involuntariness of wrongdoing significantly develops, becoming inseparable from a new moral psychology and his theory of Forms. In addition, the ethical views also become embedded in a political context. Plato starts fresh in the Republic by introducing a tripartite division, which is based on the principle that the same thing cannot do or undergo contraries regarding the same aspect of itself (Resp. 436b, see Stalley 1975). Plato describes the different capacities or parts of the soul as different subjects or even different agents having different motivations, beliefs, and desires.4 Socrates distinguishes the rational part or capacity (logistikon), which reasons and desires all knowledge, including judgements about what is good and bad, from the purely appetitive element (epithymetikon), which does not go by one name, because it is the source of blind desires that can be about anything. The spirited part (thymoeides) is Plato’s invention that allows for a new and highly dynamic moral psychology: unlike the desires of the purely appetitive part, those of the spirited part are evaluative insofar as they rest on and are responsive to judgements about goodness and badness. Anger, for example, is viewed as a strong motivational force that is supposed to help the reasonable part by being angry over injustices and providing courage when it comes to suppressive blind desires that fear that which is not bad. Socrates then discusses four virtues that later become the cardinal virtues: wisdom (sophia) is the virtue of the rational part, which is concerned with all knowledge. It is facilitating control over the other parts and one’s actions (Resp. 442c). Courage (andreia) is the virtue of the spirited part, it is about anger and other emotions which are virtues if the part follows the knowledge of the rational part (Resp. 442b). Unlike the first two virtues, temperance (sophrosyne) is not the virtue of one specific part, but describes the harmony of the entire soul, which ensures the ruling of the rational part, with the help of the spirited part, over the appetitive part (Resp. 431e-432a). There can be no virtue of the appetitive part, because this would imply that the pure desire can relate to them themselves or even control themselves. Justice (dikaiosyne) is not simply a further virtue, but the virtue necessary and sufficient for the other virtues and the harmony of the soul as well as the unity of the virtues, because justice is defined as the power that enables each part to do its own work (to ta hautou prattein, Resp. 441d, 444c-d). But although Plato’s new and more complex moral psychology tempted some scholars to assume that Plato, contrary to the Protagoras, now accepts that there is akrasia (for a discussion see Shields 2007), there can be little doubt that his ethics remain intellectualistic. The new account of the soul and virtues merely shows that the relation of virtue and knowledge is no longer merely a matter of pure cognitive training, resulting only in wisdom, which is now the only purely intellectual virtue. Plato continues to claim that virtue is knowledge, which now depends on the right ordering or harmony of the soul. Accordingly, vice is considered to be a result of the soul being in conflict (stasis), which is caused by an ignorant and dominant appetitive part. Accordingly, Plato now views wrongdoing as the effect of an unhealthy soul that is in turmoil because the blind desires rule instead of knowledge. From the end of book V on, Socrates develops a new account of knowledge that is meant to guarantee an orderly ruled society and soul and therefore right individual and collective 24

Plato on vice

action. The claim that philosophers should rule leads to the question of what knowledge is and how it can be distinguished from belief. Thus, Socrates starts to develop an epistemology intertwined with ontology. He distinguishes (Resp. 474b-480a) three kinds of cognitive states, namely ignorance, belief, and knowledge. Each state is located in one part of the soul, namely the appetitive, spirited, and rational part; and each concerns a different ontological class, namely non-being, becoming, and the pure being of the Forms. The correspondence of epistemology and ontology is further elaborated in the similes of the Sun, Line, and Cave (Resp. 508a-518b): Socrates introduces the ultimate object of the master science dialectics (dialektikê technê), the Form of the Good. This Form is the cause of all being and all goodness (Resp. 517c) and thus the ultimate object of knowledge that guarantees successful virtuous acting and well-being for individual souls and the society ruled by that knowledge (517c-519a). In sum, Plato maintains a version of moral realism of a very special kind, because that which is normative is more real than the physical world and transcendental. Plato’s theory of Forms in the middle dialogues does not only provide a new underpinning for the relation of ignorance and wrongdoing, but also for the claim that all humans strive for what is really good. According to the theory of Forms, only Forms are true beings and a completely good, that is, full representation of norms (paradeigmata). In the Phaedo (104–105), the soul is said to differ from the body and physical objects in general because it is (more) like the Forms (or even one), and this affinity also explains why the soul strives for what is really good. Two other dialogues discuss this relation of the soul, actions, the Forms and the really good under the heading of love (eros): the Symposion presents a ladder of love, scala amoris, according to which that what we ultimately love, the Form of the beautiful and good, is a contribution to our well-being and the motivation for all of our actions. All humans long for self-completion, which explains their actions, and the intermediate steps are viewed as lower and manifold manifestations of the single Form of the beautiful and the good, for example, in objects, human bodies or souls. In the later Phaedrus, Plato connects the idea of a self-completing love with three main themes of the Republic: first, with the idea of the tripartite soul, which is, in the Phaedrus, described as a charioteer (the rational part) with two winged horses, an obedient white one (the spirted part) and an unruly black one (the appetitive part). Second, the acquisition of knowledge and ignorance resulting in wrongdoing is now more clearly explained as being caused by desires: the bad horse is the reason why the charioteer (the rational part) struggles to ascend to and grasp the Forms; if, however, he has already done so, the desires, the bad horse, cause him to fall from the acquired intellectual level and veil his knowledge. Thirdly, the Phaedrus reveals Plato’s increasing interest in a specification of dialectics, the master science, as methodological knowledge, which he pursues in many late dialogues: knowledge understands language and concepts in terms of the Forms, allowing the unification of the plurality of perceptions via reason (Phaed. 249b-c). Dihaeretic knowledge influences taxonomy by distinguishing between concepts and sub-concepts until a definition is reached. In the Timaeus, a very late dialogue mainly concerned with cosmology, Plato extends the debate regarding the causes of viciousness. In previous dialogues, he had discussed the vice of the soul by assuming that it is analogous to diseases of the body (see Prot. 323c-d; Resp. 404e, 407c, 556e, 609c-d). Now, he additionally seems to suggest that turbulent and unhealthy physical movements of the body can lead to vice (Tm. 86b; see Sassi 2013). Again, this new explanation does not mean that Plato gave up his intellectualistic approach to vice. These physical factors, such as an uncorrected upbringing or bad 25

Marcel van Ackeren

influences in a society (Tm. 86d), do not directly cause vice; they lead to vice when they obscure the intellect (see Renaut 2019). In sum, the main changes in the middle and late dialogues are no departure from the claims of the earlier dialogues, but Plato developed substantially new theories providing a new basis for his intellectualistic view about virtue and vice.

III  Ignorance, involuntariness, and responsibility Let me now consider the implications of Plato’s views for the contemporary debate on responsibility. Does Plato’s claim that vice is caused by ignorance and is therefore involuntary imply that the vicious are not responsible for their actions? Clear-cut answers are very difficult, because the concept of responsibility is a later one. Also, there is a “welter” of distinguishable meanings of responsibility, as Hart famously observed – and that was more than 50 years ago, that is before a flood of new literature and theories. Thus, the remaining section cannot replace a substantial study on Plato’s relation to the current literature on responsibility. We have, very roughly, two options to answer and each answer poses a different challenge to the contemporary debate. According to the first answer, Plato regards no wrongdoer to be responsible because wrongdoing is always based on ignorance and involuntary. According to the second answer, Plato thinks wrongdoers are responsible despite acting involuntarily. The first answer, holding that, according to Plato, the vicious are not responsible, has been argued for by McTighe (1984): what makes Socrates’ views in the Gorgias (466–468) and the Meno (77–78) a paradox is the implication that all involuntary wrongdoing is exempt from blame because of the involuntariness and that any state of ignorance needs to be considered involuntary. In a similar vein, Mackenzie (1981) has discussed the legal aspects and argues that, according to Plato, a criminal is not the agent of crimes, but more like a passive victim of a bad constitution or a disease or the failure of others to rectify this disposition by education. According to Mackenzie, Plato maintains that wrongdoers should not be blamed, but rather pitied and benevolently (or paternalistically) improved (Mackenzie 1981, 156, see also Kamtekar 2017, 106–107 and critically Meyer 2021, 232–233). Another route to the same conclusion might draw on Susan Wolf’s view that there is a moral competence condition (Wolf 1987), according to which responsibility presupposes an ability to recognize and act according to moral considerations. This seems to fit Plato’s view that the knowledge leading to virtue is a technê, a competence, craft, or skill and the claim in the Gorgias that the ignorant and vicious lack the power to do what they wish, namely that which is really good. Thus, Plato might be taken to express the view that wrongdoers are not responsible because of their lack of decisive moral competence, which is based on lack of moral knowledge, making it unreasonable to expect virtuous behavior. Without moral knowledge, the wrongdoer lacks the ‘ability to do otherwise’ (Wolf 1980) and therefore is not responsible.5 As we will see when discussing the other answer, there is another way to draw on Susan Wolf, but it will reach the opposite conclusion, but first we need to conclude that those who accept the first answer provide a certain challenge to the current debate. By choosing this answer we imply that there might be two differences between Plato and the current debate, which, for a very long time, was focused on the control or freedom condition, but then the epistemic condition gained more and more attention. As recent literature (see Robichaud & 26

Plato on vice

Wieland 2017: 3) exemplifies, the debate still tends to discuss the two conditions separately. Not so Plato, who argues for a very strong relation: to him the involuntariness of wrongdoing, that is the lack of control or freedom to do what one wishes to do, depend on the epistemic condition, that is ignorance. According to this interpretation of Plato, the two conditions are causally related, so that both conditions are fulfilled or none: Knowledge not only guarantees control, but is also the only way to obtain control. Thus, a person can have relevant control if and only if they have relevant knowledge. Now, assuming such a strong relation between the epistemic condition and the control condition in Plato is not the defining feature of the first answer, because the second answer, assuming that wrongdoers are responsible despite ignorance and involuntariness and that I will discuss shortly, might be based on that relation too. What makes the core of the first answer is the claim that this relation is so strong that the ignorance leading to involuntariness makes all wrongdoers in each and every case of wrongdoing not responsible. In short, the first answer holds a peculiar and implausible form of eliminativism or asymmetry about responsibility. Though the first answer can rely on powerful and widespread intuitions according to which ignorance and involuntariness can diminish responsibility, it assumes an extreme form of these intuitions, leading to a striking asymmetry with regard to responsibility: all virtuous persons are always and fully responsible for their virtuous acts and there are no cases of responsible wrongdoing, because no wrongdoer is ever responsible because of the ignorance and involuntariness. From the viewpoint of the modern debate, it can be doubted whether it makes sense to assume that the concept of responsible wrongdoing is really without any instantiation and therefore no wrongdoer is blameworthy. Note that, even on this interpretation, Plato’s position would not be a form of hard incompatibilism because it only rejects responsibility for wrongful conduct but not responsibility for virtuous conduct. In other words, according to the first answer Plato does not hold fully-fledged elemintivism about responsibility, but only a peculiar asymmetrical version of it,6 Now, let us turn to the second answer which wants us to accept the idea that Plato indeed viewed wrongdoers to be responsible for their actions, despite the involuntariness. In this case Plato challenges the modern debate by assuming that those who go wrong because of ignorance and also involuntarily remain responsible. First, Plato always speaks of the wrongdoers’ acts as theirs. Mackenzie (see previous) argued that wrongdoers are more like passive victims of their ignorance and insanity, but Plato, whenever he discusses the punishment of vicious persons by courts or in the afterlife, regards them to be the agent of the deeds in question, thus these passages do not support the view that the vicious cannot be responsible agents: according to the Gorgias and then the Laws, wrongdoers are blameworthy and responsible in a strict legal and moral sense and need to be punished without any excuse or exemption (see Weiss 1985).7 Also, in some passages it is explicitly said that ignorance and wrongdoing is blameworthy, for instance, Ap. 29d-30a: “If I find him that he does not possess virtue, but says he does, I shall rebuke (oneidio) him for scorning the things that are of most importance and caring more for what is of less worth.” Plato thinks that vice is ignorance and involuntary, but also that the vicious are nonetheless blameworthy and responsible. If that is the case, Plato poses a challenge to a currently widely held view, namely that ignorance and involuntariness undermine responsibility. Second, there is a defense of this view in Plato. Let us go back to the passage in Gorgias (466–468): when Socrates says that we do not want to act viciously, he only talks about 27

Marcel van Ackeren

our natural wish (boulesthai), and this special sense of involuntariness does amount to an excuse or an exemption because in this case we choose to act according to what seems best to us and what we desire, even when we act wrongly (for a recent study, see Kamtekar, 2006). This special sense of involuntariness of wrongdoing does not imply the view that wrongdoers have no capacity to choose at all. So, the second answer differs from the first, by assuming that wrongdoers indeed have some “ability to do otherwise” (Wolf). To Susan Wolf, the incapacity to act morally diminishes responsibility only if it implies an incapacity to choose. To Plato, this is not the case, and we remain capable of choosing, and thus remain responsible because even though wrongdoing is involuntary and based on ignorance, Plato maintains that we can change. Thus, Socrates (Ap. 29e-30a, quoted previously) can rebuke somebody for caring about the wrong kind of goods and for not knowing what ones does not know about virtue and moral goods. Plato seems have in mind what Susan Wolf calls “deep responsibility” (1990: 37).8 Wolf’s ‘Deep Self View’ holds that “in order to be responsible for one’s actions, one must be responsible for the self that performs these actions” (Wolf 1987: 375). Contrary to Wolf, who thinks that we are not completely responsible for our deep selves, Plato seems to assume that we are indeed responsible in that sense. Such responsibility is deeper, because it concerns two second-order aspects. One the one hand, there is the question of what kind of desires and character we would like to have. According to Plato, we are responsible for the soul or character we have, because we can influence our volitional structure, for example, by becoming less desiredriven and more temperate. On the other hand, it concerns epistemic aspects. Socratic cross-examinations target secondorder ignorance, that is ignorance about ignorance: there are cases in which the agent does not know that they do not have the relevant moral knowledge. Plato thereby stresses that acting ignorantly is blameworthy, for you do not know that you do not know, but you are responsible for acting based on this double ignorance. For if you can know that you are not a moral expert, it is reprehensible to act as if you were. To Plato this ignorance is a form of negligence, or recklessness, and therefore makes one culpable (for a modern take on this see King 2017). And if we are responsible for the ignorance that makes us vicious, we are responsible for the wrongdoing, even though it is involuntary in the sense of not being according to our real and natural wishes. However, even deep responsibility in Plato has its limits. According to the myth of Er which concludes the Republic (X, 614–621) with an account of our afterlife, our character is not only a matter of choice, but also luck.

Acknowledgements I am very grateful to Maximilian Kiener for many insightful and helpful comments. Also, I would like to thank the participants of a workshop for their questions and comments, especially Mark Alznauer, Rachana Kamtekar, Maximilian Kiener, Thomas Pink, and Michael Schefczyk. Roger Crisp also made very valuable comments.

Notes 1 For example, the Laches discusses the nature and definition of courage, the Euthyphro piety, the Charmides moderation, the Greater Hippias beauty. The Protagoras is about the unity of virtue and the Meno raises questions about the acquisition of virtue.

28

Plato on vice 2 For example, an expert in mathematics must know all four basic calculating operations, not just one or two. 3 Claims 1 and 2 are often called Socratic paradoxes in the literature because they are literally paradoxical, that is “against widely shared views.” 4 For an account of the partition and the parts, see Stalley 2007; Moss 2007; Kamtekar 2017, ch. 4 and 5. 5 Note that Plato, at the end of the Lesser Hippias, concludes that only a virtuous person would be able to willingly act unjustly. Likely, Socrates means to say that the virtuous, theoretically, are able to do otherwise, but practically always will choose the virtuous course of action. 6 I am very grateful to Roger Crisp, who pointed this out to me. 7 This, however, leads to a vexing question that troubles interpreters: why does Plato think wrongdoers need to be punished rather than educated, or is the punishment something like a necessary intermediate step before education? 8 By drawing on a concept by Wolf here, I do not want to claim that Plato’s and Wolf’s positions are identical or close. I just make use of the idea that there is another level of responsibility.

Further reading Irwin, T. (1995). Plato’s Ethics. New York: Oxford University Press is a landmark publication on the development of Plato’s ethics from the early Socratic dialogues to the Republic. Kamtekar, R. (2017) Plato’s Moral Psychology. New York: Oxford University Press chapter 3, is an up to date discussion on the unwillingniss of wrongdoing. Meyer, S. S. (2021). “Involuntary Wrongdoing and Responsibility in Plato,” Philosophy and Phenomenological Research, 103, 228–233 provides a critical look on Kamtekar. Pangle, L. S. (2009). “Moral and Criminal Responsibility in Plato’s ‘Laws’,” The American Political Science Review, 103(3), 456–473 focuses on Plato’s last work.

References Bluck, R. S. (1961). Plato’s Meno. Cambridge: Cambridge University Press. Devereux, D. T. (1992). “The Unity of the Virtues in Plato’s Protagoras and Laches,” Philosophical Review, 101, 765–789. Gould, J. (1955). The Development of Plato’s Ethics. Cambridge: Cambridge University Press. Irwin, T. (1995). Plato’s Ethics. Oxford: Oxford University Press. Kamtekar, R. (2006). “Plato on the Attribution of Conative Attitudes,” Archiv für Geschichte der Philosophie, 88, 127–162. ———. (2017). Plato’s Moral Psychology. New York: Oxford University Press. King, M. (2017). “Tracing the Epistemic Condition,” in P. Robichaud & J. W. Wiland (eds.), Responsibility: The Epistemic Condition. Oxford: Oxford University Press. Mackenzie, M. M. (1981). Plato on Punishment. Berkeley: University of California Press. McTighe, K. (1984). “Socrates on Desire for the Good and the Involuntariness of Wrongdoing: Gorgias 466a-468e,” Phronesis, 29, 193–236. Meyer, S. S. (2021). “Involuntary Wrondoing and Responsibilty in Plato,” Philosophy and Phenomenological Research, 228–233. Moss, J. (2007). “Appearances and Calculations: Plato’s Division of the Soul,” Oxford Studies in Ancient Philosophy, 34, 63–89. Penner, T. (1973). “The Unity of Virtue,” Philosophical Review, 80, 35–68. ———. (1991). “Desire and Power in Socrates: The Argument in ‘Gorgias’ 466A-468E That Orators and Tyrants Have No Power in the City,” Apeiron, 24, 147–202. Penner, T., & Rowe, C. J. (1994). “The Desire for Good: Is the ‘Meno’ Inconsistent with the ‘Gorgias’?” Phronesis, 39, 1–25. Renaut, O. (2019). “The Analogy Between Vice and Disease from the Republic to the Timaeus,” in L. Pitteloud & E. Keeling (eds.), Psychology and Ontology in Plato. Heidelberg: Springer. Roochnik, D. (1996). Of Art and Wisdom. Plato’s Understanding of Techne. University Park, Pennsylvania: Pennsylvania State University Press.

29

Marcel van Ackeren Santas, G. (1979). Socrates. London: Routledge. Sassi, M. M. (2013). “Mental Illness, Moral Error, and Responsibilty in Late Plato,” in W. V. Harris (ed.), Mental Disorders in the Classical World (pp. 413–426). Leiden and Boston: Brill. Shields, C. (2007). “Unified Agency and Akrasia in Platos Republic,” in C. Bobonich & P. Destrée (eds.), Akrasia in Greek Philosophy (pp. 61–86). Boston: Brill. Stalley, R. (1975). “Plato’s Argument for the Division of the Reasoning and Appetitive Elements within the Soul,” Phronesis, 20, 110–128. ———. (2007). “Persuasion and the Tripartite Soul in Plato’s Republic,” Oxford Studies in Ancient Philosophy, 32, 63–89. Vlastos, G. (1956). “Introduction,” in G. Vlastos (ed.), Plato’s Protagoras. New York: Liberal Arts Press. ———. (1972). “The Unity of the Virtues in the Protagoras,” Review of Metaphysics, 25, 415–458. ———. (1991). Socrates. Ironist and Moral Philosopher. Cambridge: Cambridge University Press. Weiss, R. (1985). “Ignorance, Involuntariness and Innocence: A Reply to McTighe,” Phronesis, 30, 314–322. Wolf, S. (1980). “Assymetrical Freedom,” The Journal of Philosophy, 77, 151–166. ———. (1987). “Sanity and the Metaphysics of Responsibility,” in F. Schoeman (ed.), Responsibility, Character, and the Emotions: New Essays in Moral Psychology (pp. 372–387). Cambridge: Cambridge University Press. ———. (1990). Freedom Within Reason. New York: Oxford University Press.

30

2 HEGEL ON GUILT Mark Alznauer

Hegel appears to conceive of responsibility (Schuld) in a traditional way. He thinks someone is responsible (schuldig) for what they do if they are a fitting target of praise or blame – though as we will see his primary interest is in those cases where someone has violated the rights of someone else, and so is also a fitting target of some kind of punishment.1 He accepts the commonsense view that adult humans are typically responsible in this sense, and that animals, children and so forth are more or less innocent (unschuldig), having absent or diminished responsibility. He even seems to agree with the current philosophical consensus that responsibility involves both a control condition, which is satisfied if what happened was subject to the agent’s discretion, and an epistemic condition, which is satisfied if the agent had the right kind of awareness of what she was doing or allowing to happen. He characterizes something that satisfies the control condition as the deed (Tat), and something that satisfies both the control and epistemic conditions as action (Handlung). His view, at least as stated in the “Morality” section of the Philosophy of Right, is that agent is only fully responsible or accountable for her actions: for what she has brought about by her activity under the description that she knowingly willed it. Hegel diverges from the tradition, however, in his understanding of the main philosophical threat to responsibility. In his own time, and even until very recently, most attention to the problem of responsibility focused on the control condition, which was regarded as the more suspect of the two.2 It was thought that in order for agents to have right sort of control over their actions, they need to be free to have act otherwise, but that this sort of freedom is at odds with a modern, scientific view of the world, which appears to be committed to deterministic explanations of all natural phenomena including human behavior. But although Hegel briefly alludes to the traditional free will problem early in the Philosophy of Right (dismissing it as resting on a defective conception of freedom), his own treatment of responsibility centers on a problem generated by the epistemic condition, a problem he thinks threatens the foundations of modern social life. He worries that our characteristically modern conviction that individuals can only be rightly held accountable for what they have done knowingly and willingly creates a potential breach between the subjective realm of insight and motive and the objectivity of the external deed, jeopardizing our capacity to hold each other responsible for rights violations.3 The main problem he is 31

DOI: 10.4324/9781003282242-5

Mark Alznauer

concerned with in his own defense of responsibility is not finding room for counter-causal freedom within nature, but reconciling the agent’s right to reject responsibility for any aspect of her deed that was not “inwardly present” to her with the political imperative that all violations of right be punished.4 Hegel’s solution to this problem is quite provocative; he argues that we must reject the formal and individualistic conceptions of agency that are characteristic of the modern natural right tradition in favor of an intrinsically social conception of agency. This means giving up the idea the inner will of the agent is a merely private, subjective matter – something she has privileged, indefeasible access to – in favor of a conception of the inner will as something that must prove itself by becoming expressed in external reality, where it is legible to others who share the same form of life. On his view, the agent’s immediate consciousness of her intention or motive is downgraded to a mere appearance: a provisional take on what she has willed. It must be tested against what she accomplishes, which may reveal aspects of her volition of which she was not immediately cognizant.5 Hegel’s claim is that only a social re-conception of agency can fully close the gulf between the subjective will and the objective deed, making possible intersubjectively valid determinations of what the agent is responsible for. Although it is generally accepted that Hegel’s reflections on agency move in this direction, there is little consensus about the specifics of his position or how he argues for it. Many of these disagreements can be traced to differing ways of accommodating Hegel’s commitment to a dialectical mode of procedure.6 One of the chief features of the dialectical method is that the very terms or concepts it deals with are not regarded as fixed points settled by definition at the beginning of the inquiry, but as fluid moments that change over the course of the investigation. To disambiguate an argument of this form requires that we carefully differentiate the defective form that the target concept initially takes from the revised and improved form it takes at the end of the inquiry, and that we provide some account why it needs to be reconceptualized along these lines. To see how an argument like this works, it is worth beginning with a short discussion of the logical background of Hegel’s theory of responsibility.

I  The logical background For Hegel, the characteristically modern insistence on the primacy of the inner will in determining responsibility is an instance of a general tendency to wrongly oppose the inner essence of something to its outer appearances, privileging the former over the latter. Hegel offers a general treatment of the inner-outer problem in his logic, which has come down to us in two forms, the shorter Encyclopedia Logic and the longer Science of Logic.7 The central ambition of Hegel’s logic is to provide a complete derivation of our most basic categories, and the relationship of inner and outer is one of these categories. But what it is perhaps most distinctive about Hegel’s way of deriving the categories is that it is accompanied by a critique of these categories as they are standardly used. For Hegel, the “usual mistake of reflection” with regard to the categories of inner and outer is regarding the essence of a thing as a kind of inner nature that may or may not reveal itself in outer existence (EL §140 R). On such an account, the inner is regarded as having a different content than the outer, a content that might entirely fail to be expressed in the outer, and it is also regarded as having priority over the outer: when they conflict, the inner is the truth of the matter, and the outer is mere appearance.

32

Hegel on guilt

Hegel is only targeting a certain specifically philosophical way of opposing inner to outer. He does not deny that in many ordinary cases the inner nature of a phenomena might very well fail to make itself manifest externally; sometimes well-laid plans do indeed come to nothing and good seed falls on unsuitable ground. Hegel’s concern is exclusively with attempts to take this everyday contrast to mark a “hard and fast opposition” between inner and outer (EL §142 A). One place he sees inner and outer becoming opposed in this inflexible way is when it is asserted that the inner side of nature is intrinsically unknowable, having an inner essence that is independent enough from nature as it appears to us that it might entirely fail to manifest itself. When we use the opposition of inner and outer to mark a putatively absolute or fixed distinction, we fail to keep in mind that the only reason we have for regarding something as the inner essence of some phenomena in the first place is that, under the right circumstances, it manifests itself externally. Once this is noted, we can see that the very idea of an inner that never manifests itself, or which has content that is completely independent of its external expressions, is strictly incoherent – there could be no such thing. An inner essence that never reveals itself externally would be like a cause that never has any effects. Hegel concludes that the idea of such an inner is self-contradictory. He illustrates his claim with a provocative analogy. Although a good poet might write a bad poem in some given instance, her inner nature remaining hidden from view, Hegel thinks the only grounds we could have for regarding someone as a good poet at all is that they have shown some external signs of talent. His point is not merely epistemic; he is not just saying that by writing good poetry someone makes it possible for others to judge that she is a good poet. His claim is that writing good poetry (at least sometimes, or under the right conditions) is precisely what makes someone a good poet. It follows that if an unaccomplished poet demands that he should be judged based on his high ideals and intentions rather than on his unimpressive output, his demand is “rightly dismissed as empty and unjustified” (EL §140 A). Consistently writing bad poetry is not a failure to give evidence of being a good poet, it is a failure to be a good poet. All of this would need to be defended in more detail to be entirely convincing, but it gives us some idea of the basic shape of a dialectical argument. It shows what it means to move from a defective conception of a given concept to an improved successor notion. The defective conception of inner and outer regards them as having potentially distinct content. This makes it possible for the outer to flatly oppose the inner, and for the inner to be prioritized over the outer. Hegel wants to replace this with a conception of inner and outer as identical in a distinctively speculative sense.8 According to this view, inner and outer share the same content and differ only in form (just as the claim that someone is a good poet and has written good poetry are not two distinct matters but two ways of saying the same thing). On this revised conception of the inner–outer relation, it is impossible for the inner and outer to flatly oppose each other, and so impossible for the inner to have a one-sided priority over the outer. The inner–outer dichotomy takes a wide variety of forms in Hegel’s writings, and most of these are irrelevant to our concerns here, but even in his logic, Hegel alludes to the special significance it has for action: As a human being is externally, i.e. in his actions (not, of course, in his merely corporeal externality), so he is internally; and if he is only internally virtuous, moral, and

33

Mark Alznauer

so forth, i.e. only in intentions and sentiments and his outer life is not identical with them, then the one is as hollow and empty as the other. (EL §140 R) In this context, the mistaken tendency to prioritize the inner over the outer shows up as an over-valuing of the moral “intentions and sentiments” of an agent at the expense of what she accomplishes in her outer life or external actions. Hegel’s view is that this involves the same defective conception of inner and outer that we have just explored, the view according to which the inner essence of a thing can differ in content from its phenomenal expressions. Against this, he thinks that a person’s internal intentions and dispositions are only as good as their outer life – this is where their value must prove itself. He summarizes his ultimate position by saying “a human being is nothing other than the series of his actions” (EL §140 A) – a slogan that re-appears in his official treatment of action in the Philosophy of Right.

II  The historical context Much contemporary interest in Hegel’s theory of action in the Anglophone world stems from a recognition that he anticipates a similar line of thought about inner and outer in post-Wittgensteinian action theory.9 For Charles Taylor, Hegel’s crucial idea is that an agent’s intention is not something that can be held apart from the deed, and thus something that could be said to cause it (as Donald Davidson claims) – the deed is just the expression of the intention (as G. E. M. Anscombe claims).10 For Robert Pippin, Hegel’s position anticipates Anscombe’s claim that if we want to know an agent’s intentions we should not start with a search into the contents of his mind; instead we are best served by looking at what he actually does.11 More recently, Robert Brandom has argued that Hegel builds on a basically Davidsonian distinction between intentional and consequential specifications of what was done on the way to arguing for a view according to which agents must accept responsibility even for those “features of the doing that stem from its contingent, unforeseen, indeed unforeseeable consequences.”12 These well-known interventions have focused attention on Hegel’s treatment of action in the Phenomenology of Spirit rather than on the more systematic treatment of action that can be found in the Philosophy of Right.13 It is not hard to understand why: Hegel’s treatment of action in the latter text is deeply interwoven with his account of the legal and institutional structure of the rational state, raising questions in legal and political theory that might seem far removed from the broadly psychological issues central to postWittgensteinian discussions of agency.14 But some attention to this general historical context of Hegel’s intervention can help explain why Hegel views the problem the way he does: as just as much a legal and political issue as a psychological and ethical one. For Hegel, what is most characteristic of modernity is its insight that all human beings are intrinsically free, that no one can be rightly treated in a way that they could not rationally assent to. Hegel thinks the origins of this belief can be found Christianity  – in the conviction that all men are equal before God  – and that modernity involves an attempt to translate this originally spiritual ideal into a political and social program. But although Hegel clearly sees this as an achievement of enormous value, he thinks our modern commitment to freedom hides a contradiction within itself, one that that was revealed in a particularly glaring way by the tragic aftermath of the French Revolution. It is this contradiction that his own philosophy of right is intended to resolve. 34

Hegel on guilt

The implicit contradiction is that between objective and subjective aspects of freedom.15 On the objective side, the ideal of freedom entails certain specific laws and political institutions, those that can be shown to be necessary to secure our individual freedom; it requires, for example, that individuals be protected from any violations of their personal and property rights, that there be courts to support these rights, and so forth. But the modern ideal of freedom also requires what he calls “subjective freedom,” that individuals must be able to recognize the social order as justified. This entails that the laws and customs of the state cannot be legitimate if its authority cannot be “apprehended by cognition as something rational for itself.” To the degree that subjective freedom is conceptualized as distinct or independent from objective freedom – fairly or not, Hegel blames Rousseau for making this mistake – it is possible for it come into opposition with objective freedom. Individuals may to subjectively fail to recognize objectively rational laws and institutions as legitimate. This is what Hegel sees as happening in the French Revolution: it was a “terrible and drastic event” in which all existing social and political institutions were overthrown in the name of subjective freedom.16 For Hegel the main challenge the French Revolution bequeaths to modern political philosophy is to find a way to reconcile the seemingly conflicting demands of objective and subjective freedom. If we fail, he thinks we will be left with two unpalatable alternatives: either a republican fanaticism that gives us subjective freedom without rule of law (think Robespierre and the Jacobins) or some form of reactionary restorationism that is willing to sacrifice subjective freedom to satisfy the imperatives of political stability and order (a possibility we might associate with Edmund Burke or Joseph de Maistre, but that Hegel identified with Karl von Haller). Hegel’s defense of responsibility in the Philosophy of Right directly reflects this complex political crisis. On his view, objective freedom requires individuals be held accountable for any violations of each other’s rights, but subjective freedom implies right to refuse responsibility for any act that was not done knowingly, intentionally, and with insight into its rightness or wrongness. If we cannot find some way to prevent these requirements from falling into contradiction, we are forced to choose between them: either granting that subjectivity has absolute authority over what it regards as good and evil or entirely giving up on the right of the subjective will and returning to some premodern confession. So the stakes of Hegel’s discussion of responsibility are unusually high – if the problem cannot be resolved, then we have no reason to think that the modern, bourgeois idea of freedom can be realized in a stable institutional order.

III  Responsibility in the Philosophy of Right With this background in place, we can turn to Hegel’s treatment of responsibility in the Philosophy of Right. In this work, Hegel’s strategy is to provide a kind of idealized genealogy of right – a series of increasingly adequate shapes (Gestalten) of right that culminates in a fully adequate picture of what it would require to institutionalize freedom. Something counts as “right” in the most general sense when it is a mode of interaction between individuals that realizes their mutual freedom. The various shapes of right that he considers – “Abstract Right,” “Morality,” and “Ethical Life” – each correspond to different conceptions of freedom, and each of these conceptions of freedom entails a different picture how human relations must be structured in order for this sort of freedom to be realized. Each shape also implies a different conception of responsibility. So just the notion of freedom is dialectically 35

Mark Alznauer

refined as we progress through the three parts of the book, so too is the concept of responsibility; it is only in the final section of the book that we arrive at a fully adequate conception of freedom and a correspondingly adequate conception of responsibility. This is the general argument that we need to reconstruct.17 Abstract Right, the first shape, is predicated on a conception of freedom as a matter of having a sphere of external activity in which others may not interfere; it is comprised of those norms and rules that are necessary to secure freedom in this sense and these are primarily rights of property and personhood. According to Abstract Right, we are acting rightly when we respect the personal and property rights of other individuals and acting wrongly when we violate those rights. This conception of freedom implies a minimal, external notion of responsibility; at this stage, an individual is blameworthy only for activity that infringes upon the rights of others. The notion of responsibility at play here is merely external in the sense that it does not yet take the agent’s own intentions into account.18 But there is a problem. Hegel thinks that someone whose rights cannot be enforced does not really have rights at all, but if we remain within the framework of Abstract Right we cannot address and remedy rights violations without generating further rights violations. He considers the possibility that each individual might enforce their own rights via revenge, but he finds that each act of revenge would be correctly perceived as fresh infringement of the original perpetrator’s rights. So it turns out that if freedom is simply a matter of respecting the rights of others – i.e., not violating their property or persons – then it cannot be realized in a stable institutional order. Hegel argues that the only solution is to recognize a further, higher conception of freedom. This new notion of freedom is termed subjective freedom, which gives rise to the next shape of right, Morality. Subjective freedom exists in an internal domain, the domain of the will and consciousness of the individual, as opposed to the external domain of personal and property rights. A person is free in this second sense only when he knowingly wills what is morally demanded of him. This means it is no longer sufficient to simply refrain from violations of external rights. In Morality, Hegel attempts to show what further duties and rights this involves – for example, he shows that it requires that we take the welfare of others into account in our actions. This subjective notion of freedom brings with it a revised, more restricted conception of responsibility. For Morality, it is not enough to show that we violated someone else’s external sphere of freedom to establish our blameworthiness. It must also be shown that we knew what we were doing (the right of knowledge), that we intended to do it (the right of intention), and that we knew it was wrong (the right of insight into the good).19 These are the rights of the subjective will; they collectively entail that an agent need not accept responsibility for any aspects of her deed which she did not know about, did not intend, or did not recognize to be wrong. Hegel never wavers from his claim that a recognition of these rights is a great historical achievement; he considers this the “pivotal and focal point in the difference between antiquity and the modern age” (PR §124 R). But he also thinks that these rights fall into contradiction with our abstract right to undisturbed personhood when they conceive the subjective will in the wrong way: as something present in an abstract inwardness, as something that could have an entirely different content than our external deed. Such a view of the subjective will entails the conceptual possibility of a crime committed with good intentions, a crime for which we cannot hold the agent accountable. 36

Hegel on guilt

For Hegel, the tendency to privilege of the subjective dimensions of freedom at the expense of its objective dimensions is a characteristic pathology of the modern age. We have already seen the basic form that his response to this will take in our discussion of the inner– outer problem. With regard to each of the rights of the subjective will, Hegel attempts to show that it is a mistake to view the inner aspect of agency (our knowledge, intention, or moral insight) as something that could be entirely opposed to the outer aspect (the external deed).20 This requires reconceiving the inner will and the outer deed such that they are two forms that same content takes. What could that mean? The details of Hegel’s argument get complicated quickly, but one of the examples he uses is helpful in getting the gist of his response. Hegel often tells of an arsonist who sets fire to one part of a house but who does not consciously anticipate or intend for the fire to spread further, say to the rest of the house or to other buildings in the vicinity (PR §119 R, §132 R). It is Hegel’s claim, and surely common sense, that such an arsonist must accept responsibility for all the consequences of the fire, even consequences he did not foresee. The usual way to understand this is as a case where the arsonist is to be held legally responsible for burning down the whole neighborhood despite the fact that he did not actually will that outcome. On this view, we balance the need for external accountability against the need to respect the criminal’s rights and find that, in this case at least, the former is more important than the latter. We hold him morally and legally accountable for full scope of what he did despite acknowledging the fact that he only willed a small part of this. Hegel rejects this solution to the problem because it regards the subjective and objective sides of the action as if they were “separable” or “indifferent” to one another, which he takes to be the paradigmatic mistake of Morality (PR §140; see also §107A). For Hegel, holding the arsonist accountable for consequences he did not foresee is not in any tension at all with the right of the subjective will once that is properly understood (PR §132 R). What we need is a way to understand the agent’s volition – his knowledge, intention, and insight – in a less ‘formal’ way. For example, Hegel thinks that if the arsonist is aware that what he is doing falls under the concept of arson, and if he is familiar with the laws concerning arson, then he can be said to know that committing an act under that description makes even unforeseen consequences “an existence of [his] own volition” (PR §119 A). This does not need to have been clearly present in his mind in the very moment he was setting fire to the house for him to be said to ‘know’ it in the relevant sense. Once we allow for this kind of reconceptualization of the subjective will, we can see holding the reckless arsonist responsible for what he has done does not require overriding the right of the subjective will or counterbalancing it with other considerations. Hegel’s dialectical reformulation of the right of the subjective will requires careful unpacking – and it is not clear that this can be done in way that conforms to all of our ordinary intuitions about these matters.21 But although Hegel’s approach is revisionist with regard to some of our intuitions, his solution is not simply a verbal sleight of hand; he is not declaring that whatever the agent brought about, she must have willed. For an agent to count as having willed something in the revised Hegelian sense, certain substantive conditions must be in place: we must be able to assume that the agent has both a familiarity with the general norms, customs, and laws of her society and that she accepts these norms, customs, and laws as genuinely obligatory. For Hegel, it is only if the arsonist can be assumed to be familiar with the fact that legal blame extends even to unforeseen consequences of any act of arson that holding him responsible for such consequences is consistent with the right of the subjective will. Hegel is explicit that these conditions are not met if the arsonist 37

Mark Alznauer

suffers from certain mental disabilities or is too young to appreciate or understand the consequences of his own actions. He is also clear that they are not met if the relevant laws and norms are not publicly promulgated or lack legitimacy or rational warrant. But although these conditions are substantive they are not unrealizable. Hegel clearly thinks that in modern states these conditions can be fully met and that, in such circumstances, there is no problematic gulf or breach that can open up between what an agent thinks she done and what her ethical community holds her accountable for. This solution requires positing a third conception of freedom as underlying the other two: one according to which a person is not free unless she fully identifies with the specific, historically given laws and institutions of her own country (PR §142 and §144). With this distinctively Hegelian notion of ethical freedom, we have made the transition to the final shape of right, Ethical Life. The question of what social relations and institutions are required for the full realization of ethical freedom is a complicated one, and it takes the rest of the Philosophy of Right for Hegel to provide his answer. But the ordinary individual in a rational state does not need to understand any of this complicated philosophical reasoning in order to know her obligations. He says: For the everyday contingencies of private life, definitions of good and bad are supplied by the laws and customs of each state, and there is no great difficulty in recognizing them. . . . The individual’s morality will consist in fulfilling the duties imposed upon him by his social station [Stand]. (RH 94) Hegel does not offer an explicit account of the notion of responsibility that corresponds to ethical freedom, and this has led to competing accounts of it, but the basic implication of his view is clear. In Ethical Life, the individual is responsible for a bundle of obligations or duties associated with her station as those are understood in her society. Hegel thinks parents in a modern society, for example, are responsible for bringing up children in a way that makes it possible for them to become self-sufficient members of such a society (PR §175). If they fail to meet the obligations attendant on their station, say they neglect some important need of their child, they can be blamed for it without checking to see if they were consciously flouting this duty. Hegel thinks blaming parents for being negligent of their duties is no violation of the rights of the subjective will because parents in a rational state can be assumed to be familiar with, and to accept, the general standards that pertain to their role (even if they might fail to realize what is required in some given instance).

IV Conclusion Hegel does not put it in just this way, but his account shows that the control and epistemic conditions are not jointly sufficient to determine someone’s responsibility for what she has done. He argues that the satisfaction of these two conditions still leaves us with potential gulf between what the agent regards herself as having done, her inner subjective will, and the objectivity of the deed, and that this gap can only be closed if the agent both knows and can be assumed to know that she is answerable to shared standards of action that are independent of her current state of mind and subjective opinions (PR §144).22 The responsible individual, in other words, must belong to a living ethical community – she must accept and identify with shared customary norms that attach to her ethical station. This implies 38

Hegel on guilt

that outside the context of a living ethical community, some degree of alienation from our external deeds is more or less inevitable. Hegel’s ‘social’ solution to the paradoxes of responsibility still has its defenders, but it has also been challenged by several important post-Hegelian thinkers. In his pseudonymous writings, Kierkegaard argues that Hegel exaggerates the degree to which any subjective act of the will could ever be made publicly available; he overlooks the possibility of an inner that is never expressed in the outer world, one that is strictly incommensurable with externality.23 A different line of critique is prominent among readers influenced by Marx. They suggest that the kind of full identification with the norms and customs of modern society that Hegel’s account depends on would be a form of false consciousness in our circumstances  – it would require overlooking the manifest irrationalities and contradictions of modern capitalism. Although Hegel’s emphasis on the social dimensions of human agency generates as many new problems as it resolves, it is perhaps all the more interesting for that.24

Notes 1 The summary of Hegel’s position in this paragraph is drawn from PR §§115–120. For a detailed and historically sensitive account of the ambiguities of Hegel’s use of the word Schuld, see Battistoni (2020a). Hegel’s framework certainly allows for cases of merely moral blame – cases where no legally enforceable right has been violated (see, for example, PR §213 A) – but all of the examples in the “Morality” section of the Philosophy of Right, which is the primary locus of his discussion of responsibility, concern crimes (like murder, arson, or theft). 2 Robichaud and Wieland (2017) is clear sign that this is changing. The potential paradoxes involved in the epistemic condition on responsibility have now been developed in great detail in the Anglophone literature, and several solutions have been offered (see Rudy-Hiller (2018) for a helpful summary of some of these). These discussions are clearly pertinent to the issue Hegel is addressing since they raise the question of whether Hegel has overlooked ways of solving the problem he is worried about that do not require a radical re-conception of agency as a social status. 3 See PR §121 A. He thinks the same gulf is also manifest in attempts to derogate the actions of heroic individuals because their motives were supposedly self-interested  – so it can be used to increase blame, not just reduce it. 4 PR §114 A. 5 PR §108. 6 What is characteristic of non-dialectical treatments of Hegel’s theory of responsibility is the view that the inner and outer aspects of the action are conceptually distinct, that these aspects can be in genuine (not merely apparent) conflict, and that a solution to any particular conflict must involve overriding the claims associated with one of these in favor of the claims associated with the other. Laitinen and Sandis (2019) offers a recent defense of an explicitly non-dialectical reading. More dialectical interpretations, as we will see, differ on all of these points. 7 SL 460–64; 11:364–368. EL §137–41. In emphasizing the importance of the inner–outer problem for Hegel’s theory of responsibility I am following Pippin (2008). See Yeomans (2018) for an interesting critique of Pippin’s view. 8 There is a helpful discussion of the role that speculative identity claims play in Hegel’s theory of action in Pippin (2008: 165–166). In pursuing a criticism of Brandom’s interpretation of Hegel, Levine (2022) offers a provocative alternative account of the speculative identity at stake here as a matter of a kind of hermeneutical circle – the inner is interpreted by means of the outer and vice versa. This nicely captures the conceptual interdependence of inner and outer that Hegel insists upon, though perhaps at the risk of a little anachronism. 9 Other commentators who have drawn on this connection include: Forster (1998), Pendlebury (2006), Quante (2004), Mead (2009), Deligiorgi (2010), Brandom (2013), Cargnello (2014), and Torralba (2015). See Yeomans (2010) for a helpful general summary of Hegel’s philosophy of action. 10 Taylor (2010). For criticism, see Maraguat (2019)

39

Mark Alznauer 1 Pippin (2008: 151), f. 11. For criticism, see Laitinen (2004). 1 12 Brandom (2019: 465). For criticism, see Levine (2022). 13 Outside the context of the Anglophone reception of Hegel, the treatment of action in the Philosophy of Right has always taken pride of place (see, for example, the classical treatments by Derbolav (1975) and Menegoni (1982)). 14 Quante (2004) attempts to disentangle the “Morality” section from its context in the Philosophy of Right. For criticism, see Houlgate (2010) and Alznauer (2012). 15 The quotations in the following are drawn from PR §258 R. 16 On Hegel’s ambivalent attitude towards the French Revolution, see Bourke (2022). 17 I offer a more detailed account of the following in Alznauer (2015). 18 He says “[W]ith right in the strict sense, it made no difference what my principle or intention was” (PR §106 A). Passages like these have suggested to many that external responsibility is coextensive with causal responsibility, and many if not most discussions of Hegel’s theory of responsibility proceed on that assumption (see Quante (2004), Battistoni (2020a), Brandom (2019), Meyer (2020)). My own view is that one is externally responsible only for what Hegel calls the “gerichtliche Handlung” in PR §113: this is the event as determined by rules of abstract right (see Alznauer (2015)). For a contrasting reading of PR §113, see Meyer (2017). 19 See PR §117, PR §120, PR §132 & R. 20 I deal with the specific problems that each of these rights give rise to in Alznauer (2015), Chapter  4. But for a different approach to the same material see Quante (2010) or Yeomans (2017). 21 On this objection to dialectical interpretations of Hegel’s theory of responsibility, see Laitinen and Sandis (2019). 22 Here I am in basic agreement with the account given in Battistoni (2020b). 23 See Alznauer (2019). 24 I want to thank Steven Levine and Maximillian Kiener for valuable comments.

Further reading The best short introduction to this topic is Wood, A. W. (2010). “Hegel on Responsibility for Actions and Consequences,” in A. Laitinen & C. Sandis (eds.), Hegel on Action (pp. 176–188). New York: Palgrave Macmillan. The same volume includes several other influential treatments of related themes in Hegel’s practical philosophy, like C. Taylor “Hegel and the Philosophy of Action” (pp. 22–41) and R. Pippin “Hegel’s Social Theory of Agency: The ‘Inner-Outer’ Problem” (pp. 59–78). Alznauer, M. (2015). Hegel’s Theory of Responsibility. Cambridge: Cambridge University Press and B. Caspers ‘Schuld’ in Kontext der Handlungslehre Hegels (Felix Meiner, 2012) are the only book-length treatments. A recent important critical synopsis of work in this area is Laitinen, A., & Sandis, C. (2019). “Hegel on Purpose,” Hegel Bulletin, 40(3), Winter, 444–463. G. Battistoni has recently published several interesting studies of these issues, most notably “The Polyvalency of Schuld in Hegel’s Morality: A Contribution to Translation Studies about a Fundamental Concept of Hegel’s Philosophy of Action” in Teoria: Revista di filosofia, Vol. 40, No. 2, pp. 209–220.

References Alznauer, Mark. (2012). “The Role of ‘Morality’ in Hegel’s Theory of Action,” The Owl of Minerva, 44(1/2), 67–92. ———. (2015). Hegel’s Theory of Responsibility. Cambridge: Cambridge University Press. ———. (2019). “Kierkegaard on Hegel’s Inner-Outer Thesis,” The Heythrop Journal, March, 60(2), 260–274. Battistoni, Guilia. (2020a). “The Polyvalency of Schuld in Hegel’s Morality: A Contribution to Translation Studies About a Fundamental Concept of Hegel’s Philosophy of Action,” Teoria: Revista di filosofia, 40(2), 209–220. ———. (2020b). “Wissen und Wollen: il fondamento dell’imputazione della responsabilità in G. W. F. Hegel,” in G. Battistoni (ed.), Fondamenti per un agire responsabile. Riflessioni a partire dalla filosofia classica tedesca (pp. 101–122). Milano: FrancoAngeli.

40

Hegel on guilt Bourke, Richard. (2022). “Hegel and the French Revolution,” History of European Ideas. DOI: 10.1080/01916599.2022.2095754. Brandom, Robert. (2013). “Some Post-Davidsonian Elements of Hegel’s Theory of Agency,” in S. Rödl & H. Tegtmeyer (eds.), Sinnkritisches Philosophieren (pp. 63–82). Berlin: De Gruyter. ———. (2019). A Spirit of Trust. Cambridge: Harvard University Press. Cargnello, David. (2014). “Beyond Morality: Intentional Action in Hegel’s Philosophy of Right,” Mind, 123(491), July, 671–706. Deligiorgi, Katerina. (2010). “Doing Without Agency: Hegel’s Social Theory of Action,” in Arto Laitinenen  & Constantine Sandis (eds.), Hegel on Action (pp.  97–118). New York: Palgrave-Macmillan. Derbolav, J. (1975). “Hegel’s Theorie der Handlung,” in Materialen zu Hegel’s Rechtsphilosophie (pp. 201–216). Frankfurt am Main: Suhrkamp. Forster, Michael. (1998). Hegel’s Idea of a Phenomenology of Spirit. Chicago: University of Chicago. Hegel, G. W. F. (1953). (RH), Reason in History, trans. R. Hartmann. New York: MacMillian. ———. (1991). (PR), Elements of the Philosophy of Right, trans. H. B. Nisbet. Cambridge: Cambridge University Press. ———. (2010a). (EL), Encyclopedia of the Philosophical Sciences in Outline, trans. and ed. K. Brinkmann & D. Dahlstrom. Cambridge: Cambridge University Press. ———. (2010b). (SL), The Science of Logic, trans. George di Giovanni. Cambridge: Cambridge University Press. Houlgate, Stephen. (2010). “Action, Right, and Morality in Hegel’s Philosophy of Right,” in A. Laitinen & C. Sandis (eds.), Hegel on Action (pp. 155–175). New York: Palgrave Macmillan. Laitinen, Arto. (2004). “Hegel on intersubjective and retrospective determination of intention,” Bulletin of the Hegel Society of Great Britain, 49, 54–72. Laitinen, Arto, & Sandis, Constantine. (2019). “Hegel on Purpose,” Hegel Bulletin, 40(3), Winter, 444–463. Levine, Steven. (2022). “Brandom on Hegel and the Retrospective Determination of Intention,” Hegel Bulletin. https://doi.org/10.1017/hgl.2022.22. Maraguat, Edgar. (2019). “Hegel on the Productivity of Action: Metaphysical Questions, NonMatephysical Answers, and Metaphysical Answers,” Hegel Bulletin, 40(3), 425–443. Mead, Aaron. (2009). “Hegel and Externalism About Intentions,” The Owl of Minerva, 41(1/2), 107–142. Menegoni, Francesca. (1982). Moralità e Morale in Hegel. Padova: Liviana Editrice. Meyer, Thomas. (2017). “Review of Hegel’s Theory of Responsibility by Mark Alznauer,” Zeitschrift für philosophische Literatur, 5(4), 38–47. ———. (2020). Verantwortung und Verursachung: Eine moral- und rechtphilosophische Studie zu Hegel. Hamburg: Felix Meiner Verlag. Pendlebury, Gary. (2006). Action and Ethics in Aristotle and Hegel: Escaping the Malign Influence of Kant. Burlington: Ashgate. Pippin, Robert. (2008). Hegel’s Practical Philosophy: Rational Agency as Ethical Life. Cambridge: Cambridge University Press. Quante, Michael. (2004). Hegel’s Concept of Action, trans. Dean Moyar. Cambridge: Cambridge University Press. ———. (2010). “Hegel’s Planning Theory of Agency,” in A. Laitinen & C. Sandis (eds.), Hegel on Action (pp. 212–231). New York: Palgrave Macmillan. Robichaud, Philip, & Wieland, Jan Willem (eds.). (2017). Responsibility: The Epistemic Condition. Oxford: Oxford University Press. Rudy-Hiller, Fernando. (2018). “The Epistemic Condition for Moral Responsibility,” Stanford Encyclopedia of Philosophy, accessed September  2022, from https://plato.stanford.edu/entries/ moral-responsibility-epistemic/. Taylor, Charles. (2010). “Hegel and the Philosophy of Action,” in A. Laitinen & C. Sandis (eds.), Hegel on Action (pp. 22–41). New York: Palgrave Macmillan. Torralba, José. (2015). “Hegel on Action as Expression: An Anscombean Reading of the Notion of Brute Facts,” Hegel-Jahrbuch, 2015(1), 379–384. Yeomans, Christopher. (2010). “Hegel’s Philosophy of Action,” in D. Moyar (ed.), The Oxford Handbook of Hegel (pp. 475–495). Oxford: Oxford University Press. ———. (2018). “Modernity and the Inner-Outer Problem,” in Australasian Philosophical Review, 2(4), 403–411.

41

Section 2: Responsibility and determinism

3 THE STOICS What kind of responsibility is compatible with divine providence? Rachana Kamtekar

Introduction Contemporary philosophers gloss determinism as ‘the thesis that everything that occurs is the inevitable result of the laws of nature and the state of the world in the distant past’ (McKenna & Coates 2021). And contemporary compatibilists maintain that determinism is compatible with moral responsibility in the basic desert sense, i.e., with praise and blame accruing to you in virtue of your control over your action, or your reasons–responsiveness, rather than for forward-looking consequentialist reasons (Pereboom 2021: 11–12). These philosophers argue for the compatibility of moral responsibility and determinism by: (1) denying that in order to be morally responsible for an action it must have been possible for you to do otherwise at the time of action, on the grounds that moral responsibility only requires that your character or your judgment or your intention was the actual cause of your action (following Frankfurt 1969); and/or (2) affirming that determinism is consistent with the ability to do otherwise on a correct understanding of that ability (e.g. Vivelhin 2004 and the new dispositionalists); and/or (3) arguing that whether or not determinism is true, our intimate interpersonal relationships are partly constituted by reactive attitudes such as resentment and gratitude, a presupposition of which is that we are morally responsible (following Strawson 1962). Since ancient Greek and Roman Stoics claimed that everything that comes to be does so by fate, and that agents are appropriate targets of praise and blame for their actions, they too are considered compatibilists. However, while recent scholarship focuses on whether or not the Stoics or their critics considered the ability to do otherwise a necessary condition of freedom and moral responsibility, this chapter will also examine their arguments over the kind of moral responsibility, or just praise and blame, that is compatible with fate: is this moral responsibility in the basic desert or some other sense? Let’s begin with a brief review of recent scholarship. Barring some scholarly disagreement about the interpretation and relative importance of (2) for the Stoic position, Susanne Bobzien’s landmark Determinism and Freedom in Stoic Philosophy (1998) established a consensus: for 45

DOI: 10.4324/9781003282242-7

Rachana Kamtekar

the Stoics, moral responsibility requires (1) that the agent be a cause of her action in the right kind of way, viz., by the mind’s rational assent to an impulsive impression, for which the ability to do otherwise, in the same circumstances, and with the same desires and beliefs, is irrelevant – even if that mind’s dispositions were shaped by genetic and environmental factors (276–290). If this is right, then ancient Stoics are Frankfurtians (Salles 2005: 63–68 makes the comparison explicit). Bobzien explains that since neither the Stoics, nor their critics for most of the duration of Hellenistic philosophy, conceived of moral responsibility as presupposing the agent’s ability to do otherwise than he did in the very same circumstance (1998: 277), when the Stoics speak about assent or action being ‘up to us’ or ‘in our power’ (eph’ hêmin), they only have in mind a general ‘two-sided potestative capacity’ to assent or not, or to engage in alternative kinds of action, such as to walk or not, in different circumstances (281). What matters for moral responsibility, on this view, is that you, and not someone or something else, knowingly did the action in question, and that it wasn’t necessary for you to do that action. Now while the Stoics clearly claim that the fatedness of all things is compatible with actions following assents being attributable to agents, their critics question whether praise and blame are just in response to fated actions, a question that is not answered simply by the observation that agents themselves are among the causes of their actions.1 Further, at least some Stoics explicitly reject attitudes of blame and anger (e.g., the resentment that is supposed to be constitutive of our interpersonal relations in (3)) for wrongdoing. It is, however, consistent with this to claim that the practices of blame, such as reproach and punishment, are just. This is for forward-looking reasons: beings with a faculty of assent, which is explanatorily idle in today’s mainstream accounts of Stoic compatibilism, are sensitive to and can learn from such practices, and thereby improve in the future. In what follows, Section 1 sets out Stoic theories of causation and of fate, and reviews their arguments for attributing actions to their agents rather than to fate. Section 2 shows how some Stoics reject attitudes of blame and anger, but can justify punishment and blaming practices, for forward-looking reasons.

1  How ‘everything happens by fate’ but some things are not necessary 1.1  Causation and fate For the Stoics, causation is a three-place relation, between (a) a body that is a cause to another body, (b) the body affected, and (c) an effect predicable of the second body. T1 The Stoics state that every cause is a body that becomes a cause, to a body, of something incorporeal; as the scalpel, which is a body, becomes a cause to the flesh, which is a body, of the incorporeal predicable ‘being cut’. And again, the fire, a body, becomes a cause to the wood, a body, of the incorporeal predicable ‘being burnt’ (Sextus M 9.211 = von Arnim 1903 [= SVF] 2.341, tr. after Bobzien 1998: 18, 53 n. 97; cf. Stobaeus 1.138.23–139.2, = SVF 2.336). Bodies have the causal power (aitia) they do in virtue of their so-called sustaining (sunektikon) cause, the air–fire blend (pneuma) that holds them together so that they continue in their current state, and this also gives rise to their ability to effect changes in other bodies. The affected body is receptive of the effect in virtue of its constitution, but this is its constitution as (passive) matter, the water–earth blend held together by its air–fire blend, with the latter being the cause of its resistance to change (Clement, Misc. 8.9.30, = LS 55C, = SVF 46

The Stoics

2.349, and Galen On Sustaining Causes 1.1–2.4, = LS 55F). The identification of causes as active or productive bodies and causal power as the active principle in them, as well as the contrast between active and passive power central to this account of causation, set it apart from the common contemporary account of causes as conditions, whether necessary, or sufficient, or necessary and sufficient, or ‘insufficient but necessary part of an unnecessary but sufficient set of conditions’ (Mackie 1965). The conception of cause as active has implications for how we understand the Stoic accounts of fate. Fate does not, at least in the first instance, connect events or facts to one another; rather, it connects bodies qua causes via their causal power, which is the active principle in them, to other bodies. This is why fate just is causal power, distributed among the bodies in the cosmos. Nothing happens except by fate. And all things happen by the coordination of these distributed portions of causal power. In other words, fate is: T2 a sequence of causes, that is, an inescapable ordering and interconnection (SVF 2.917 = Aetius Plac. I. 28, 4) T3 by fate, I mean what the Greeks call heimarmenê – an ordering and sequence of causes, since it is the connection of cause to cause, which out of itself produces anything. It is everlasting truth, flowing from all eternity. Consequently nothing has happened which was not going to be, and likewise nothing is going to be of which nature does not contain causes working to bring that very thing about. This makes it intelligible that fate should be, not the ‘fate’ of superstition, but that of physics, an everlasting cause of things – why past things happened, why present things are now happening, and why future things will be (Long & Sedley 1987 vol. 1 [=LS] 55L = SVF 2.921 = Cicero, On Divination 1.125–26). As a result, when the Stoics speak of natural law, they conceive of that law’s bearing on what occurs in a very different way than do contemporary accounts according to which the laws (plural) of nature are exceptionless generalizations about natural relations. For according to the Stoics the singular law of nature, reason, is the active principle moving, shaping, and ordering passive matter. Another name for this principle is God, and God is providential, intelligently ordering everything for the best. Now on the one hand, from the point of view of a particular natural body, the law’s determination of its movements need be no different from its own nature’s determination of those movements. But on the other hand, since this law determines the movements of all other natural bodies too, and the natural movements of two natural bodies seem to be able to conflict, as when they bump into one another, some of what happens to a natural body will seem like compulsion or contrary-to-nature movement. Should we say then that these movements are only apparently contrary to nature because there is really only one nature, that of the world as a whole as crafted by God (cf. Epictetus Discourses 2.5.24)? Or should we allow for conflicts between global and local natures? The problem replicates in the case of the distribution of reason or active principle among gods and humans. On the one hand, despite the fatedness of everything that happens, human rationality allows human agents to be causes of our actions via our assents to appearances and consequent impulses. Since the law of nature by which these actions are fated is reason, but our reason is a part of that reason, and our reason is expressed in our impulses, our actions result from impulses that are no less causally efficacious with respect to those actions than are God’s impulses. (This counters the ‘Lazy Argument’ according to which fate makes it pointless for you to act since ‘what will be, will be’; to this argument 47

Rachana Kamtekar

the Stoics answer: what will be is co-fated with your actions, cf. Cicero On Fate 28–29.) On the other hand, the question arises whether these actions belong to us any more than they belong to God, and if so, why.

1.2  The compatibility of fate with moral responsibility The Stoics maintain that everything’s coming about by fate is compatible with individual moral responsibility. But let’s first review their critics’ argument that these are incompatible: T4a ‘If all things come about through fate, all things come about through an antecedent cause. And if impulses [adpetitus, presumably by this the critics mean practical impressions (phantasiai hormētikai) which give rise to impulses to act when they are assented to, since they present acts of assent as ‘consequent’ upon them] do this, so do the things which are consequent upon impulse; therefore so do acts of assent. But if the cause of impulse is not located in us, neither is impulse itself in our power. If that is so, not even the results of impulse are in our power. Therefore neither acts of assent nor actions are in our power. The result is that neither commendations nor reproofs nor honours nor punishments are just’. Since this argument is unsound, they think it a plausible inference that not all events come about through fate. (Cicero On Fate 40–43 = tr. LS 62C, continued shortly as T4b and T4c; my interpolations in square brackets). T4a’s critics use a reductio ad absurdum to argue against the Stoic claim that all things happen by fate. They suppose that it is absurd (or at least false or implausible) that commendations and reproofs, etc., are unjust, and that assents and actions are not in our power. And since they suppose that these are consequences of the premise that everything comes about by fate, they conclude that this must be false. Framing their criticism as a misunderstanding of his causal claim, Chrysippus draws a distinction, between complete and primary causes on the one hand, and auxiliary and proximate ones on the other. T4b But Chrysippus, disapproving of necessity and at the same time wanting nothing to happen without antecedent causes, distinguishes between kinds of cause, in order to escape necessity while retaining fate. ‘Of causes’, he explains, ‘some are complete and primary (perfectae et principales), others auxiliary and proximate (adiuvantes et proximae). Hence when we say that all things come about by fate by antecedent causes (omnia fato fieri causis antecedentibus), we do not mean this to be understood as ‘by complete and primary causes’ but ‘by auxiliary and proximate causes’. Although we might think every cause is antecedent, the Stoics don’t; for example, the sustaining cause is contemporaneous with its effect. It seems that the antecedent causes because of which everything happens by fate are restricted to auxiliary (Greek sunerga, helping or intensifying) and proximate (Greek prokatarktika, initiating) causes. We should interpret Chrysippus’ claim ‘all things happen by fate by antecedent causes’ (‘omnia fato fieri causis antecedentibus’) so that it is consistent with his description of fate as inclusive of all causes in other passages (e.g., T2, T3, as well as T8 below, cf. SVF 2.917, 921, 945). So, presumably, Chrysippus thinks that all things happen by fate by antecedent causes in conjunction with other causes, including us. Perhaps his idea is that fate both uses antecedent causes to 48

The Stoics

coordinate primary causes with each other (for even primary causes need co-causes or helpers or circumstances), and also is the sum of all these causes. Chrysippus’ distinction among causes allows him to argue that antecedent causes’ not being in our power does not take our assents and actions out of our power, as long as the former are not primary and complete causes of these assents and actions: T4c ‘If all things come about by fate, it does follow that all things come about by prior causes – not however by primary and complete causes, but by auxiliary and proximate causes. If these latter are not in our power, it does not follow that not even impulse [adpetitus, presumably Chrysippus uses ‘impulse’ (hormē) for the action-directed movement that follows assent] is in our power. If, on the other hand, we said that all things come about by complete and primary causes, it would follow that, since these causes were not in our power, impulse would not be in our power either. . . . He does not want assent, at least, to be able to occur without the stimulus of some external force (for assent must be prompted by an impression). But he resorts to his cylinder and spinning-top: these cannot begin to move without a push; but once that has happened, he holds that it is thereafter through their own nature that the cylinder rolls and the top spins. Hence, he says, just as the person who pushed the cylinder gave it its beginning of motion but not its capacity for rolling, likewise, although the impression encountered will print, and, as it were, emblazon its appearance on the mind, assent will be in our power. And assent, just as we said in the case of the cylinder, although prompted from outside, will thereafter move through its own force and nature. If something were brought about without an antecedent cause, it would be untrue that all things come about through fate. But if it is plausible that all events have an antecedent cause, what ground can be offered for not conceding that all things come about through fate? It is enough to understand what distinction and difference obtains between causes’. When Chrysippus says that if the antecedent causes were also primary or complete causes of everything that came to be, then indeed our assents and actions would not be in our power, he implies that (i) if a cause other than ourselves is sufficient to produce an effect, then the effect is not in our power, but also (ii) even if the cause is not sufficient, if it is primary, then the effect is not in our power.2 Chrysippus does not in fact say that our assents or impulses are the primary or complete causes of our actions (cf. Bobzien 1998: 261). Perhaps, given the Stoics’ three-place model of causation, no single item (aside from God) could be the complete (autoteles) cause of a change to a body, although a single item could be of its remaining in the same condition (e.g., the cylinder’s continuing to roll straight down as opposed to rolling in place). And presumably some cases of human action have assent as their primary cause but others (e.g., actions that are due to madness like Heracles’ slaying his children) don’t. Bobzien interprets Chrysippus in T4c as saying that that the nature of the mind is the ‘decisive’ factor in an agent’s assents and actions, for the faculty of assent ‘makes it possible for adult human beings to become the controlling factor of their actions, in that it enables them to respond to externally induced impressions in accordance with the individual nature of their mind’ (1998: 269). Chrysippus’ main argument for this is that different agents respond to the same externally induced impressions differently. There’s a cake on the table, and I reach for it but you do not. This shows that the primary (assuming this is the same as ‘decisive’) cause of my reaching for the cake is my (greedy or intemperate) disposition, and 49

Rachana Kamtekar

the primary cause of your not reaching is your (health-seeking or temperate) disposition. But to establish which causal factor is the primary one, Chrysippus would need to consider more contrasting cases: what about the fact that I don’t reach out for a plate of kale that’s in front of me? What if neither you nor I reach for the plate of kale – does that show that when it comes to kale, it is the primary cause of its not being eaten?3

1.3  Possibility and non-necessity Now in T4c Chrysippus does not deny that all the co-causes of our actions jointly necessitate our actions. He simply draws attention to the distribution of causal responsibility for our actions and observes that our actions (also) depend on us in the sense that we are among their actual causes. Why does Chrysippus ‘disapprove of necessity’ while maintaining that all things are fated and that fate is inescapable? According to Bobzien, Hellenistic philosophers agree ‘that a prerequisite for something’s depending on us is that it is both possible and non-necessary’ (97).4 The Stoics give the following account of possible and non-necessary propositions: T5 Possible is that which admits of being true and which is not prevented by external factors from being true, such as ‘Diocles is alive’. . . . Non-necessary is that which both is true and is capable of being false, and is not prevented by external factors from being false, such as ‘Dion is walking’ (DL 7.75). How do we relate these accounts of possibility and non-necessity for propositions to the possibility-and-non-necessity of token effects that would be required for them to depend on us? On Bobzien’s account of the relationship between propositions and the states of affairs they describe, a given proposition is true at a time if a corresponding predicable (e.g., being alive, walking) obtains at a given body during that time (1998: 25–26). On this model we can say that a given proposition predicating a predicable of a body is possible if the predicable is not prevented from obtaining at the body, either by the body’s nature or by anything external to the body, at some time from now on. And that proposition is non-necessary if the corresponding predicable is not prevented from failing to obtain at that body, either by the body’s nature or by anything external to the body, at some time from now on (cf. 1998: 115). So, for example, ‘Dion walks’ is possible if nothing about Dion’s nature rules it out and Dion is not tied down, injured in the legs, etc., at some point from now on. And ‘Dion walks’ is not necessary if Dion’s nature is consistent with his staying still and nothing external is forcing Dion to keep walking, at some point from now on. The two conjuncts in this account (‘in the body’s nature’ and ‘external to the body’) pick up on two sources of necessity: the necessity of a predicable given a body or its nature, and compulsion or force (being moved contrary to nature or hindered from natural movement). When we plug these back into the causal network of fate, we can see that many token effects are possible and non-necessary, because many causal powers of bodies can only be exercised in relation to other bodies that are receptive of them. Effects which depend on us will be a subset of these. This suggests that our question, ‘why does Chrysippus disapprove of necessity while maintaining that fate is inescapable?’ is really about why Chrysippus disapproves of local but not global necessity. A plausible answer is that global necessity is the result of providence coordinating the bodies in the world in the best way possible, but local necessity, if it 50

The Stoics

is from a thing’s nature, rules out changes to that thing, and if it is from something external, implies compulsion, i.e., change or inhibition of change which is contrary to that thing’s nature. In the context of a science of nature(s) (‘physics’), Chrysippus’ causes should allow for change that has its source within a body and that is in accordance with its nature while at the same time being providentially governed. But however adequate Chrysippus’ distinction may be to his physics, his arguments against local necessity secure for the agent only some contribution via impulse to a particular action given some other antecedent cause. If we think that what threatens moral responsibility is not some particular body the causal power of which would compel the agent or deprive her of alternatives, but rather the system of individually necessary and collectively sufficient causes that constitute fate, it’s hard to see how anyone could be morally responsible aside from God or cosmic Reason (which acts as it does because intelligence by its own nature causes what there is best reason to cause).

1.4  What is ‘the agent’? Why is assent important? Brennan (2005: 263–265) complains about Chrysippus’ account of responsibility: is the responsible agent the agent considered along with or without her dispositions? When Chrysippus says that the agent who assents to an impression is not necessitated because she is capable of not assenting, ‘agent’ refers to the bare faculty of assent, abstracted from any dispositions. But when Chrysippus likens the different ways that a cylinder and a cone roll, even if they both receive the same initial push, to the different responses of two agents to the same impression, then, in order to assign the agents causal and moral responsibility for their assent and impulse, ‘agent’ includes their dispositions. And it’s never open to our faculty of assent, given our mental dispositions, to assent or not in the very same circumstance (300–301). It’s worth noting that the previous account of the Stoic justification of praise and blame and reward and punishment has made little use of the faculty of assent – aside from mysteriously assigning it the ‘decisive’ role in action. Salles (2005: 51–68) reconstructs a separate argument according to which acting on a judgment based on reasoning (krisis) is sufficient for moral responsibility, and suggests that this is a Stoic reply to the Aristotelian claim that moral responsibility requires the agent to have had the ability to do otherwise at the moment of action (87–88). The textual basis for Salles’ reconstruction reads: T6 Those who hold that both that which depends on us and that which is in accordance with fate are preserved (for something in accordance with fate is given to each thing, just as cooling to water, bearing this sort of fruit to each of the plants, moving downwards to the stone, moving upwards to fire, so too giving assent and exercising impulse to animals; when nothing of the things external and according to those resists this impulse, then walking fully depends on us and we will definitely walk) those who say these things (they are Chrysippus and Philopator and many other famous ) do not prove anything else but that everything happens in accordance with fate (Nemesius On the Nature of Man 106, 6-ff., tr. Salles, cf. Alexander On Fate 13–14). In T6, ‘depending on us’ consists in ‘depending on our nature’, which in the case of human beings involves rational assent. But even here the connection between rational assent and 51

Rachana Kamtekar

praise or blame is unclear. For unless something further is filled in, the argument seems to license praising the water for cooling and the plant for fruiting, and blaming them if they fail to do these things. Below, I argue that the late Stoic Epictetus conceives of moral responsibility in a way that makes use of the fact that the faculty of assent is an intelligent cause that can be affected by praise and blame, reward and punishment.

2  Why punishment but not anger is compatible with determinism 2.1  Epictetus on assent, appearance, and the character of the mind Previously we saw that the Stoics’ critics in T4a are not merely questioning the causal attributability of an action to an agent. They are complaining that if fate is the cause of everything, then praise and blame are not just. If the justice of praise and blame were settled by causal attribution, should we not also blame the other co-causes, for example, the antecedent causes? If not, what else is required? You may answer that these other causes lack the faculty of rational assent, but exactly why should only this be subject to praise and blame? Even if reason is uniquely critical, or uniquely involves self-awareness, unlike the perceptual capacities of other animals, how exactly does this bear on the justice of praise and blame? The later Stoic Epictetus says that all blame, including anger at wrongdoers, should be avoided. This is because it expresses impiety: if all things are fated, then we ought to appreciate everything that happens, including others’ wrongdoings, as part of the providential government of the cosmos (3.26.18–19, cf. 2.19.26, 2.23.42, 3.24.79, 4.1.109). When it comes to blaming wrongdoers specifically, Epictetus says we should not be angry at them, for they are assenting on the basis of how things seem to them, as they must. Here is Epictetus on the connection between appearance and assent: T7 What is the cause of assenting to something? Its appearing that it’s so. Therefore, it is not possible to assent to what doesn’t appear to be so. Why? Because this is the nature of thought – to say yes to truths, to say no to falsehoods, and to suspend about unclear things. What is the proof of this? “Feel, if you can, that it is now night.” That is not possible. “Put away the feeling that it is day.” That is not possible. “Either feel or put away the feeling that the stars are even in number.” That is not possible (Epictetus Discourses 1.28.1–4, my tr. modified from Oldfather; cf. 1.18.21). Epictetus seems to think that appearances compel assent or suspension, where ‘appearance’ refers to the joint product of external impression and prior beliefs (cf. Brittain 2014). If we must assent or suspend depending on our appearances, he thinks, then instead of blaming and being angry at wrongdoers, we should try to change their beliefs. For example, Epictetus says about Medea, who killed her children in order to take vengeance on her traitor husband: T8 Show her clearly that she is deceived, and she will not do it; but as long as you do not show it, what can she follow other than what appears to her? Nothing. Why then are you angry with her, because the wretch has gone astray about the greatest matters, and has become a viper instead of a human being? And why not rather pity her, if possible, as we pity the blind and the lame, so too [pity] those who are blinded and maimed in the most authoritative faculties (ta kuriôtata)? (1.28.8–9, cf. 1.18). 52

The Stoics

Rather than grounding praise and blame (per Bobzien 1998: 269), assent’s dependence on our appearances and hence (as the parallel with the blind and the lame suggests) on the character of our mind seems to be, for Epictetus, a reason to refrain from blame and anger at wrongdoing. Now that a Stoic should refrain from anger against wrongdoing also follows from the Stoic account of the passions as based on false value–judgments. For example, my anger at you for slighting me in public is based on the false judgment that my public image is a good, when in fact it is indifferent to my happiness. The false judgment is a hindrance to my knowledge, virtue and happiness, and so is to be eliminated. However, Epictetus recommends that, unless we can completely eradicate our passions, we should pity the wrongdoer rather than be angry at her. Pity is also a passion, but it’s preferable to anger. So there must be something wrong with anger in addition to its being based on false value judgments. In T7 and T8 Epictetus suggests that because the wrongdoer cannot but follow her false appearances, anger at her wrongdoing is inapt. If the presupposition of anger is that she could have done otherwise than she did, Epictetus argues, that presupposition is false, for she – given her epistemic condition – could not. Epictetus, then, seems to stick firmly to one of the alternatives Brennan finds Chrysippus to slide between (section 1.4): it is the agent considered along with her dispositions that would be morally responsible (or not) for her actions.5

2.2  Chrysippus on the fatedness of actions, given character However, Epictetus’ account of the relationship between assent and the character of the mind is anticipated in another of Chrysippus’ replies to an incompatibilist argument: T9 In On Providence book 4 he [Chrysippus] says ‘Fate is a certain natural everlasting ordering of the whole: one set of things follows on and succeeds another, and the interconnection is inviolable . . . [from which his critics infer that] the wrongdoing and fault of humans should not be subjected to anger (suscensenda) nor should these be attributed to their wishes, but to that necessity and impulse that arises from fate. ‘[Chrysippus replies that] Although it is true . . . that all things are constrained and connected by fate by a certain necessary and principal reason, nevertheless, the natural dispositions of our minds are subject to fate in accordance with their own peculiarity and quality. For if they were initially made healthy and useful, they pass over, without obstruction and opposition, every force that comes in from the outside by fate. But if they are rough, uneducated and rude, and supported by none of the supports of good skills, they fall into continual faults and errors by their own wickedness and voluntary impulse, even though pressed by a slight or no blow of a fated disadvantage. And that this very thing happens is the product of that natural and necessary consequence of things which is called ‘fate’. For that bad dispositions should not be free of wrongdoings and errors is so to speak fated and consequent from its own kind’ (Aulus Gellius Attic Nights 7.2.3–10, = SVF II.1000, my tr.). Unlike the critics in T4a, the critics in T9 allege only that anger at the agent for her wrongdoing and attribution of the wrongdoing to her wishes (i.e., judging her action voluntary) are not justified if all things come about by fate – they do not mention reward and punishment, and they are not giving a reductio ad absurdum.6 53

Rachana Kamtekar

Who are these critics? It’s likely that in a work called On Providence Chrysippus would have engaged the views of other philosophers on providence. Platonists and Stoics both drew on Plato’s Timaeus, which describes a providentially designed cosmos and sketches a position similar to the critics of T9. Timaeus provides an aetiology of vice, which he says is unwilling (86d-e) and arises in the soul from bad bodily conditions, e.g., excessive pleasures and pains that compromise our judgment and learning (86c, 87a), and from bad constitutions that produce bad culture and education (87b). Timaeus says that it is our begetters and nurturers that are blameworthy (aitiateon) for our badness rather than we ourselves. Instead of blame, he recommends a bodily and psychic regimen to cure us of our psychic illness (87c-89d).7 Chrysippus’ statement in T9 that the wrongdoings that follow from a bad character are ‘so to speak fated and consequent from its own kind’ anticipates Epictetus’ claim in T7 that we cannot but follow what appears to us. Given that On Providence will have argued that God made the best possible world, and that the bad that exists is a necessary concomitant of Providence’s producing good (cf. Attic Nights 7.1: it’s not possible for good to exist without bad), it makes sense for Chrysippus to add that bad characters, necessary concomitants of good ones, can’t but produce bad actions. If we step back and view the wrongdoing from a global perspective, we will resign ourselves to wrongdoing’s existence as an inevitable consequence of a necessary concomitant. But this is very far from seeing blame and anger at wrongdoing as just. We noted previously that the objectors in T9 say that we should not be angry at wrongdoing and vice, not that wrongdoers should not be punished. Renouncing the attitudes of blame and anger is compatible with retaining the practices of punishment and reproof (all mentioned in T4a as ‘not just’ if everything is due to fate) as corrective to bad character. This is where the faculty of assent, linchpin of our rationality, seems most relevant. Since the faculty of assent responds to appearances, the appearance of reproof or punishment in response to a wrongdoing can lead the agent to reassess the value of actions of that sort. This is a forward-looking justification for punishment, of course, and while it shows how punishment can be compatible with determinism, it is silent about how attitudes of praise and blame, or moral responsibility in the basic desert sense, can be compatible with it. Some scholars of Stoicism claim that it is ‘common ground’ between the Stoics and their critics that ‘some actions . . . deserve praise and blame’ (Brennan 2005: 243, my emphasis). But if some Platonists and some Stoics reject attitudes of blame and anger in response to wrongdoing, and approve of punishment for forward-looking reasons, then assumptions about desert are not common ground. In light of this we should reconsider the often-cited quip: T10 The story goes that Zeno was flogging a slave for stealing. ‘I was fated to steal’, said the slave. ‘And to be flogged’, was Zeno’s reply (Diogenes Laertius 7.23 = LS 61E). Zeno’s response to his slave’s plea tells us only that fate is not grounds for being excused from punishment; it does not license blame as deserved.8

Notes 1 A caveat: I take praise and blame to be directed at agents for actions. I do not think that this varies between moral outlooks (contrary to Watson 2004; Shoemaker 2015, according to whom the concern with attributability goes with aretaic assessment and is of agents rather than actions, whereas accountability goes with an interpersonal concern with agents’ quality of will).

54

The Stoics 2 Sedley (1993: 323–325) proposes a distinction between ‘primary’ and ‘complete’: ‘primary’ causes are salient; examples include ‘external’ environmental factors as well as ‘internal’ assents. 3 Brennan (2005: 261–262) raises another objection: if I, who weigh 250 lbs, am in a lifeboat with a 300 lb capacity and you, who weigh 100 lbs, jump in, can you really deny that you caused the lifeboat to sink on the grounds that if I hadn’t been in it, it wouldn’t have sunk? 4 However, Bobzien 1998: 136–143 segregates, as belonging to separate parts of Chrysippus’ philosophy, the logical concept of the necessary (to anangkaion), according to which not everything that happens is necessary, from the cosmological concept of necessity (anangke), according to which everything that is fated happens necessarily (cf. Salles 2005: 82–84). Bobzien stresses that the logical concept of the necessary does not involve any views about how what must come about does so, and that the cosmological concept of necessity doesn’t involve logical necessity (140). But if these belong to different parts of his philosophy, why does what depends on us – a causal notion – require the modal status of possible and non-necessary? Bobzien admits that the consistency achieved by the segregation has as a consequence that ‘the problem of the integration of that which depends on us, and of moral responsibility, into Stoic theory of fate is not even touched upon’ (142). 5 Braicovich 2010 argues that Epictetus’ injunction to refrain from punishment and censure given that we cannot but do what we judge best (which he calls ‘epistemic determinism’) is consistent with our being responsible for our actions just because it was we rather than someone else who did them (214, on 3.26.28) because the latter is a conception of causal rather than moral responsibility (217). Braicovich finds a notion of moral responsibility in the idea that we are accountable (hupeuthunos, 1.12.32) to God for how we use our impressions, viz., for our assents and opinions (215), but in this context Epictetus’ point is that God has made us only accountable for this, and our happiness depends on nothing else since the consequence of our use of impressions is our character. 6 So I find Bobzien’s reconstruction of T9 (1998: 243–245) along the lines of T4a, supplying to it (i) an affirmation of the justice of attributing actions to and being angry at wrongdoing agents, and (ii) the conclusion that fate is not the cause of everything that happens, to be unwarranted. 7 I discuss Plato’s position of the unwillingness of wrongdoing and vice, and the inappropriateness of anger and blame in response to it, in this passage and Plato’s Laws in greater detail in Kamtekar 2019. 8 I’m most indebted to Maximilian Kiener for comments on previous drafts of this chapter. I’m also grateful to participants of the Oxford Moral Responsibility Workshop and the Cornell Ancient Philosophy Workshop in Fall 2022, and to an anonymous reader, for their feedback.

Further reading Bobzien, S. (1998). Determinism and Freedom in Stoic Philosophy. Oxford: Oxford University Press is the classic scholarly reconstruction of Stoic thinking on moral responsibility and determinism in its ancient context. Bobzien, S. (1998). “The Inadvertent Conception and Late Birth of the FreeWill Problem,” Phronesis, 43(2), 133–175, argues that it is the 3rd century Peripatetic Alexander of Aphrodisias who introduces the indeterminist interpretation of ‘up to us’ as ‘could have done otherwise’ into the debate. Part IV of Brennan, T. (2005). The Stoic Life. Oxford: Oxford University Press is an accessible and critical discussion of the Stoic position and arguments for and against.

References Bobzien, S. (1998). Determinism and Freedom in Stoic Philosophy. Oxford: Clarendon Press. Braicovich, R. S. (2010). “Freedom and Determinism in Epictetus’ Discourses,” Classical Quarterly, 60(1), 202–220. Brennan, T. (2005). The Stoic Life. Oxford. Brittain, C. (2014). “The Compulsions of Stoic Assent,” in M. Lee (ed.), Strategies of Argument: Essays in Ancient Ethics, Epistemology, and Logic (pp. 332–355). Oxford. Epictetus, tr. Oldfather, W. A. (1925–28). Discourses, Manual and Fragments. 2 vols. Harvard.

55

Rachana Kamtekar Frankfurt, H. (1969). “Alternate Possibilities and Moral Responsibility,” Journal of Philosophy, 66, 829–839. Kamtekar, R. (2019). “Aristotle contra Plato on the Voluntariness of Vice,” Phronesis, 64, 57–83. Long, A. A., & Sedley, D. N. (1987). The Hellenistic Philosophers. 2 vols. Cambridge. Mackie, J. L. (1965). “Causes and Conditions,” American Philosophical Quarterly, 2, 245–264. McKenna, M., & Coates, D. J. (2021). “Compatibilism,” in Edward N. Zalta (ed.), The Stanford Encyclopedia of Philosophy  (Fall 2021 ed.). https://plato.stanford.edu/archives/fall2021/entries/ compatibilism/. Pereboom, D. (2021.) Wrongdoing and the Moral Emotions. Oxford. Salles, R. (2005). The Stoics on Determinism and Compatibilism. Ashgate. Sedley, D. N. (1993). “Chrysippus on Psychophysical Causality,” in J. Brunschwig & M. C. Nussbaum (eds.), Passions and Perceptions: Proceedings of the Fifth Symposium Hellenesticum (pp. 313–331). Cambridge. Shoemaker, D. (2015). Responsibility from the Margins. Oxford. Strawson, P. (1962). “Freedom and Resentment,” Proceedings of the British Academy, 48, 1–22. Vivelhin, K. (2004). “Free Will Demystified: A  Dispositional Account,” Philosophical Topics, 32, 427–450. Von Arnim, H. (1903–05). Stoicorum Veterum Fragmenta. 4 vols. Teubner. Watson, G. (2004). Agency and Answerability: Selected Essays. Oxford.

56

4 HOBBES AGAINST BRAMHALL Moral responsibility, free will, and mechanistic determination Thomas Pink

1  Hobbes’s debate with Bramhall As he worked on the political theory that culminated in Leviathan, Hobbes was developing a new theory of freedom. Freedom was no longer as generally conceived by the philosophers and theologians of Hobbes’s time. Freedom was no longer peculiar to beings capable of rationality, such as humans, as a power to determine for themselves what they do. It was instead a phenomenon found throughout nature – not a distinctive kind of power but an absence of external obstacles to power. As Hobbes insisted, a river may flow freely if the force or power of its current is unimpeded by any dam. Similarly we humans act freely if there is no external obstacle to the power of our will to cause us to act as willed. Hobbes denied any freedom of the will itself; freedom is simply an unimpeded capacity to act as willed and is entirely consistent with the necessitation of our will by antecedent causes. In The Questions Concerning Liberty, Necessity and Chance,1 the debate about freedom between Hobbes and John Bramhall, his opponent defended the opposite: that freedom exists as liberum arbitrium, a freedom of the will itself, a power to determine for ourselves what actions we will or decide to perform, and a power that is inconsistent with the determination of our will by antecedent necessitating causes.

2  Moral responsibility and power before Hobbes Hobbes proposed his view by invoking Reformation Protestantism, accusing Bramhall of being a fellow traveller of Catholicism. But in fact Hobbes’s theory was radically transforming of the entire Reformation debate about free will – of the Protestant side as well as of the Catholic. All parties to that debate had assumed a theory of responsibility which I shall term the imputation model. In moral blame we not only criticise an agent’s action as bad but impute the badness of the action to the agent on the basis that the agent had a power to determine the occurrence of the bad action for themselves. Not only was the action faulty; but its occurrence was the agent’s own fault. In an influential exposition of the model, Aquinas called this power to determine one’s action for oneself dominium or dominion,

57

DOI: 10.4324/9781003282242-8

Thomas Pink

and regarded it as a power exercised by the agent over the operation of their will – of their capacity to decide how they would act: For an action is said to be blameworthy or praiseworthy which is imputed to the agent. For praise and blame is nothing other than for the goodness or badness of his action to be imputed to someone. Now an action is imputed to an agent when it is in his power, so that he has dominion over the act. But this is the case with all voluntary actions: for it is through the will that man has dominion over his action.2 Liberum arbitrium was the term commonly used for this power over decision or choice, and the reality and general nature of this power was common ground to theologians both Catholic and Protestant. So too was the identity of the will as the primary locus of human agency. Action as involving the exercise of liberum arbitrium occurs not simply as the moving of a hand but as a decision or choice to move one’s hand. All parties assumed too that the power of liberum arbitrium could be exercised contingently. Ordinary efficient causes operate by necessity. Given its mass a brick hurled at a sufficiently unreinforced window will break it; the brick’s operation to cause the breakage is necessitated by its own power and the circumstances of its operation. A  brick is a necessary cause. But our power of liberum arbitrium to determine our actions does not similarly operate by necessity. We can possess a power to raise our hand, and there may be no obstacles to its successful exercise. But because it is up to us whether we raise our hand, it remains contingent whether we exercise our power to raise our hand, or to refrain from raising it, or even to lower our hand instead. Liberum arbitrium is a kind of efficient causal power – but it is efficient causation in a form that can operate contingently as a power over alternatives. Besides freedom as liberum arbitrium, the theologians generally supposed that human life and action involve another distinctive kind of power that is also not operative in ordinary nature – this time not a power exercised by the agent, but a power to which the agent is subject. This is a power of reason to move the agent to believe and decide in ways that are reasonable or justified. This power is normative in the following sense. It is the power of a value, of truth or of goodness, to move us by supporting objects of our thought – possible objects of belief or of decision – and thereby, in so far as we are indeed rational, produce attitudes directed at those objects. So the power of truth moves us to form beliefs directed at claims that are true, and the power of goodness moves to decide to perform actions that are good. The power of goodness to move the will to decide to act in ways that are good was identified with Aristotle’s final causation; it is the power of a goal, an end or finis, entertained as an object of our thought to move us to decide on and pursue it. Freedom as a power exercised by us over the will was called ‘physical’, not because the power-bearer is material or the power exercised purely over matter – the will is an immaterial faculty – but because the power and its exercise depends on the real existence of its bearer. Whereas a final cause operates ‘metaphorically’, not because its power is not real – the power is real and genuinely productive, attracting us to decide on it – but because the cause exercising the power lacks real or extra-mental existence. Though the goal or finis really does move us to decide on its attainment, it may never be realized. The final cause may remain forever merely a goal – merely an object of our thought: Thus they say one and the same action of the will is caused by the end and by the will itself, and in so far as it is caused by the will, the causation is efficient, in so far 58

Hobbes against Bramhall

as by the end, the causation is final, and in respect of the former the motion is real and proper, since such an action comes from the power as from a properly physical principle, and in respect of the latter, the motion is metaphorical since it comes from an object attracting and drawing the will towards it.3 Freedom understood as a power over alternatives that operates contingently was generally held to depend on our possession of a capacity for reason. Appeal was made to Aristotle’s Metaphysics: Secondly it should be said that all causes in so far as they operate without the use of reason operate by the above-mentioned necessity. This is taken from Aristotle, book 9 of the Metaphysics, chapter  2, where he makes this distinction between rational and non-rational capacities, that non-rational capacities are determined to one, but rational capacities are indifferent between opposites.  .  .  . Which can be confirmed inductively; for it is confirmed by experience at every level of thing, up to brute animals.4 The debate between Catholic and Protestant took this metaphysics for granted and was conducted in its terms.5 The controversy concerned the impact of Adam’s original sin and fall on our rationality. All agreed that this original event, the Fall, had degraded human nature and that, in particular, it had damaged our capacity to respond properly to the force of goodness – to decide as was justified. But how far did the damage extend, and what were the implications of the damage for our responsibility? All parties agreed that we retained some capacity to exercise liberum arbitrium contingently. It remains up to us whether, for example, we decide to steal or to refrain from stealing. The Catholics further insisted that it also remains contingent whether we decide in a way that is morally good – whether in deciding to refrain from stealing we still respond to and aim at the bonum honestum, the genuinely moral good, as opposed to responding to and aiming simply at our own selfish good. Moreover, on the Catholic view, this contingency is essential both to the very character of liberum arbitrium as a power of selfdetermination, and so also to our moral responsibility for our choice of moral good or bad. By contrast, Protestants who adhered closely to Luther and Calvin in their view of moral responsibility claimed that the Fall had entirely removed our capacity to respond to the bonum honestum. Whatever we now decide and do, we now decide and do sinfully, for sinful ends such as our own pleasure and the like. If we avoid stealing it may simply be to avoid the penalty of being caught. We now sin of necessity, out of a corrupted nature – but when we sin we still determine for ourselves that we do so. Contingency is, they admitted, still a common feature of the exercise of liberum arbitrium – but only in respect of aspects of our agency that are ‘civil’ rather than ‘moral’, that is only in respects that leave our actions still sinful no matter how conforming to civil or positive law in other respects. Contingency moreover is not essential to liberum arbitrium and to its character as a mode of self-determination. What is essential to liberum arbitrium is instead something entirely consistent with necessity – the absence of coactio or compulsion. Coactio is the imposition of an outcome on us against our will. I might compel you – impose coactio – by tying your hand down, hence blocking your decision to raise it. But coactio of the will itself, all parties agreed, is impossible. For decisions or choices are by their very nature made willingly – they are by their very nature exercises of the will, not impositions contrary to it. Protestants 59

Thomas Pink

insisted that the absence of coactio is by itself enough to leave our decisions and actions self-determined and so base an imputation model of our responsibility. Contingency might remain a feature of the exercise of liberum arbitrium at the civil or not strictly moral level; but it is not essential to its character as a power to determine our decisions or choices for ourselves.

3

Hobbes against the imputation model and contingent causation

Hobbes’s most fundamental departure from the shared metaphysics of responsibility was to deny outright the imputation model of responsibility:‘[Why do we blame someone?] I answer because they please us not. I might ask him, whether blaming be any thing else but saying the thing blamed is ill or imperfect’.6 Just as we may criticize an ugly or leaky vase as bad without attributing to the vase any power over itself, so too we may criticize a human agent as bad, simply expressing our displeasure at the agent without attributing to them any special power over their agency. And that negative expression of displeasure is all that blame involves. Hobbes further denied the very possibility of contingent causation. For Hobbes all power in nature is ordinary efficient causation, operating by necessity like a brick hitting a window. If a cause has power sufficient to produce an outcome, and all the conditions necessary for the outcome’s production are met (there are no obstacles preventing the power’s successful operation) the cause must produce the outcome. To suppose that freedom is a power sufficient under one and the same circumstances to produce any one of a variety of alternative outcomes is, therefore, to commit ourselves to the contradictory absurdity of their simultaneous production. Referring to a ‘free’ or ‘contingent’ cause abusively as an ‘indetermination’, Hobbes observed: But that the indetermination can make it happen or not happen, is absurd; for indetermination maketh it equally to happen or not to happen; and therefore both; which is a contradiction. Therefore indetermination doth nothing, and whatsoever causes do, is necessary.7 Why did Hobbes deny contingent causation? Why, in his view, must a cause sufficient to produce an outcome operate of necessity to produce it? For there seems no obvious contradiction in the supposition that a cause might possess power sufficient to produce an outcome – but still not operate to produce it. Key to Hobbes’s denial of contingent causation was a reductive account of our conception of power. Power as we conceive it is no more than material motion that produces or prevents further motion. Hobbes expounded this reductive view of our conception of power in 1655 in De Corpore: the Power of the Agent is the same thing with the Efficient Cause. From which it may be understood, that all Active Power consists in Motion also; and that Power is not a certain Accident which differs from all Acts, but is indeed an Act, namely Motion, which is therefore called Power, because another Act shall be produced by it afterwards. For example, if of three Bodies the first put forwards the second, and this the third, the Motion of the second in respect of the first which produceth it, is the Act of the second Body, but in respect of the third it is the Active Power of the second Body.8

60

Hobbes against Bramhall

Correspondingly, rest or lack of motion is simply lack of power: ‘Rest does nothing at all, nor is of any efficacy; and . . . nothing but Motion gives Motion to such things as be at Rest, and takes it from things moved’.9 We have no conception of power without motion affecting motion – ‘Motion, which is therefore called Power, because another Act shall be produced by it afterwards’. But, very plainly, power operative contingently presupposes the possibility of a power to move unaccompanied by actual motion. On Hobbes’s view, then, we simply have no conception of contingency as the mode of operation of a power. By its very nature, we conceive power to operate by necessity. For power to be present and unimpeded just is for it to operate – to produce or prevent further motion. On the other hand we do still talk of outcomes as ‘contingent’. So, if not a mode of operation of a power, what conception does this use of the term ‘contingent’ express? In Hobbes’s view, the conception is entirely epistemic. We use ‘contingent’ for cases where either the outcome itself or at least its causes are still unknown to us: ‘For by contingent, men do not mean that which hath no cause, but which hath not for cause any thing which we perceive’.10 Freedom in the case of human action could not then be some distinctively contingent form of power. It could only be what Hobbes supposed it to be  – simply an absence of obstacles to the only possible form of power – to power operating by necessity, as the unimpeded force of an appetite of our will to cause us to act as willed.

4

Hobbes against self-determination

But then we meet a puzzle. Hobbes did sometimes write as if people altogether lack any conception of freedom of the kind proposed by Bramhall – as a distinctive kind of power operative through the will. Bramhall’s theory would then, strictly speaking, be no more than contentless verbiage. It would express no genuine conception at all. But that was not in fact Hobbes’s more considered view. Rather he allowed that we do have a conception of freedom or liberty as ‘productive’ – as a kind of power. We conceive liberty to be ‘productive’ just in those cases where we lack experience of some externally originating power operating on our will. Whenever we do experience the operation of such a power, we always understand it as moving us by necessity. It is where such experience is lacking that we suppose our actions to be produced not by necessitating causes, but by ‘liberty’. Commonly when we see and know the strength that moves us, we acknowledge necessity, but when we see not, or mark not the force that moves us, we then think there is none, and that it is not Causes but Liberty that produceth the action.11 But what is it to conceive of liberty as ‘productive’, if not as a power operating contingently? Hobbes could have followed the Protestant theologians and retained a conception of freedom as a power of self-determination that could operate through necessity. But he did not do this. He also denied the very possibility of self-determination. Hobbes viewed the claim that an agent could determine his own action as involving a vicious regress. He taunted Bramhall with the threat of it:‘And if a man determine himself, the question will still remain what determined him to determine himself in that manner’.12 But why did Hobbes suppose self-determination to be such a regressive idea? Why should ‘the question still remain what determined him to determine himself in that manner’? It is important that Hobbes identified the idea of self-determination not as common to

61

Thomas Pink

theologians generally (which in fact it was) but as supposedly peculiar to theology that was Catholic or allied to Catholicism, involving what Hobbes alleged to be a peculiarly Catholic view of the will. This took the will to be a determinant of action independent of other causes: But for some ages past, the Doctors of the Roman Church have exempted from this dominion of God’s Will, the Will of Man; and brought in a Doctrine that not onely Man, but also his Will is Free, and determined to this or that action, not by the Will of God, nor necessary causes, but by the power of the Will it Self. And though by the reformed Churches instructed by Luther, Calvin and others, this opinion was cast out, yet not many years since it began again to be reduced by Arminius and his followers, and became the readiest way to Ecclesiastical promotion.13 This view of the will was not in fact treated by Hobbes as entirely peculiar to a sinister party within theology. Just as he described all of us as naturally inclined, when unaware of the precise causes of our actions, to ascribe those actions to liberty as a productive force, so he also suggested a rather general tendency to ascribe those same actions to our own will operating as a cause independently of external causes. In replying to Bramhall’s assertion of freedom as implying an absence of antecedent necessity at the point of the will, Hobbes observed: the later I deny, and have shown that he ought to have proved that a man is Free to Wil. For that which he sayes, any thing else whatsoever, would think, if it knew it were moved, and did not know what moved it. A woodden Top that is lasht by the Boyes, and runs about sometimes to one Wall, sometimes to another, sometimes spinning, sometimes hitting men on the shins, if it were sensible of its own motion, would think it proceeded from its own Will, unless it felt what lasht it. And is a man any wiser, when he runns to one place for a Benefice, to another for a Bargain, and troubles the world with writing errors, and requiring answers, because he thinks he doth it without other cause than his own Will, and seeth not what are the lashings that cause his Will?14 For Hobbes, then, the idea of liberty as productive was closely associated with the idea of a will operating as an externally uncaused cause. Moreover the idea of self-determination, which Hobbes viewed as unrealisable because viciously regressive, involved for him this same idea of the will as externally uncaused cause. This is the key to understanding how, for Hobbes, we could ever come to think, however mistakenly, of liberty as something productive. The association of self-determination with a determination of action by the will alone brings in a vicious regress; but it also allows some content, even on Hobbes’s terms, to the claim that liberty is productive. How can we explain an agent’s coming to move in one way rather than another other than appealing to something external acting on the agent – other than by appeal to necessitation from without? The agent would have to have a power to move itself. But since as ‘liberty’ this power of the agent to move itself would have to be a power to move itself in multiple directions; and since in the absence of contingency as a mode of operation, if possessed unimpeded this power to move in multiple directions would absurdly have to operate in all these directions at once, something internal to the agent would have to block 62

Hobbes against Bramhall

its operation in all ways but one. For the agent to move one way something internal to the agent would have to suspend or block the operation of the agent’s equal power to move all other ways. The agent would have to exercise a further power over their power to move. Variation in the exercise of the agent’s power to move would have to be explained in terms of the operation of a further power over the power. Of course, the postulation of that further power to exercise the power would raise the same problems in turn, and so be viciously regressive. But it would at least allow content to the thought that variation in human action arises other than through an externally imposed necessity – a content that does not directly involve a conception of contingency as a power’s mode of operation, a conception of contingency that Hobbes very clearly denied we possessed. Liberty would instead be a power of an agent to move that depended on a prior power of the agent to exercise the power. Hobbes clearly viewed belief in an agency of will as viciously regressive. It was regressive as transforming a conception of a power to act into an impossible capacity to exercise power over how power was exercised, and in the form of a regressive sequence of willings and willings to will. In The Elements of Law Hobbes had already insisted, as a point very much worth making, that the will is not voluntary as are the actions willed that it explains: Appetite, fear, hope, and the rest of the passions are not called voluntary; for they proceed not from, but are the will; and the will is not voluntary. For a man can no more say he will will, than he will will will, and so make an infinite repetition of the word will; which is absurd, and insignificant.15 Hobbes saw the formulations both of willing to will and of self-determination as in themselves absurd and as implied by the erroneous psychology of liberum arbitrium. This was a theory that took the will both to exercise power and to be the immediate object of the same power’s exercise, so that the will determines its own operation. The mistake, in Hobbes’s view, was to treat an agent’s power to do something as if it were itself an agent. The will is only the agent’s power to act; but the theory of self-determination involves turning the will into a quasi-agent whose own power to act would in turn have to be provided by a further will – and so on ad infinitum: As if it were not Freedome enough for a man to do what he will, unless his will also have power over his Will, and that his will be not the power itself, but must have another power within it to do all voluntary acts.16 The will could no more perform actions than our power or capacity to dance could be supposed to go in for dancing: Secondly, you may observe, that actus elicitus, is exemplified by these words, to will, to elect, to choose, which are all one, and so to will here is made an act of the will; and indeed, as the will is a faculty or power of a man’s soul, so to will is an act of it, according to that power. But as it is absurdly said, that to dance is an act allowed or drawn by fair means out of the ability to dance; so it is also to say, that to will is an act allowed or drawn out of the power to will, which power is commonly called, the will.17 63

Thomas Pink

5

Hobbes against normative power

Hobbes also denied power in normative form and so denied too that rationality consists in being moved by any such power. Bramhall referred to motivation by a force of goodness as ‘moral determination’, by contrast to ‘natural’ determination by efficient causes. For Hobbes, by contrast, all power or ‘efficacy’ was ‘natural: ‘Natural efficacy of objects does determine voluntary Agents, and necessitates the Will, and consequently the Action; but for moral efficacy, I understand not what he means by it’.18 A more developed account of how ‘moral efficacy’ was to be excluded was given by Hobbes in his critical Examinatio of Thomas White’s De Mundo. Since, in Hobbes’s view, the production of anything, including any action, must be by an efficient cause, insofar as it did play a productive role final causation could amount to nothing more than efficient causation. A final cause in so far as it can be conceived by man, is entirely the same as an efficient cause, for from an agreeable object there arises an imagination of enjoyment, from an imagination of enjoyment arises an imagination of a way, from an imagination of a way there arises a motion towards the desired object; in which series of productions the object or end is an agent, for which reason the act of the object which is the end is an efficient cause of our motion towards the end.19 An end as genuinely productive could only be some sensorily presented object or at least a mental state of experiencing or imagining such, in either case something that did really exist prior to the action that it explained, and that could function as an efficient cause of our wanting to act. This of course was not final causation as understood by the scholastics. Hobbes’s end was always some already existent stimulus that triggered the pursuit of a goal – like an inviting bottle entering a visual field. But, as we have seen, the scholastics’ final cause was the potential goal, and so an as yet unrealised and purely mental object – such as that the bottle is painted by us as a still life, or its contents savoured, or that it is returned to the cupboard to be saved for the Christmas party. And for Hobbes this end as the as yet unattained goal of the action, so not the sensorily presented thing that causes desire but the desired outcome, was not a cause at all. It was simply something imagined, the content of the ‘imagination of enjoyment’, which was Hobbes’s description of a desire or appetite with that content. It was that desire alone, a psychological occurrence and so a possible efficient cause, that then gave rise to action, supplying through its content the action with its goal. But that content, what was desired – the goal and its goodness – played no causal role, either in the production of the desire itself or, through that desire, of the action desired. The goal itself was inert; only the desire for it was productive, a motion in matter and an ordinary cause. For Hobbes, by contrast, though we might talk of the objects of our appetites as being good, goodness was not an attractive force operating through those objects to explain those appetites. Far from explaining our appetites, our talk of their objects as good is expressive simply of the fact that those appetites are already present. For, what is it else to praise, but to say a thing is good? Good, I say for me, or for somebody else, or for the State and Commonwealth. And what is it to say an action is good, but to say, it is as I would wish, or as another would have it, or according to the will of the State, that is to say, according to Law?20 64

Hobbes against Bramhall

Hobbes attacked the possibility of what he termed a ‘metaphysical goodness’ that could explain or produce our psychological motivations or actions, and any associated doctrine that goodness was convertible with being: There hath been in the Schooles derived from Aristotles Metaphysicks an old Proverb rather than an Axiome, Ens, Bonum, et verum convertuntur [being, the good and the true are convertible]. From hence the Bishop hath taken this notion of a Metaphysical goodness, and his doctrine, that whatsoever hath a being is good; and by this interpreteth the words of Gen. 1. God saw all that he had made, and it was very good. But the reason of those words is that Good is relative to those that are pleased with it, and not of absolute signification to all men. God therefore saith that all that he had made was very good, because he was pleased with the Creatures of his own making.21 Hobbes’s allegiance to this account of value was consistent. He defended the same view of the relativity of good to appetite in Leviathan in a famous passage: But whatsoever is the object of any mans Appetite or Desire, that is it, which he for his part calleth Good; and the object of his Hate, and Aversion, Evill; and of his Contempt, Vile and Inconsiderable. For these words of Good, Evill, and Contemptible, are ever used with relation to the person that useth them: There being nothing simply and absolutely so; nor any common Rule of Good and Evill, to be taken from the nature of the objects themselves.22 The denial of motivation by ‘moral efficacy’, by a normative form of power, came with an account of agency and its psychology that further applied Hobbes’s rejection of a distinctive metaphysics of rational nature. Since there is no ‘moral determination’ of attitudes operating through intellectually presented objects, there can be no special class of motivations of the will that are peculiarly responsive to reason and its moral force: For I do not fear it will be thought too hot for my fingers, to shew the vanity of words such as these, Intellectual appetite, conformity of the appetite to the object, rational will, elective power of the rational will; nor understand I how reason can be the root of true liberty, if the Bishop (as he saith in the beginning) had the liberty to write this discourse. I  understand how objects, and the conveniences and inconveniences of them, may be represented to a man by the help of his senses; but how reason representeth anything to the will, I [do not] understand.23 Since there are no distinctively reason-responsive motivations, decisions or choices must be no more than a form of appetite or passion – the last appetite or passion in deliberating, the appetite that finally determines how we act. The motivation of action must then be reconceived. What moves an agent to act in pursuit of a goal? It can no longer be the goal itself as a mental object motivating a decision to attain it. Mental objects are now inert, and any motivating force belongs instead only to what can function as efficient causes, the motivating attitudes directed at those objects. So voluntary action, what we are moved to do willingly, begins not in our motivating attitudes themselves, but only when those attitudes as efficient causes move or produce actions as their effects. Voluntary action no longer occurs as an action of the will itself, but only as a willed effect of a prior will or appetite so to act. 65

Thomas Pink

Since all motivations are passive appetites, the idea of a freedom of will is as absurd as the idea of a freedom of passion, since the former idea reduces to the second, ‘nor can a man more determine his will than any other appetite; that is, more than he can determine when he shall be hungry and when not’.24 Hobbes thereby assimilated agency to what scholasticism had regarded only as a secondary form of agency – the willed actions that the will produced. Hobbes was quite clear: voluntary acts did not occur decisions or choices, what scholastics termed the ‘elicited’ actions of the will itself, but were simply what scholastics had termed ‘commanded’ acts – though Hobbes had serious reservations about the use here of terms such as commanded’ or ‘imperatus’: Wherein letting pass that Metaphoricall speech of attributing command and subjection to the faculties of the soul, as if they made a commonwealth or family among themselves, and could speak to one another, which is very improper in searching the truth of this question; you may observe first that to compell a voluntary act, is nothing else, but to will it; for it is all one to say, my will commands the shutting of mine eyes, or the doing of any other action, and to say, I have the will to shut my eyes. So that actus imperatus here, might as easily have been said in English, a voluntary action, but that they that invented the term, understood not anything it signified.25 Thus we see that Hobbes’s restriction of agency to willed or commanded action, with denial of an elicited agency of the will itself, followed from each of two distinct lines of argument. The first was Hobbes’s sceptical attack on a power of self-determination, an elicited agency of choice or decision being the supposed medium for its exercise. The second was Hobbes’s attack on a normative power of goodness exercised by mental objects, a force productive of attitudes to those objects with elicited actions of choice or decision being the supposed form of agency immediately responsive to its motivating force. Neither line of argument nor their conclusions were to be found in the Protestant theologians whom Hobbes invoked in his own support.

Notes 1 The Questions Concerning Liberty, Necessity and Chance, clearly stated between Dr Bramhall Bishop of Derry, and Thomas Hobbes of Malmesbury was published by Hobbes in London in 1656. A private debate between Hobbes and Bramhall about free will in Paris in 1645 had been followed by a series of opposing publications. Questions contains Hobbes’s initial published account of free will, Of Libertie and Necessitie, with an interweaving of Bramhall’s subsequent responses and then further replies by Hobbes. 2 Thomas Aquinas Summa Theologiae 1, 2 q21 a 2, resp, ed. Pietro Caramello (Rome: Marietti,1950) p112. 3 Francisco Suarez, Disputationes Metaphysicae, disp 23, sect 4, §8, ed. Charles Berton, Opera vol. 25 (Paris: Vives, 1866) p861. 4 Suarez, Disputationes Metaphysicae, Disp 19 sect 1, §12, ed. Charles Berton, Opera vol. 25 (Paris: Vives, 1866) p692. 5 For Reformation theology that (very typically) assumed this shared metaphysics, consider the widely read and cited work of the eminent Girolamo Zanchi, a convert to reformed Protestantism who held a chair at Heidelberg in the Calvinist Palatinate – see his Tractationum Theologicarum Volumen (Neustadt 1603) and Opera Theologica (Geneva 1619). Hobbes invoked Zanchi in the Questions as one important Protestant authority for the consistency of freedom with necessity; but then he explained this consistency by a radically different metaphysics. 6 Hobbes Questions p39.

66

Hobbes against Bramhall

Further reading Pink, Thomas. (2004). “Suarez, Hobbes and the Scholastic Tradition in Action Theory,” in Thomas Pink  & Martin Stone (eds.), The Will and Human Action: From Antiquity to the Present Day (pp. 127–153). London: Routledge gives a detailed account of the very differing theories of human agency given by scholastic Aristotelians and Thomas Hobbes. Pink, Thomas. (2018). “Agents, Objects and Their Powers in Suarez and Hobbes,” in Constantine Sandis (ed.), Philosophy of Action from Suarez to Anscombe (pp. 3–24). London: Routledge gives a detailed account of the very differing theories of power proposed by scholastic Aristotelians and Thomas Hobbes. Tuck, Richard. (1989). Hobbes. Oxford: Oxford University Press gives a general account of Hobbes’s life and work.

References Aquinas, Thomas. (1950). Summa Theologiae, ed. Pietro Caramello. Rome: Marietti. Hobbes, Thomas. (1950). Critique du De Mundo de Thomas White, eds. J. Jacquot & H. W. Jones. Paris: Vrin. ———. (1999). De Corpore, ed. Karl Schuhmann. Paris: Vrin. ———. (2012). Leviathan, ed. Noel Malcolm. Oxford: Clarendon Press. ———. (1969). Elements of Law, ed. Ferdinand Tönnies. London: Cass. Hobbes, Thomas,  & Bramhall, John. (1656). The Questions Concerning Liberty, Necessity and Chance, Clearly Stated Between Dr Bramhall Bishop of Derry, and Thomas Hobbes of Malmesbury. London. Suarez, Francisco. (1866). Disputationes Metaphysicae, ed. Charles Berton, Opera vol. 25. Paris: Vives. Zanchi, Girolamo. (1603). Tractationum Theologicarum Volumen. Neustadt. ———. (1619). Opera Theologica. Geneva.

67

5 HUME ON FREE WILL AND MORAL RESPONSIBILITY Peter Millican

David Hume’s views on morality  – the topic of both Book  3 of his Treatise of Human Nature (1740), and of his Enquiry concerning the Principles of Morals (1751) – are widely familiar and much discussed, but his view of moral responsibility is far less well known, and even rather obscure. To piece this view together, we must examine his theory of intentional agency, which he expounds under the heading “Of Liberty and Necessity”. But there are significant discrepancies between his discussions of this topic in Book 2 of the Treatise of Human Nature (1739) and the Enquiry concerning Human Understanding (1748). And his theory has been widely misunderstood for a number of reasons, including confusion about his theory of causal necessity and a tendency to view him through the lens of subsequent writers in the empiricist tradition. So although Hume’s writings “of liberty and necessity” are amongst the most widely known in the philosophical canon, achieving a reliable interpretation of his settled opinion on these topics – and on the closely related topic of moral responsibility – requires careful analysis.

1

Hume’s theory of causation

Hume famously analyses causal relations in terms of mere “constant conjunction” between types of event, where the cause is regularly followed by the effect.1 Many of his readers down the years have interpreted this theory as implicitly denying that there is any genuine causal necessity between a cause and its effect, thus perhaps suggesting skepticism about causal relations themselves. But such readings are incorrect if we understand the relevant terms in accordance with his own theory. First, Hume clearly believes in causal relations as understood on the regularity account, and indeed frames explicit rules to identify them: it may be proper to fix some general rules, by which we may know when [objects] really are . . . causes or effects to each other (T 1.3.15.2, cf. T 1.3.13.11) all objects, which are found to be constantly conjoin’d, are upon that account only to be regarded as causes and effects (T 1.4.5.32)

DOI: 10.4324/9781003282242-9

68

Hume on free will and moral responsibility

Secondly, Hume insists – both before and after developing this theory – that necessity is essential to causation. Indeed, his long investigation into the origin (and hence nature) of our “idea of necessary connexion” takes off from this premise, when he asks what more is needed for the causal relation to hold between two “objects”, besides (single-case) contiguity and temporal priority: An object may be contiguous and prior to another, without being consider’d as its cause. There is a necessary connexion to be taken into consideration; and that relation is of much greater importance, than any of the other two above-mention’d. (T 1.3.2.11) To achieve a clearer understanding of this key idea, Hume employs what is commonly known as his Copy Principle, for which he has previously argued (in T 1.1.1 and EHU 2): It seems a proposition, which will not admit of much dispute, that all our ideas are nothing but copies of our impressions, or, in other words, that it is impossible for us to think of any thing, which we have not antecedently felt, either by our external or internal senses. (EHU 7.4, cf. T 1.3.14.1) This is the core of his conceptual empiricism, whereby all the contents of our thoughts are derived either from external sense experience (sight, touch, hearing etc.), or from internal “reflection” (feelings, emotions, etc.). Hume eventually (T  1.3.14.20, EHU  7.28) tracks down the “impression of necessary connexion” to something that we understand through reflection, namely customary transition of the mind from observed cause A to expected effect B, in response to repeated observation of As followed by Bs  – what we now call inductive inference. His declared success in this quest clearly indicates that he takes the corresponding idea to be legitimate, since it has a genuine – though surprising – impressionsource, rather than being a confused “fiction”.2 Hume then frames on this basis two “definitions of cause”, the first couched in terms of constant conjunction and the second in terms of mental inference: Similar objects are always conjoined with similar. Of this we have experience. Suitably to this experience, therefore, we may define a cause to be an object, followed by another, and where all the objects, similar to the first, are followed by objects similar to the second. . .. The appearance of a cause always conveys the mind, by a customary transition, to the idea of the effect. Of this also we have experience. We may, therefore, suitably to this experience, form another definition of cause; and call it, an object followed by another, and whose appearance always conveys the thought to that other. (EHU 7.29, cf. T 1.3.14.31) The prominent role of mental inference in these discussions has led many interpreters to presume that he is a subjectivist about causal necessity (in the sense of considering such necessity to be observer-relative), and hence denies it as an objective factor within causal relations. But this is to ignore that in both works – and almost immediately in the

69

Peter Millican

Enquiry – Hume goes on to provide two parallel “definitions of necessity”, again in terms of constant conjunction and mental inference. These serve to vindicate his claim that necessity is essential to causality, on both definitions: According to my definitions, necessity makes an essential part of causation (T 2.3.1.18, cf. EHU 8.25) Necessity may be defined two ways, conformably to the two definitions of cause, of which it makes an essential part. (EHU 8.27, cf. T 2.3.2.4) So Hume is neither skeptical about causal relations nor about causal necessity, as long as these are understood in accordance with his definitions. Let us now go on to see how Hume applies these definitions in arguing for the deterministic thesis which he calls “the doctrine of necessity”.

2  The “doctrine of necessity” Hume’s discussions entitled “Of liberty and necessity”  – in both Treatise 2.3.1‑2 and Enquiry 8 – are largely devoted to establishing what he calls the doctrine of necessity in the sphere of human behavior. This doctrine states that all such behavior is subject to causal necessity. On Hume’s view as explained previously, such causal necessity is simply a matter of events’ occurring in conformity with deterministic causal laws, from which, therefore, those events could be predicted at least in principle (i.e., by a hypothetical being who knows all the relevant initial conditions and laws, and is able to calculate their workings in detail). But previous philosophical orthodoxy, as represented by the influential Newtonian Samuel Clarke, took causal necessity to involve more than mere conformity with predictive laws. For Clarke, real necessity is to be understood on the model of mechanical impulse – when one body in motion smashes into another and forces it to move – so it is a mistake to think of intelligent agents – acting on reasons or motives – as subject to such necessity. We might call such agents’ behavior “morally necessary” in the sense of being entirely predictable in principle (i.e., the minimal deterministic “necessity” that Hume endorses), but if so, that is “merely a figurative Manner of Speaking” and “not indeed any Necessity at all” (Clarke 1717, 15–17).3 One of the main aims of Hume’s discussions “Of liberty and necessity” is to undermine this alleged distinction between physical and moral necessity, by arguing that our only possible understanding of causal necessity is in terms of his two definitions, and that these apply equally to the physical and moral (i.e., human) worlds. So Hume’s primary motivation in applying his Copy Principle to track down the “impression of necessary connexion” seems to be to circumscribe the limits of what we can possibly mean by causal necessity. On this basis, he insists that those such as Clarke who claim that “there is something else in the operations of matter” – some form of supposed genuine necessity that goes beyond anything in human behavior – are using terms that have no corresponding idea, and are therefore “unintelligible” (T 2.3.2.4, cf. A 34, EHU 8.22).4 Hume’s argument for the doctrine of necessity (i.e., determinism) within the human sphere aims first to establish and then build on these conceptual claims. He starts from what he takes to be the “universally” agreed predictability of “the operations of external bodies”, in which “Every object is determin’d by an absolute fate to a certain degree and 70

Hume on free will and moral responsibility

direction of its motion” with “not the least traces of indifference or liberty” (T 2.3.1.3), and “every natural effect is so precisely determined by the energy of its cause, that no other effect, in such particular circumstances, could possibly have resulted from it” (EHU 8.4). Then he refers back to his two definitions of cause, and draws from them two parallel criteria of causal necessity, which he aims to prove apply to the human world as much as to the physical: “Here then are two particulars, which we are to consider as essential to necessity, viz. the constant union and the inference of the mind; and wherever we discover these we must acknowledge a necessity” (T 2.3.1.4). The version in the Enquiry is more explicit about the conceptual or semantic aspect of this argument – that it hinges on questions of meaning as determined by the limits of our ideas: Our idea, therefore, of necessity and causation arises entirely from the uniformity, observable in the operations of nature; where similar objects are constantly conjoined together, and the mind is determined by custom to infer the one from the appearance of the other. These two circumstances form the whole of that necessity, which we ascribe to matter. Beyond the constant conjunction of similar objects, and the consequent inference from one to the other, we have no notion of any necessity, or connexion. (EHU 8.5) An entry in Hume’s own index to the Enquiry – “Necessity, its definition” – makes explicit that he understands this last statement as defining necessity (citing also the passage from EHU 8.27 quoted earlier). The index has a quite separate entry for “Cause and Effect . . . Its Definition”, referring to the two definitions of cause at EHU 7.29 (and the footnote at EHU 8.25). So Hume clearly distinguishes his definitions of necessity from these earlier and more famous two definitions of cause, though they are obviously closely related. As we have seen, Hume casts this argument in semantic terms – as restricting what we can possibly understand or mean by causal necessity – in order to undermine the orthodox claim that “moral” and “physical” causation are of a different nature. But he also implicitly appeals to this semantic theme when attempting to dissolve the debate as having turned hitherto on a misunderstanding: If it appear, therefore, that all mankind have ever allowed, without any doubt or hesitation, that these two circumstances take place in the voluntary actions of men, and in the operations of mind; it must follow, that all mankind have ever agreed in the doctrine of necessity, and that they have hitherto disputed, merely for not understanding each other. (EHU 8.5‑6) This attempted dissolution might seem a bit far-fetched with regard to “the most contentious question, of metaphysics, the most contentious science” (EHU 8.23), but Hume suggests that the debate’s very intractability is indicative of some persisting misunderstanding (EHU 8.1‑3). The bulk of Hume’s subsequent argument, however, appeals to considerations that seem observational more than semantic, because they emphasize not so much people’s assumptions about the regularity of human behavior – and the fact that they make inferences accordingly – but rather, that such regularity actually obtains, so that such inference is in fact reliable. Accordingly, the first part of his strategy is to “prove from experience that our actions have a constant union with our motives, tempers, and circumstances” (T 2.3.1.4, cf. EHU 8.7). 71

Peter Millican

After giving a variety of examples to support and illustrate this claim (T  2.3.1.5‑10, EHU  8.7‑15), Hume concludes “that the conjunction between motives and voluntary actions is as regular and uniform, as that between the cause and effect in any part of nature” (EHU 8.16, cf. T 2.3.1.12). In the Enquiry, he also at this point claims to have shown “that this regular conjunction has been universally acknowledged among mankind”, in which case “it may seem superfluous to prove, that this experienced uniformity in human actions is a source whence we draw inferences concerning them” (EHU 8.16), though he goes on to illustrate the latter anyway (EHU 8.17‑20). In the Treatise, he separates out this last task of showing that as the union betwixt motives and actions has the same constancy, as that in any natural operations, so its influence on the understanding is also the same, in determining us to infer the existence of one from that of another (T 2.3.1.14). His argument for this mixes illustration (T 2.3.1.15, 17) with appeal to his theory of necessity (T 2.3.1.16, 18).

3  Determinism, chance, and some reservations Hume’s argument for “the doctrine of necessity” is by no means compelling. The complexity of the human brain and associated faculties (both sensory and motor) render it hopelessly unfeasible to establish, in the sphere of human behavior, that there are “not the least traces of indifference”, or that “every effect is so precisely determined, that no other effect, in such particular circumstances, could possibly have resulted” (cf. T 2.3.1.3 and EHU 8.4, as quoted previously). Even in the physical world, such claims are dubious (as has been emphasized by quantum mechanics), but at least there it is possible to test and measure isolated systems with relatively great accuracy, thus giving a plausible basis for extrapolation to more complex phenomena. But there is obviously no way of even beginning to study human thought and behavior in isolation from the unfathomable complexity of our bodies and our environment (both physical and social). Determinism about the mind and behavior may be plausible, but support for it is likely to come less from direct experimental confirmation than from more general considerations – and perhaps prejudices – such as the desire for explanatory completeness (thus ruling out randomness), the belief that mental phenomena are dependent on a deterministic physical substrate (as strongly suggested by evolutionary considerations), and the conceptual difficulty of envisaging a non-random alternative to causal explanation.5 In the 18th century, following Newton’s impressive achievements, physical determinism was widely taken for granted. Those inclined to materialism, such as Hobbes and Hume, would naturally extend determinism equally to the “moral” sphere of human behavior, on the basis that humans – like everything else in the world – are purely material beings. But Christians such as Clarke, believing in a realm of immaterial spirits, typically considered mental causation to operate quite differently. Both sides, however, were equally committed to the Causal Maxim discussed in Treatise 1.3.3, that “whatever begins to exist, must have a cause of existence”. And the Maxim was also commonly understood as extending to any change, not just new beginnings of existence. So both materialists and Christians were typically committed to universal causation, but differed in respect of the kinds of causation they countenanced, and in particular, whether human 72

Hume on free will and moral responsibility

behavior involved purposive causation by spirits (as opposed to mechanical causation by bodies). Neither of these sides would typically understand causation as involving randomness, because while materialists tended to attribute all causation to deterministic matter, Christians opposed the “Epicurean” hypothesis of chance as dangerously atheistic. Hume was therefore reflecting this consensus when he reported as “commonly allow’d by philosophers, that what the vulgar call chance is nothing by a secret and conceal’d cause” (T 1.3.12.1).6 All this helps to explain why Hume’s case for the “doctrine of necessity” – understood as equivalent to universal determinism – can seem to us rather vague and complacent. At best it shows that determinism about human behavior might plausibly be true, because just as we presume that unpredictable changes in our health (EHU 8.14) and the weather (EHU 8.15) are explicable by underlying physical causes (even when these causes may be unknown, and are anyway far too complex for us to work out in detail); so “The philosopher, if he be consistent, must apply the same reasoning to the actions and volitions of intelligent agents” (EHU 8.15). In the case of the physical world, however, Hume is able to offer rather more, because here, not only can “the philosopher” reasonably consider it “at least possible [that] the contrariety of events [proceeds] . . . from the secret operation of contrary causes”; but also, “This possibility [may be] converted into certainty by farther observation; when . . . upon an exact scrutiny, a contrariety of effects always betrays a contrariety of causes”. Hume gives the example of a clock or watch that stops unexpectedly, and an artisan who, on examining it, finds a grain of dust interfering with the mechanism. This discussion is the closest that Hume comes to giving solid evidence for the truth of determinism. But there is no such relative solidity to be had in the world of mind and behavior, where hidden mechanisms cannot be opened up and examined in detail. So again, Hume’s case looks less than totally convincing. The same can be said for his claim “that the conjunction between motives and voluntary actions is as regular and uniform, as that between the cause and effect in any part of nature” (EHU 8.16). For motives result in actions only through complex causal chains involving bodies and other physical things, with many potential obstacles on the way. Indeed, Hume had observed in just the previous paragraph how physical circumstances can impact on us, potentially affecting both our motives and our ability to act on them:7 “A  person of an obliging disposition gives a peevish answer: But he has the toothake, or has not dined” (EHU 8.15). Hence his specific focus on “the conjunction between motives and voluntary actions” seems to be a slip (cf. Botterill 2002: 285–286). His overall view would leave him free to acknowledge the complex interplay between mental and physical, and to expect deterministic patterns across the entire system, but not necessarily within the “moral” sphere considered separately.

4  Introducing “hypothetical liberty” In his discussion “Of liberty and necessity” in Treatise 2.3.1‑2, Hume mostly uses the term “liberty” not as applying specifically to human action, but as meaning simply chance or absence of necessity (e.g. T  2.3.2.2, 2.3.2.6‑7)  – something which, as we have seen, he rejects. He there refers to “free-will” only once, equating it with both “liberty” and “chance”, and accordingly denying its existence also (T 2.3.1.18). This might suggest that Hume takes a “hard determinist” position, denying moral freedom on the basis that we are causally determined. But on the contrary, at T 2.3.2.5‑7 he argues that his deterministic 73

Peter Millican

position, so far from undermining morality, is essential to it; then in Book 3 he goes on to develop in detail his own positive theory of morality. When he came to write the first Enquiry, Hume corrected the false impression of hard determinism by endorsing “liberty”, having defined it quite differently from his usage in the Treatise, and as entirely consistent with determinism. By liberty . . . we can only mean a power of acting or not acting, according to the determinations of the will; that is, if we chuse to remain at rest, we may; if we chuse to move, we also may. Now this hypothetical liberty is universally allowed to belong to every one, who is not a prisoner and in chains. (EHU 8.23) The same paragraph famously starts by describing this discussion as a “reconciling project with regard to the question of liberty and necessity”, so Hume’s compatibilist intention is entirely explicit. But the precise nature of his compatibilist theory is not so clear. Looking back on some occasion when I have moved or stayed at rest – to follow Hume’s example – a minimal interpretation of his “hypothetical liberty” might simply require that the act I performed was in fact in accordance with my will: Either I willed to move, and did move; or I willed to stay at rest, and did stay at rest. But a more fulsome interpretation would require also a counterfactual condition, such as: If I had willed differently, then my action would have been correspondingly different. It might be suggested that Hume ought to favor the minimal interpretation, on the basis that his determinism would rule out the counterfactual possibility of my having willed differently, or to respect the puzzling T 1.3.14.34 corollary to his definitions of cause, that “The distinction, which we often make betwixt power and the exercise of it, is . . . without foundation” (cf. Penelhum 2000: 162). But determinists are generally happy with counterfactuals corresponding to causal laws, while Hume’s commitment to his implausible corollary weakens when he considers our passions and probabilistic thinking (T 2.1.10.4, 6, 10; 2.2.5.7). Indeed the corollary is absent from the Enquiry, where Hume’s chosen term – “hypothetical liberty” – strongly suggests a counterfactual interpretation. Moreover his choice of example here  – moving or not moving  – is suggestively reminiscent of the discussion in Locke’s Essay, which seems explicitly to require the two-way (counterfactual) power of the more fulsome interpretation: so far as a Man has a power to think, or not to think; to move, or not to move, according to the preference or direction of his own mind, so far is a Man Free. Whereever any performance or forbearance are not equally in a Man’s power; where-ever doing or not doing, will not equally follow upon the preference of his mind directing it, there he is not Free, though perhaps the Action may be voluntary. (Essay II xxi 8) Given Hume’s Lockean background, and “the absence of any Humean signals to the contrary”, Bricke suggests that “it is reasonable to read Hume’s doctrine in the standard . . . way”, as involving two-way “implementability of relevant alternatives” (2008: 208). This 74

Hume on free will and moral responsibility

seems right, but before confidently drawing this conclusion, we need to examine some further textual complications.

5  “Liberty of spontaneity” and a puzzling mismatch In Treatise 2.3.2, Hume attempts to explain why people are falsely tempted to think of their actions as not determined, and argues against the concern that determinism would undermine morality. Much of this discussion is reproduced in Enquiry 8 (21‑2 and 26‑31), but the best-known passage is confined to the Treatise: Few are capable of distinguishing betwixt the liberty of spontaneity, as it is call’d in the schools, and the liberty of indifference; betwixt that which is oppos’d to violence, and that which means a negation of necessity and causes. The first is even the most common sense of the word; and as ’tis only that species of liberty, which it concerns us to preserve, our thoughts have been principally turn’d towards it (T 2.3.2.1) Here Hume gives a clear impression of endorsing the “liberty of spontaneity” as his favored account of free will, though the term does not appear at all in the Enquiry. This quite strongly suggests that it is another name for his “hypothetical liberty”, especially when both are contrasted with just one other sense of “liberty”, namely indifference or chance, and in the light of Hume’s emphatic statement at EHU 8.23 that if we speak of “liberty” in any reasonable sense, “we can only mean” his hypothetical liberty. But the first sentence of this initial paragraph in the Treatise significantly muddies the water: “I believe we may assign the three following reasons for the prevalence of the doctrine of liberty, however absurd it may be in one sense, and unintelligible in any other” (T 2.3.2.1). This seems to be describing the doctrine that we have “liberty” as absurd if that is understood as indifference, and unintelligible otherwise. So Hume apparently considers “liberty of spontaneity” to be unintelligible, even though he then goes on to refer to it as “the most common sense of the word; and . . . that species of liberty, which it concerns us to preserve”. But how can it concern us to preserve something that is unintelligible? And if, on the other hand, “liberty of spontaneity” is indeed intelligible, then isn’t it perverse that elsewhere in the Treatise Hume prefers to use “liberty” in something other than its “most common sense”, and indeed another sense that he himself calls “absurd”? But putting aside any such perversity, and assuming that the liberty of spontaneity of the Treatise is indeed intelligible, does this really equate with the hypothetical liberty of the Enquiry?8 Most previous scholars have indeed affirmed this identification, but it seems hard to square with Hume’s statements that his hypothetical liberty is “universally allowed to belong to every one, who is not a prisoner and in chains”, whereas by contrast spontaneity is “that species of liberty, which it concerns us to preserve” and which “our thoughts [are] principally turn’d towards” (thus suggesting that it is easily lost, and very far from universal). People anxious about their freedom are most commonly thinking about such things as political, religious, social or family pressures that constrain what they can do for fear of punishment or opprobrium; but they are relatively rarely in danger of being chained up in prison. If, for example, a gang boss threatens that my factory will be burned down if I continue to compete with his business, or a community elder threatens me with social ostracism if I continue my relationship with some outsider, then I am likely to be seriously concerned by this impairment of my freedom. But neither of these apparently deprives me 75

Peter Millican

of the minimal “hypothetical liberty” explicated in the Enquiry, which seems to amount to little more than a capacity for voluntary agency: to apply my will in whatever situation I am placed, with all its implied consequences and constraints. For in these situations, if I choose to do as I have been ordered, I may; and if I choose to defy the instruction (and take the consequences), I also may. So apparently I have “hypothetical liberty”, but not the sort of freedom “which it concerns me to preserve”. Which, then, is the genuine Humean notion?

6  A traditional misreading, and a resolution of the puzzle I suspect that previous scholars have often viewed Hume through the lens of later thinkers in the empiricist compatibilist tradition, such as A. J. Ayer in his classic essay “Freedom and Necessity”: it is not, I think, causality that freedom is to be contrasted with, but constraint. . .. If I am constrained, I do not act freely. . .. An obvious instance is the case in which I am compelled by another person to do what he wants . . . the compulsion need not be such as to deprive one of the power of choice. . .. [But] if . . . no reasonable person would be expected to choose the other alternative, then the action that I am made to do is not one for which I am held to be morally responsible. (Ayer 1954: 278‑9) Ayer’s identification of freedom with lack of constraint (rather than lack of causality) seems to echo Hume: the idea of necessity [seems] to imply something of force, and violence, and constraint, of which we are not sensible (T 2.3.2.1) if the definition [of cause] above mentioned be admitted; liberty, when opposed to necessity, not to constraint, is the same thing with chance; which is universally allowed to have no existence (EHU 8.25) The first of these passages immediately precedes Hume’s mention of liberty of spontaneity in the Treatise, which as we saw earlier, talks of such liberty as “oppos’d to violence”. The second passage appears, in a similar spirit, to be distinguishing between liberty of indifference – as “opposed to necessity” – and Hume’s hypothetical liberty – as “opposed . . . to constraint”. There is, however, a very different way of interpreting Hume’s language, which avoids the puzzling mismatch between “spontaneity” understood in Ayer’s way and “hypothetical liberty” as described in the Enquiry. This alternative interpretation is strongly supported by a later passage in the Enquiry, where Hume explains why the latter notion is appropriate to moral appraisal: It will be equally easy to prove, and from the same arguments, that liberty, according to that definition above mentioned, in which all men agree, is also essential to morality, and that no human actions, where it is wanting, are susceptible of any moral qualities, or can be the objects either of approbation or dislike. For as actions are objects of our moral sentiment, so far only as they are indications of the internal character, passions, and affections; it is impossible that they can give rise either to praise or 76

Hume on free will and moral responsibility

blame, where they proceed not from these principles, but are derived altogether from external violence. (EHU 8.31) An action that is “derived altogether from external violence” (e.g., being thrown downstairs by thugs) is quite different from an action performed from fear of violence (e.g., complying with a gang boss’s commands for fear of being thrown downstairs). And if Hume’s talk of “violence” in these contexts is thus understood, then this in turn suggests that when he speaks of “force, and violence, and constraint” in the Treatise (T  2.3.2.1), he has in mind physical force, violence, and constraint – which fits precisely with his reference in the Enquiry to the prisoner in chains. I conclude, therefore, that whereas Ayer’s paradigm of “constraint” is coercion by threat, Hume’s paradigm is physical restraint such as by chains, or a straitjacket, or physical force. Understood in this way, Hume’s “liberty of spontaneity” as “oppos’d to violence” (T 2.3.2.1) fits closely with “hypothetical liberty” as “universally allowed to belong to every one, who is not a prisoner and in chains” (EHU 8.23).

7  Two-way liberty and what we seek to preserve If the sorts of constraints that Hume has in mind as thwarting hypothetical liberty are indeed physical constraints such as prison walls and chains, then this gives extra support to the two-way or counterfactual interpretation of Hume’s “hypothetical liberty” that was tentatively supported in section 4. When thus constrained, details about the causal operations of our decision-making mental processes seem beside the point: our liberty is undermined in a far more straightforward way. So I think Hume would agree with Locke (Essay II xxi 10) that a man who is locked in a room where he wishes to stay (owing to the “desirable Company”), may be staying voluntarily – at least if he is unaware of the locked door – but he is not free with regard to staying or leaving. Regardless of the operation of his will, or whatever might determine it, he does not physically have the power to leave. Of course the man in the locked room is still free with regard to other choices (e.g., whether or not to speak, to move, or to attempt to leave). So this brings to light an issue implicit in Hume’s definition of hypothetical liberty, but not explicitly noted by him (nor by most commentators); namely, that the “power of acting or not acting” can be understood as relative to the choice in question.9 That being so, it might not after all be unreasonable to see this kind of liberty as one that we are anxious to preserve, not just minimally (e.g., moving our finger or not), but with as much scope as possible, so that we have the power to do, or not to do, a multitude of different things. In this way, our desire to remain out of prison and unchained can indeed be seen as concerned with maximizing our “power of acting or not acting”. If Hume was thinking along these lines, then he might after all have had the hypothetical liberty of the Enquiry in mind when, in the Treatise, he described liberty of spontaneity as “that species of liberty, which it concerns us to preserve”. This would yield an interpretation that neatly reconciles Hume’s favored notions of liberty in the two works.

8  Moral responsibility Just one passage in Hume’s philosophical writings – at T 2.3.2.6 – talks about the conditions for responsibility using that very term.10 But the points it makes are echoed elsewhere (notably EHU 8.29‑30), and follow familiar themes in his theory of moral 77

Peter Millican

appraisal, which focuses on enduring qualities of mind in the agent that may be judged as either virtues or vices: The constant and universal object of hatred or anger is a person or creature endow’d with thought and consciousness; and when any criminal or injurious actions excite that passion, ’tis only by their relation to the person or connexion with him. But according to the doctrine of liberty or chance, this connexion is reduc’d to nothing. . . . Actions are by their very nature temporary and perishing; and where they proceed not from some cause in the characters and disposition of the person, who perform’d them, they . . . can neither redound to his honour, if good, nor infamy, if evil. The action itself may be blameable. . . . But the person is not responsible for it; and as it proceeded from nothing in him, that is durable or constant . . .’tis impossible he can, upon its account, become the object of punishment or vengeance. ’Tis only upon the principles of necessity, that a person acquires any merit or demerit from his actions, however the common opinion may incline to the contrary.11 (T 2.3.2.6, emphasis added) According to Hume’s virtue-ethical theory, we judge behavior morally in terms of the qualities of mind or character that the relevant actions evince, rather than the type of action (T 3.3.1.4, EPM 1.10). And we judge these qualities of mind by their general tendencies rather than specific consequences (T 3.3.1.30, EPM 9.12). In its ultimate form in the second Enquiry, this constitutes something like a utilitarian virtue ethics,12 with reason playing a crucial role in assessing these general tendencies of characters, while sentiment generates our approval of beneficial outcomes and disapproval of bad (EPM  Appx 1.2‑3). In the Treatise, by contrast, Hume downplays the role of reason, and devotes far more attention to our moral sentiments, explaining how the mechanism of sympathy plays a crucial role by causing the ideas that we have of other people’s pains and pleasures to become enlivened into impressions that we ourselves feel (T 2.1.11.2‑7).13 This mechanism is non-rational – just part of our natural psychology – as is our tendency to feel distinctive moral sentiments of approval and disapproval when we consider certain characters of mind and their good or bad tendencies from a “general” view (T 3.3.1.23, 27‑30). Many scholars – notably Paul Russell in his influential 1995 book Freedom & Moral Sentiment – have accordingly seen the key to Humean moral responsibility as residing in these natural moral sentiments. But there is an important distinction between moral responsibility and moral culpability. Suppose that an intentional action exhibits some quality of mind – maybe it is kind, or sadistic, or cautious. If the action was kind or sadistic, then we would appropriately judge it (respectively) as virtuous or vicious, but if the action was cautious, then our judgment might well be morally neutral. Though some people are indeed more cautious than others, this “quality of mind” is neither an obvious virtue nor an obvious vice: in different contexts, different degrees of caution can be appropriate. But judging an action as morally neutral does not remove the agent’s responsibility; it simply means that their responsibility for the action brings them neither credit nor discredit. In other words, it is one thing to be responsible for an action; it is quite another to be morally praiseworthy or morally culpable for doing that action. Moral assessment of an action presupposes responsibility, but an agent can be responsible without being either praiseworthy or culpable. I suggest that Hume’s account of hypothetical liberty is intended to explain what it is to be responsible for an action, whereas his account of moral sentiments is intended to explain what it is for a responsible action to have a moral valence, either positive or negative. I am 78

Hume on free will and moral responsibility

free in Hume’s sense with respect to action A if my situation is such that, whether I choose to do A or to refrain from A, I  shall be able to carry through that choice. The crucial “determination” here is in the causal chain between my volition and the action (rather than the causation of my volition), so whether A occurs or not is suitably determined by my decision, and faithfully reflects it. Thus I am the voluntary author of what happens, straightforwardly responsible for it because my volition was a determining factor in its causation. And if that volition reflects an enduring feature of my character, then the action will also reflect my enduring character, thus providing an appropriate basis for morally judging me if the character thus revealed is virtuous or vicious.14 In such judgement, the moral sentiments indeed play a crucial role (as Russell emphasizes), but they are not crucial for assessing whether or not I am responsible; only whether the action would redound to my credit or my discredit. This division of labor makes good sense of Hume’s text, and explains why he says so little to connect moral sentiments with responsibility.15 It also provides a much cleaner way of analysing some cases, as compared with Ayer’s very broad appeal to “constraint” which treats coercion as a paradigm. For if a gangster threatens me with a gun and orders me to apply my thumb to open my employer’s digital cash register (say), then this need not undermine my ability to think rationally and decide what to do in the given situation. Ayer would say that my moral freedom is lost, but a better Humean response is to say that my opening of the cash register remains morally free (as long as the gangster is not physically forcing my thumb against the detector), because it is under my voluntary control. Both Ayer and Hume can agree that I would not be morally culpable for this action, but whereas Ayer would attribute this to a lack of moral freedom, Hume’s response is more straightforward: in that situation, complying with the order was the right (or at least permissible) thing to do. In this way, Hume’s account can avoid the major difficulty for classic compatibilists such as Ayer, of distinguishing those “constraints” that supposedly undermine moral freedom from the various risks, threats, and limitations that are just part of normal life. But Hume’s very thin conception of moral freedom as simple voluntariness faces well-known difficulties with regard to more complex, non-rational factors that may determine our choices, such as addiction, indoctrination, and coercive control. These might be addressed in a Humean spirit by appeal to higher-order desires (as in Frankfurt 1971), but a fully adequate theory would probably have to move beyond the relatively crude “Humean” model of desiredriven behavior, and take more account of the messy realities of human psychology, in the study of which Hume himself was such a pioneer.16,17

Notes 1 For detailed discussion of Hume’s theory of causation, see in particular Millican (2021) and (forthcoming). Both make reference to numerous earlier interpretive debates, which are mainly ignored here. 2 In contrast with the bogus notions of substance, accident, and inherence, for example, rejected at T 1.4.3.3‑8 and 1.4.5.2‑6. 3 Clarke appears to accept that human behavior is predictable in principle, since he appeals to this in explaining God’s foreknowledge of our actions – for discussion, see Millican (2010: 619). 4 Quite apart from any doubts about Hume’s Copy Principle and his determinism, one might reasonably be suspicious of his apparent implicit denial of a distinction between physical and moral causation, in so far as the former concerns events that are explicable only in terms of physical laws, whereas the latter concerns actions whose explanation involves reference to motives, intentions, plans, strategies, or other forward-looking considerations. “Moral causation” in this sense need

79

Peter Millican not, however, be confined to actions of conscious agents, as illustrated by AI systems (such as chess computers) and arguably by evolutionary phenomena. 5 For related remarks about the popular “causal closure of physics” doctrine, see Millican 2010: 636–637. 6 See also T  2.3.1.18, EHU  6.1 and  8.25, and for detailed discussion of Hume’s determinism, Millican (2010). 7 Note also that Hume attacks supposed theoretical objections to the possibility of mental/physical causal interaction by appeal to his theory of causation, in an important pro-materialist argument at T 1.4.5.30‑33. 8 Contemporary usage leaves this unclear, because some took “spontaneity” to mean simple voluntariness (e.g. Watts 1732: 5–6; Ramsay 1751: 107, 295), whereas others saw it as requiring more rational control (Bayle 1697, “Rorarius” note F, pp. 228–229; Voltaire 1764: 31). 9 This point is easy to overlook when the discussion is couched – as it often is – in terms of whether humans are “free agents” in an absolute sense, as opposed to being “free to do X”. 10 It is also striking that Hume says very little about the conditions for just punishment except within his short and posthumously published essay “Of the Immortality of the Soul”, as illuminatingly discussed by Russell (1995, ch. 10). 11 This is Hume’s argument that his doctrine of necessity (i.e. determinism), so far from undermining morality as “the common opinion” presumes, is actually essential to it, ensuring the crucial link between motives and actions. 12 For more details and discussion, see Millican (2012). 13 This mechanism of sympathy apparently disappears from the second Enquiry, but is commonly assumed to remain implicit. My own view, however (see Millican 2020, §§5‑7), is that it is crucial in the Treatise precisely because Hume is there developing his theory on a predominantly egoistic basis. Its absence from the Enquiry, accordingly, reflects his new awareness – evident in the second Appendix “Of Self-love” – that we are genuinely altruistic, and hence capable of caring about others without literally sharing in their feelings. 14 This, again, implies nothing about the causation of my volition or my character, so Hume is not forced into any regress about whether we have responsibility for either of these. Indeed, he insists that the virtues and vices for which we may be morally appraised include entirely involuntary natural abilities (see T 3.3.4 and EPM Appx 4). It is only in respect of voluntary actions that moral appraisal presupposes responsibility and hence hypothetical liberty. 15 With one very important exception noted at E 8.35: that these sentiments – and hence ascriptions of blame, praise, and moral responsibility – are not undermined by abstract reflections on such metaphysical issues as the deterministic nature of the universe. 16 For illuminating discussion of these sorts of complications, see Pitson (2016: 389–392). 17 I am very grateful to Don Garrett, Hsueh Qu, Paul Russell, and especially Max Kiener, for helpful comments on an earlier draft of this paper.

Further reading Botterill (2002) gives a lively and illuminating analysis of Section 8 of the first Enquiry, distinguishing Hume’s position sharply from that of “classic” compatibilists. Pitson (2016) offers a sensitive and wide-ranging account of Hume’s position, drawing useful links with more recent literature. Russell (1995) presents his influential “naturalistic” interpretation of Hume’s position, based in part on a distinctive understanding of Hume’s view of causation. For a much briefer summary, see Russell’s article “Hume on Free Will” in the Stanford Encyclopedia of Philosophy.

References Ayer, A. J. (1954). “Freedom and Necessity,” chapter  12 of Philosophical Essays (pp.  271–284). London: Macmillan. Bayle, Pierre. (1697). Historical and Critical Dictionary: Selections, trans. Richard H. Popkin. Indianapolis: Bobbs-Merrill, 1965.

80

Hume on free will and moral responsibility Botterill, George. (2002). “Hume on Liberty and Necessity,” in Peter Millican (ed.), Reading Hume on Human Understanding (pp. 277–300). Oxford: Clarendon Press. Bricke, John. (2008). “Hume on Liberty and Necessity,” in Elizabeth S. Radcliffe (ed.), The Blackwell Companion to Hume (pp. 201–216). Oxford: Blackwell. Clarke, Samuel. (1717). Remarks Upon a Book, Entituled, A  Philosophical Enquiry Concerning Human Liberty. London: James Knapton. Frankfurt, Harry G. (1971). “Freedom of the Will and the Concept of a Person,” The Journal of Philosophy, 68, 5–20. Hume, David. (1739–40, 2007). A Treatise of Human Nature: A Critical Edition, eds. David Fate Norton and Mary J. Norton. Oxford: Oxford University Press. ———. (1748, 2007). An Enquiry Concerning Human Understanding, ed. Peter Millican. Oxford: Oxford University Press. ———. (1751, 1998). An Enquiry Concerning the Principles of Morals, ed. Tom L. Beauchamp. Oxford: Oxford University Press. Locke, John. (1690/1975). Essay Concerning Human Understanding, ed. P. H. Nidditch. Oxford: Clarendon Press. Millican, Peter. (2010). “Hume’s Determinism,” Canadian Journal of Philosophy, 40, 611–142. ———. (2012). Chapter 5 “Hume,” in Tom Angier (ed.), Key Thinkers: Ethics (pp. 105–131). London: Bloomsbury; revised in second edition 2023, pp. 89–111. ———. (2020). “The Relation Between Hume’s Two Enquiries,” in Jacqueline Taylor (ed.), Reading Hume on the Principles of Morals (pp. 273–299). Oxford: Oxford University Press. ———. (2021). “Hume on Causation and Causal Powers,” in Benjamin Hill, Henrik Lagerlund, & Stathis Psillos (eds.), Reconsidering Causal Powers: Historical and Conceptual Perspectives. Oxford: Oxford University Press. ———. (forthcoming). “Hume as Regularity Theorist – After All!”, Hume Studies, 49, (2024). Penelhum, Terence. (2000). “Hume and the Freedom of the Will,” in T. Penelhum (ed.), Themes in Hume: The Self, the Will, Religion (pp. 156–176). Oxford: Clarendon Press. Pitson, Tony. (2016). “Hume, Free Will, and Moral Responsibility,” in Paul Russell (ed.), The Oxford Handbook of Hume (pp. 380–400). Oxford: Oxford University Press. Ramsay, Andrew Michael [Chevalier]. (1751). The Philosophical Principles of Natural and Revealed Religion (2nd ed.). London. Russell, Paul. (1995). Freedom and Moral Sentiment: Hume’s Way of Naturalizing Responsibility. Oxford: Oxford University Press. Voltaire [François-Marie Arouet]. (1764). The Newtonian Philosophy Compared with That of Leibnitz, Translated from the French of M. de Voltaire. Glasgow. Watts, Issac. (1732). An Essay on the Freedom of Will in God and in Creatures. London (originally published anonymously, but included in volume 6 of Watts’ Works, 1753).

81

6 SIDGWICK ON FREE WILL AND ETHICS Anthony Skelton

In the Outlines of the History of Ethics, Henry Sidgwick writes that the subject of ethics includes “some examination of the question of human Free Will” (Sidgwick 1902: 11). When he addresses the question in The Methods of Ethics he says that he is anxious to avoid the “great” difficulties associated with it and thus to develop a view on it that keeps “as much of my subject as possible as free from their perturbing influence” (Sidgwick 1907: 57). In keeping with this aim, he attempts to argue that appearances to the contrary, resolution of the free will controversy is of “limited” importance to ethics (1907: 66; also 1889: 474–479, 1902: 261). Despite the view’s uniqueness, surprisingly little sustained attention has been paid to it. This chapter tries to remedy this situation. Part one clarifies Sidgwick’s argument for the claim that resolving the free will controversy is of only limited importance to ethics. Part two examines and tries to deflect objections to Sidgwick’s position raised by J. B. Schneewind (1977) and Roger Crisp (2015) while itself putting forward a distinct criticism. Part three raises objections to Sidgwick’s claim that “it is practically impossible to be guided, either in remunerating services or in punishing mischievous acts, by any other considerations than those which the Determinist interpretation of desert [and responsibility] would include” (Sidgwick 1907: 72). The chapter concludes by noting that although Sidgwick’s view is not without problems, his discussion forces us to consider which aspects of moral thinking (if any) survive the recognition that free will is illusory.

Sidgwick on the significance of free will to ethics Sidgwick thinks that the main question at issue in the free will problem is the following: Is the self to which I refer my deliberate volitions a self of determinate moral qualities, a definite character partly inherited, partly formed by my past actions and feelings, and by any physical influences that it may have unconsciously received; so that my voluntary action, for good or for evil, is at any moment completely caused by the determinate qualities of this character, together with my circumstances, or the

DOI: 10.4324/9781003282242-10 82

Sidgwick on free will and ethics

external influences acting on me at the moment – including under this latter term my present bodily conditions? – or is there always a possibility of my choosing to act in the manner that I now judge to be reasonable and right, whatever my previous actions and experiences may have been? (Sidgwick 1907: 61–62) Sidgwick considers what he calls libertarian and determinist answers to this question. The libertarian opts for the second disjunct above, maintaining, Sidgwick reports, that “supposing that there is no obstacle to my doing it other than the condition of my desires and voluntary habits” (Sidgwick 1907: 65), it is always possible to choose to do what I judge to be right and reasonable or not even where the antecedents to my choice remain unchanged. The libertarian finds support for this in the “immediate affirmation of consciousness in the moment of deliberate action” (Sidgwick 1907: 65; also 1889: 475–476). Involved in the usual case of deliberation about what one ought to do in light of one’s principles is, the libertarian contends, the “actual consciousness of choosing between alternatives of conduct” (1889: 476). The idea is, Sidgwick argues, that “[i]n deliberate volitions there is always a conscious selection of the result as one of two or more practical alternatives” (Sidgwick 1907: 61; also 1889: 476). Central to deliberate action is the idea of believing ourselves to have a choice of what to do, whatever has come before. Sidgwick concedes that the consciousness of free choice may be “illusory: that if I knew my own nature I might see it to be predetermined that, being so constituted and in such circumstances, I should act on the occasion contrary to my rational judgment” (Sidgwick 1889: 477, 1907: 65–66). However, accepting this would involve a fundamental alteration of my “whole conception” of what I now call “my” action (Sidgwick 1907: 66; also 1889: 476–478). If one came to view one’s actions as completely determined, one would be forced to give up the usual way in which one attributes one’s actions to one’s self, namely, as the product of deliberation and choice. Sidgwick’s idea here is that since the conception of volition previously mentioned is the only one possible and it includes the idea that, contrary to determinism, we possess freedom of choice, we have some reason to think we in fact have such freedom (1889: 477). The determinist denies what the libertarian asserts based on “a cumulative argument of great force” (Sidgwick 1907: 62), and so affirms the first disjunct above in Sidgwick’s account of the main question at issue in the free will debate, namely, that all one’s actions are the product of antecedent events outside one’s control. Sidgwick lists a number of considerations favoring determinism, the most potent of which is the following.1 All “competent thinkers” hold, Sidgwick affirms, that with the exception of human volitions “events are determinately related to the state of things immediately preceding them” (1907: 62). “Every scientific procedure assumes it: each success of science confirms it” (1907: 63). Such considerations, the determinist submits, increase “the indisposition to allow the exceptional character claimed by Libertarians for the department of human action” (1907: 63).2 Sidgwick does not provide a theoretical solution to the free will problem. He does not, for example, argue that we do or do not have contra-causal free will. Instead, he argues for a practical solution to the problem (Sidgwick 1889: 474–475). He aims to show that for “the purposes of practical reasoning the two opposed arguments [libertarian and determinist]

83

Anthony Skelton

cannot really collide” (1889: 475). The idea is that while the gap between rival solutions to the problem of free will is in theory quite wide, its effect on reasoning in practice is quite limited.3 He argues as follows (Sidgwick 1889: 478–479, 1907: 68–70): P1. If taking up a stance on the question of free will by adopting libertarianism or determinism has more than limited significance to ethics, it alters either our view of what is ultimately good or our practical conclusions as to the fittest means to securing these goods. P2. It is not the case that taking up a stance on the question of free will by adopting libertarianism or determinism alters either our view of what is ultimately good or our practical conclusions as to the fittest means to securing these goods. C1. Therefore, it is not the case that taking a stance on the question of free will by adopting libertarianism or determinism has more than limited significance for ethics. Sidgwick argues that neither the value of happiness (pleasure) nor the value of perfection (the realization to a high degree of mental capacities essential to human nature, e.g., theoretical rationality and virtuous willing) is altered by the adoption of libertarianism as opposed to determinism.4 If Happiness . . . be taken as the ultimate end of action on a Libertarian view, the adoption of a Deterministic view affords no ground for rejecting it: and if Excellence [perfection] is in itself admirable and desirable, it surely remains equally so whether any individual’s approximation to it is entirely determined by inherited nature and external influences or not. (Sidgwick 1907: 68; also 1889: 478) Furthermore, neither the adoption of determinism nor libertarianism affects our scientific accounts of the relation between means and ends. If we confine our attention to such connexion between means and ends as is scientifically cognizable, it does not appear that an act now deliberated on can be less or more a means to any ulterior end, because it is predetermined. (Sidgwick 1907: 69; also 1889: 478–479)5 Sidgwick does, however, think that some (theoretically significant) changes in our commonsense attitudes is implied by the endorsement of determinism. He says, for example, that “it must be admitted . . . that the common retributive view of punishment, and the ordinary notions of ‘merit,’ ‘demerit,’ and ‘responsibility’ . . . involve the assumption of [libertarian, contra causal] Free Will” (Sidgwick 1907: 71; also 284, 349). Therefore, if we jettison libertarianism, we are forced to divest ourselves of retributivism and our common conceptions of merit, desert, and responsibility. The assumption behind these views is that if a person’s actions are the product exclusively of “events anterior to his personal existence” and so he could not have done otherwise, he cannot really have merit or demerit, and it would be “repugnant to the common moral sense . . . to reward or punish . . . praise or blame . . . him” (1907: 66).

84

Sidgwick on free will and ethics

The determinist gives to these notions a different signification. Of the notion of responsibility, Sidgwick says if I affirm that A is responsible for a harmful act, I mean that it is right to punish . . . [A] for it; primarily, in order that the fear of punishment may prevent him and others from committing similar acts in the future. (Sidgwick 1907: 72–73) Of the notion of desert, he says when a man is said to deserve reward for any services to society, the meaning is that it is expedient to reward him, in order that he and others may be induced to render similar services by the expectation of similar rewards. (Sidgwick 1907: 284n1) Sidgwick describes these as utilitarian conceptions of responsibility and desert. Adopting determinism seems, then, to have more than limited significance for ethics. It involves rejecting retributivism about punishment and commonsense views of responsibility, desert, merit, and so on. Adopting determinism would therefore be revisionary. On the libertarian, commonsense views of responsibility and desert, the assignment of responsibility and the allotment of rewards is based solely on backward-looking considerations, on whether what happened was in some way under the control of the agent and so what an agent did. On the retributivist view, punishment of an individual is justified entirely on the basis that they deserve it for having done wrong in the past for which they are responsible. On the determinist, revisionary view, the assignment of responsibility and the allotment of rewards is based solely on forward-looking considerations, on what will happen. The determinist, on Sidgwick’s reckoning, justifies punishment entirely on the basis of future-looking considerations, including reform of offenders and general deterrence. Sidgwick concedes that the difference between the libertarian and determinist is, in the case of the justificatory basis of punishment, merit, desert, and responsibility, “theoretically very wide” (Sidgwick 1907: 72). But, he says, this theoretical difference has hardly any effect in practice. This is because “it is practically impossible to be guided, either in remunerating services or in punishing mischievous acts, by any other considerations than those which the Determinist interpretation of desert [and responsibility] would include” (1907: 72; also 285). Certain “practical exigences” force us to rely exclusively on forward-looking considerations in assigning responsibility, allotting rewards, and punishing.

Is free will of limited significance to ethics? J. B. Schneewind and Roger Crisp object to Sidgwick’s claim that what we take to be ultimately valuable is unaffected by one’s stance on the free will question. That is, they challenge P2 of Sidgwick’s argument that was outlined above. Sidgwick says that the manifestation of such virtues as courage, temperance, and justice do not “become less admirable because we can trace their antecedents in a happy balance of inherited dispositions developed by a careful education” (Sidgwick 1907: 68). Schneewind

85

Anthony Skelton

(1977: 212) objects that if what Sidgwick has in mind “points to strict determinism”, then while such virtues may not be less admirable, they would become “admirable in a different way – as we admire straight noses and golden hair, not as we admire honesty under severe temptation” (1977: 212). In reply, Sidgwick might say it is not clear that such virtues and other perfections are admirable in a different way under the conditions he describes. Honesty, temperance, and justice, among other virtues, could still be regarded as good to have for their own sake and so worthy of promotion, and celebration even if they are the product of inheritance and education. We might admire them in the same way we admire a body of knowledge a person acquires under great stress due entirely to the inherited disposition to disciplined study augmented by a solid education. Or we might admire them in the same way that we admire a small child’s artistic creativity or intelligence which result from natural endowment and instruction. Our admiration in these cases seems not to resemble our admiration of straight noses, which does not involve valuing them or celebrating them for their own sake. In response, Schneewind might concede that such virtues remain admirable as Sidgwick says but argue that what changes is our attitude to the agent manifesting the virtues. We may admire or come to regard the agent differently. We might think her less worthy of praise or of congratulation. This reply might be right. But, as we saw, Sidgwick concedes that in theory this is one implication of the endorsement of determinism, though he thinks this concession is of limited practical significance. We shall return to this shortly. Crisp concedes that Sidgwick might be right that when viewed “merely” as perfections the value of certain intellectual and physical perfections and excellences of character remains whether one is a determinist or a libertarian (Crisp 2015: 52). But, he argues, if such things are viewed as personal accomplishments, “the libertarian may see determinism as undermining their value” (2015: 52). It not clear that the value of accomplishments would be undermined by the adoption of determinism. The value of an accomplishment or achievement per se does not obviously depend on free will any more than the value of certain intellectual or physical perfections depends on free will. According to an influential account by Gwen Bradford (2015), an achievement involves a product (a cure for cancer) and a process (scientific research), where the product is competently caused by means of a difficult (intensely effortful) activity. Such things qualify as achievements, in other words, when they involve the exertion of (intense) effort and (roughly speaking) a certain amount of justified true beliefs about the (causal) relationship between one’s activities and the product of one’s actions. It is unclear that we would stop admiring, promoting, encouraging, and honoring achievements in this sense given the fact of determinism. A  determinist’s judgement about the value of achievement would seem to persist even were she to agree with Bradford that the value of an achievement rests on exercises of theoretical and practical reason and the will and is therefore perfectionist in nature (Bradford 2015). It does not seem obvious that the value lies in a “free” will and so, again, it is not clear that the adoption of determinism would undermine their value. Consider the fact that many admire the accomplishments or achievements of young children despite the fact that we do not think small children have any significant kind of free will. It is not clear, then, that our commonsense attitudes about the value of certain achievements change with knowledge that they are due entirely to education and to inherited dispositions.6 Of course, as in the case of Schneewind, we might admit that the scientists who discover cures or the individuals who build businesses become admirable in different ways and 86

Sidgwick on free will and ethics

that we might modify our view of them once we consider and adopt a deterministic point of view on their activities. But, as we saw, Sidgwick thinks that if we adopt determinism our views of responsibility and merit must be altered in theory, though this alteration does not have a significant effect in practice. It is to this point that we will turn in the next section. Crisp argues further that it is not only personal accomplishments that may be undermined by determinism. He asks us to imagine that . . . [a] libertarian had taken much pleasure in her accomplishments and her moral integrity, attributing both to free and unconstrained choices. Especially in the case of moral integrity, it is not unlikely that the belief in determinism may affect these more sophisticated, intentional pleasures, and indeed cause her a certain amount of unpleasant regret. (Crisp 2015: 51–52) Sidgwick might argue that in this case since it is not the value of, but rather the means to, happiness that is altered, the alteration is, practically speaking, of only limited importance. He might go on to suggest that libertarians could well still acquire the pleasure they relish by retaining the belief in free will as part of the best strategy for promoting happiness or pleasure. For some, this suggestion for the libertarian may constitute a practical change that is of more than limited significance. This, of course, speaks to the issue of when, on Sidgwick’s view, an alteration is of limited, as opposed to more than limited, significance. He is not clear on where lies the boundary. In any case, even if the reply is effective, it is less clear that other changes occasioned by the adoption of determinism that Sidgwick allows can be plausibly classified as limited in ethical importance. He readily admits that there are a number of cases in which adopting determinism seems to make a significant difference to our motives rather than the substance of our obligations. Sidgwick says that the acceptance of determinism makes no difference to our belief about the connection between some goal and the most efficient means to it in so far as the belief rests on “empirical or other scientific grounds” (Sidgwick 1889: 479, 1907: 69).7 However, things look different, he suggests, when “theological considerations” are introduced (1889: 479, 1907: 69). Suppose one is a theist who believes that one’s happiness in the next life is dependent on freely choosing to perform one’s duty in this world (Sidgwick 1889: 479, 1907: 69). Sidgwick concedes that if one holds this view, whether or not one has libertarian free will “becomes obviously of fundamental ethical importance” (Sidgwick 1907: 69, 1889: 479). The denial of free will in this case has the effect of removing one’s motive to discharge one’s duty. The idea is that if one’s motive to do what one ought rests on a belief that God exists and rewards virtue with happiness in the afterlife and does so only if one freely chooses virtue, then the affirmation of determinism will undermine one’s incentive to perform one’s duty. The ethical significance is therefore fundamental for a theist of this sort. Sidgwick seems to think this is not detrimental to his argument because the denial of free will has the effect of undermining one’s motive to duty, he says, only in so far as (a) one’s self-interest diverges from what morality demands apart from “theological considerations” and (b) free will (in 87

Anthony Skelton

the libertarian sense) is an “indispensable” assumption of the “theological reasoning that removes this divergence” (Sidgwick 1907: 69). Sidgwick ends the discussion here in part on grounds that (b) is not “within the scope of this treatise to discuss” (1907: 69). This is rather too convenient. Sidgwick ought to have been a bit more concerned. He believes that duty and self-interest do diverge apart from theological considerations (1907: 162–175, 496–509), he is a theist (Sidgwick 1906: 228), and even suggests an argument for God’s existence in part on the grounds that God might make duty and self-interest coincident (1906: 600–608).8 He does admit that one’s stance on the free will controversy may have a significant impact on ethics if you embrace the previous theological position. In response, he modifies his conclusion to say resolution of the free will debate is of limited significance to ethics unless you make the assumptions associated with the previous theological position (Sidgwick 1907: 71). But this qualification might, for some, look more significant than Sidgwick allows. Sidgwick seems to accept that the effect of the adoption of determinism would be of more than limited significance in the previous case because the effect would be the removal of the motive to discharge one’s duty. He seems less worried in cases where the denial of libertarian freedom occasions only modifications to one’s motives to perform one’s duty. He admits, for example, that motives to perform one’s duty may be weakened if one accepts determinism, since a man will not feel remorse for his actions, if he regards them as necessary results of causes anterior to his personal existence. I admit that so far as the sentiment of remorse implies self-blame irremovably fixed on the self blamed, it must tend to vanish from the mind of a convinced Determinist. (Sidgwick 1907: 71) In reply, Sidgwick says that the determinist will have to jettison the feeling of remorse. But he suggests that this will not much effect the convinced determinist’s motive to do their duty, since she may possess other and equally effective motives prompting her to duty. There is, he says, no reason that the imagination of the determinist should not be as “vivid, his sympathy as keen, his love of goodness as strong as a Libertarian’s”, and therefore no reason why the determinist’s “dislike for his own shortcomings and . . . mischievous qualities of . . . character which have caused bad actions in the past should not be as effective a spring of moral improvement as the sentiment of remorse would be” (Sidgwick 1907: 71). This is not an entirely effective reply. At most, Sidgwick is permitted to claim that the adoption of determinism may have only limited significance to ethics in so far as motives are concerned. It is, after all, quite possible that amongst those qualifying as unimaginative determinists, the modification of motives is quite significant in practice. They may well jettison remorse while lacking a suitable replacement for it, in which case the effect of adopting determinism will in fact weaken their motive to do their duty. Moreover, Sidgwick overlooks the difficulty that even an imaginative and sympathetic determinist may have in coming to regard moral defects as they regard intellectual or organic defects, which, Sidgwick maintains, we take great strides to “cure” despite the fact that they cause us no remorse (Sidgwick 1907: 71). It is quite possible that for some, classifying these changes to motives as of limited importance to ethics may well be difficult to accept, if not question begging.9

88

Sidgwick on free will and ethics

Are the theoretical differences between libertarian and deterministic conceptions of punishment and responsibility insignificant in practice? Let’s return to Sidgwick’s claim that “it is practically impossible to be guided, either in remunerating services or in punishing mischievous acts, by any other considerations than those which the Determinist interpretation of desert [and responsibility] would include” (Sidgwick 1907: 72). He assumes, it seems, that the Determinist interpretation of these notions must be utilitarian in nature (1907: 71, 284n2). Sidgwick provides two arguments for his practical impossibility claim. The first argument appeals to what he calls the “practical exigencies of social order and wellbeing”. These force us to assign responsibility for bad acts and to mete out legal punishment in “deterrent and reformatory rather than retributive” ways “quite apart from any Deterministic philosophy” (Sidgwick 1907: 72). He suggests that due to such practical exigencies we “punish negligence, when its effects were very grave, even when we cannot trace it to wilful disregard of duty . . . [and] rebellion and assassination . . . although we know that they were prompted by sincere desire to serve God or to benefit mankind” (1907: 72n2; also 1907: 446–447, 1908: 113–114). There are two replies to this. First, it is not clear that we do not hold people responsible in the retributive sense for harmful negligence and rebellion of the sort Sidgwick has in mind. Provided there is some sense in which they had control over what they did in being negligent or in participating in a rebellion there is room for holding people responsible and punishing them for the harm caused in ways acceptable to libertarians, that is, based on backward-looking considerations and desert. Sidgwick seems to think that the punishment of harmful negligence and rebellion, which does not rest on wilful disregard of duty or wickedness cannot rest on anything other than utilitarian considerations (that is, “prevention of mischief” (Sidgwick 1908: 114; also 110; 1907: 446–447). He seems to suppose that retributivists have grounds to punish such acts only if they rest on intentional neglects (1907: 447) or “wickedness” (1908: 114). In the absence of such neglect or wickedness they would, then, have no basis for punishing. Therefore, if punishment seems fitting, it would, Sidgwick thinks, have to rest on utilitarian considerations, that is, on preventative or reformatory grounds. But the retributivist need not concede that punishment is permissible only in cases where there is intentional neglect or wickedness. They may simply argue that the negligent, the rebel, the assassin who commits mischief may be punished provided they were able to do otherwise or had some level of control at some point over the harmful outcomes produced by their action. Second, even if Sidgwick is right that we reason as deterrence theorists where rebellion and assassination is not motivated by “immoral intention” (Sidgwick 1908: 114), it does not follow that we ought to employ this kind of reasoning in all other cases, e.g., where an individual knowingly commits a wrong such as an assault or an act of vandalism. It is not true that because we reason as deterrence theorists in some cases (rebellion in the sincere belief you are in the right) that we must do so in other cases. At most, what Sidgwick provides is a presumption in favor of his view. The other argument on which Sidgwick relies on his practical impossibility claim occurs in his discussion of the commonsense notion of justice. Sidgwick notes again that our commonsense notions of desert and merit “require material modification” if we deny the reality of free will in the libertarian sense (Sidgwick 1907: 284; also 349). But, he reiterates, the

89

Anthony Skelton

differences between libertarians and determinists with respect to these notions “can hardly have any practical effect” (1907: 285). His support in this context for this claim is that “it does not seem possible to separate in practice that part of a man’s achievements which is due strictly to his free choice from that part which is due to the original gift of nature and to favouring circumstances” (Sidgwick 1907: 285). This dismissal of libertarian notions of merit and desert seems too hasty. It might be true that it is hard to separate what is due to one’s free choice from what is due to other, distinct factors in deciding what one deserves. But it does not obviously follow that it is not possible in some cases to form at least reasonable judgements about what is due to each of these distinct factors. We might get it wrong, but surely that is not an argument against trying to form plausible views about the contributions of each factor. Sidgwick himself seems to suggest that such separation might be possible to some extent in some cases if only we were to remove  .  .  . the inequalities that are attributable to circumstances, by bringing the best education within the reach of all classes, so that all children might have an equal opportunity of being selected and trained for any functions for which they seemed to be fit. (Sidgwick 1907: 285) One might suggest here that perhaps the right approach is not to jettison libertarian notions of desert and reward, and so on, but to instead reform society so that the provision of rewards is more clearly based on what one deserves in the backward-looking sense. Sidgwick says recall that the only tenable Determinist interpretation of Desert is, in my opinion, the Utilitarian: according to which, when a man is said to deserve reward for any services to society, the meaning is that it is expedient to reward him, in order that he and others may be induced to render similar services by the expectation of similar rewards. (Sidgwick 1907: 284n1) It is not clear that it is possible to achieve clarity with respect to what will and will not function to induce people to supply services to other people or the nature of the conditions under which it expedient to reward people. On this utilitarian view of desert and responsibility, claims about what will induce people to render services and what to reward people for will depend on claims about the consequences of various incentives. However, as Sidgwick himself shows, it is hard to calculate the consequences of actions because there is difficulty associated with arriving at precise accounts of the quantitative relations between pleasures (pains) or units of happiness (unhappiness) (Sidgwick 1907: 123–130, 413). It is not obvious, then, that the deterministic forward-looking view of desert and responsibility is better in practice than the backward-looking libertarian view. In reply, Sidgwick might point out that we do not need precision to know in general what will induce people to render services. He might insist that this gives his view some advantage over the view that has the seemingly more difficult task of separating what is due to free choice and from what is due to circumstances in deciding what individuals deserve. It may not be clear how much of an advantage his view has, but perhaps he can claim that the problem he faces is more tractable. 90

Sidgwick on free will and ethics

Finally, Sidgwick seems to be of the view that either we rely on libertarian conceptions of punishment, responsibility, desert, merit, and so on, or we rely on determinist conceptions of the same concepts. He maintains in addition that utilitarian conceptions of such concepts are the only ones the determinist may tenably accept (Sidgwick 1907: 71, 284n2). And, further, as we have seen, he argues that utilitarian conceptions of responsibility, desert, punishment, and so on are the only ones suitable for use in the practice of holding people responsible and dealing with criminal conduct, among other things. However, it is not obvious that determinists must accept utilitarian accounts of responsibility or the utilitarian basis for punishing criminal behavior (that is, deterrence and reform). Determinists seem not to be forced to accept Sidgwick’s utilitarian conception of responsibility. It is possible for a determinist to reject utilitarian conceptions of responsibility and cognates and to banish talk of such concepts altogether. A determinist might, for example, think of moral agents not as responsible, praiseworthy or blameworthy for their actions, but instead as morally attractive or unattractive based on considerations other than the exercise of free will in the libertarian sense and on other than the utilitarian or extrinsic considerations Sidgwick relies on (Smilansky 1994: 361).10 It might be too drastic to eliminate responsibility talk. But even if it is right that we need to rely on responsibility talk, determinists can reject the account of responsibility Sidgwick attributes to them. W. D. Ross (1939) is a determinist who rejects the utilitarian account of responsibility. He says that: holding fast to Determinism, I  am inclined to think that the only account we can give of responsibility is this: that bad acts can never be forced on any one in spite of his character; that action is the joint product of character and circumstances and is always therefore to some extent evidence of character; that praise and blame are not (though they serve this purpose also) mere utilitarian devices for the promotion of virtue and the restraint of vice, but are the appropriate reactions to action which is good or bad in its nature just as much if it is the necessary consequence of its antecedents as it would be if the libertarian account were true. (Ross 1939: 250) If Ross is right, one can be a determinist and reject the Sidgwick’s view of responsibility. The rejection of libertarian free will need not involve a commitment to thinking that the only basis for the justification of punishment is deterrence and reform. Indeed, in recent literature on free will some have rejected both the existence of free will in the libertarian sense and the deterrence theory of punishment. It is, some have argued, possible to adopt a public health approach to dealing with criminal behavior, thinking of measures for dealing with crime not as functioning primarily to deter or prevent future mischief, but instead as functioning to protect the public from dangerous criminals in the same way that quarantining people with virulent infectious diseases works to protect the public, namely, through some form of containment (Pereboom 2001: 174–177; Caruso 2022). On this non-punitive view, serious criminals and individuals with infectious diseases are treated in an analogous way. Neither is held responsible in the libertarian sense for their conduct or condition. Nevertheless, just as we may quarantine and contain people with deadly infectious diseases to protect public health, we may quarantine those who commit serious crimes to protect the public from harm. As with those quarantined because they have a serious infectious disease, 91

Anthony Skelton

those quarantined for criminal behavior are contained only for as long as they remain a threat and in conditions no harsher than is necessary. The important point here is not whether this view is plausible.11 It is that this view is an option for dealing with criminal behavior that is open to determinists to adopt. It rivals the deterrence theory Sidgwick endorses. It seems, then, that determinists have more options for thinking about punishment and responsibility than Sidgwick allows. So it does not follow from the fact that we cannot rely in practice on libertarian conceptions of responsibility, desert, and punishment that we must, then, rely on utilitarian conceptions of responsibility, desert, and punishment in practice. Sidgwick might try to argue that practical exigencies may force us to accept his favored, utilitarian conceptions of responsibility and desert and the deterrence theory of the justification of punishment. However, to do so he will have to rely on arguments distinct from the ones he actually puts forward against the libertarian conceptions of responsibility and punishment. And without such arguments he cannot claim that the acceptance of determinism is of limited significance to ethics. Sidgwick therefore must do more to defend his claim (a) that determinists are committed to utilitarian conceptions of punishment, merit, desert, and responsibility and (b) that these conceptions are the only ones viable in practice. Without a defence of the claim that in practice we are forced to be deterrence theorists about punishment, for example, he cannot rule out that the adoption of determinism will have significant impacts on ethical practice. If we adopt determinism and have a good argument in favor of the quarantine model for dealing with serious criminals, the adoption of determinism will have potentially significant impacts on what we ought to do in practice and so will be of more than limited significance to ethics.

Conclusion Sidgwick argues that resolution of the free will controversy is of limited significance to ethics. He says that while in theory libertarians and determinists conflict over how best to understand the notions of merit, demerit, and responsibility and over the correct account of the justification of punishment, these theoretical conflicts can hardly have any practical effect. After all, in practice, we are forced for various reasons to rely on deterministic (utilitarian) conceptions of these notions. As the foregoing discussion has clarified, we should be chary of accepting Sidgwick’s view. But whatever we may think of the merits of it, Sidgwick forces us to consider the options we have respecting ethical reasoning and responsibility in the face of the denial of the existence of libertarian free will. Indeed, Sidgwick’s position is best understood as an intellectual forerunner to views attempting to establish which aspects of moral argument (if any) survive the discovery that we lack libertarian free will.12

Notes 1 I agree with Crisp (2015: 48) that this is Sidgwick’s most powerful argument in favor of determinism. 2 As Crisp notes, Sidgwick did not consider indeterminism. On some interpretations of quantum theory, which was developed only after Sidgwick died, at the most basic level matter behaves indeterministically. The implications of indeterminism for the free will debate are at present unclear. However, although the fact of indeterminism may weaken Sidgwick’s argument here, Crisp points out, “many continue to believe in determination at higher levels, and one might anyway make the

92

Sidgwick on free will and ethics position disjunctive: actions are either determined or random. And then Sidgwick’s argument goes through” (2015: 46). 3 Sidgwick says that libertarians tend to exaggerate the importance of free will to jurisprudence and ethics (Sidgwick 1907: 66). So, his main focus is on the impact on practical reason and ethics of adopting determinism. 4 Sidgwick thinks the two most commonly accepted ultimate values are happiness and perfection (1907: 9). 5 Sidgwick grants that one’s position on the free will debate will be of “fundamental ethical importance” if one is a theist who thinks the fittest means to happiness in the next life is dependent on freely choosing to perform one’s duty in this life (Sidgwick 1889: 479, 1907: 69). More on this shortly. 6 It is not clear that the adoption of a rival view of achievement yields a different result. Imagine again that achievements comprise difficulty and competent causation. But imagine further that some activity is considered difficult not when it requires the exertion of intense effort, but instead when it is an activity most adult human beings with average capabilities are likely to fail at. This conception makes possible achievements where due to skill and talent one succeeds in effortlessly doing something (e.g., writing a best-selling novel) at which most adult human beings with average capabilities are likely to fail. It’s not clear that the adoption of determinism would undermine the value of these achievements. For this conception of achievement, see von Kriegstein (2019). 7 For criticism of this point, see Crisp (2015: 53–54). 8 He does not seem to argue that free will is indispensable to this reasoning. 9 In discussing the “ought implies can” principle, Sidgwick notes one other possible alteration to our motives occasioned by the adoption of determinism (Sidgwick 1907: 67). He says the determinist holds the “commonly accepted” version of the principle according to which “ought implies can do it” if one chooses. “Ought to do” does not imply that one can choose to do. But, the libertarian asks, is it possible on the determinist view to choose what one ought? It may be true in some cases, Sidgwick says, that the determinist will agree that given one’s past experience one will “certainly” not choose what one ought (1907: 67). He allows that “[t]his being supposed it seems to be undeniable that this judgement will exclude or weaken the operation of the moral motive in the case of the act contemplated: I either shall not judge it reasonable to choose to do what I should otherwise so judge, or if I do pass the judgement, I shall also judge the conception of duty applied in it to be illusory, no less than the conception of Freedom” (1907: 67). Sidgwick says that he agrees that in such cases determinism has a “demoralising effect”. In response, he says that the situations in which one is certain one will not do what ought are rare: “Ordinarily the legitimate inference from a man’s past experience, and from his general knowledge of human nature, would not go beyond a very strong probability that he would choose to do wrong: and a mere probability – however strong – that I shall not will to do right cannot be regarded by me in deliberation as a reason for not willing” (1907: 68–69). 10 For a similar proposal, see Slote 1990: 375–379. 11 For criticism, see Smilansky 2017. 12 I wish to thank Lisa Forsberg and Max Kiener for helpful comments on previous versions of this chapter.

Further reading Baldwin, T., & Preti, C. (eds.). (2011). G. E. Moore: Early Philosophical Writings. Cambridge: Cambridge University Press contains Moore’s 1897 Trinity College Prize Fellowship dissertation entitled The Metaphysical Basis of Ethics, in which Moore discusses Sidgwick’s view on the free will question in relation to Kant. Hayward, F. H. (1901). The Ethical Philosophy of Sidgwick: Nine Essays, Critical and Expository. London: Swan Sonnenschein, chapter 4, includes a sympathetic exposition of Sidgwick’s position on free will and its relation to ethics, noting some changes in Sidgwick’s view across editions of The Methods of Ethics. Phillips, D. (2022). Sidgwick’s the Methods of Ethics: A Guide. Oxford: Oxford University Press is an excellent introduction to the main arguments in Sidgwick’s The Methods of Ethics, including (in chapter 3) Sidgwick’s argument that resolution of the free will debate is of limited importance to ethics. Sidgwick, H. (1888). “The Kantian Conception of Free Will,” Mind 13, 405–412, argues that to his detriment Kant relies on two conceptions of freedom.

93

Anthony Skelton

References Bradford, Gwen. (2015). Achievement. Oxford: Oxford University Press. Caruso, Gregg D. (2022). “The Public Health-Quarantine Model,” in Dana Nelkin  & Derk Pereboom (eds.), Oxford Handbook of Moral Responsibility. New York: Oxford University Press. Crisp, Roger. (2015). The Cosmos of Duty: Henry Sidgwick’s Methods of Ethics. Oxford: Oxford University Press. Pereboom, Derk. (2001). Living Without Free Will. Cambridge: Cambridge University Press. Ross, W. D. (1939). The Foundations of Ethics. Oxford: Oxford University Press. Schneewind, J. B. (1977). Sidgwick’s Ethics and Victorian Moral Philosophy, Oxford: Oxford University Press. Sidgwick, Henry. (1889). “Some Fundamental Ethical Controversies,” Mind, 14, 473–487. ———. (1902). Outlines of the History of Ethics for English Readers (5th ed.). London: Macmillan. ———. (1906). Henry Sidgwick: A  Memoir, eds. Arthur Sidgwick  & Eleanor Mildred Sidgwick. London: Macmillan. ———. (1981 [1907]). The Methods of Ethics (7th ed.). Indianapolis: Hackett. ———. (1908). The Elements of Politics (3rd ed.). London: Macmillan. Slote, Michael. (1990). “Ethics Without Free Will,” Social Theory and Practice, 16, 369–383. Smilansky, Saul. (1994). “The Ethical Advantages of Hard Determinism,” Philosophy and Phenomenological Research, 54, 355–363. ———. (2017). “Pereboom on Punishment: Funishment, Innocence, Motivation, and Other Difficulties,” Criminal Law and Philosophy, 11, 591–603. von Kriegstein, Hasko. (2019). “On Being Difficult: Towards an Account of the Nature of Difficulty,” Philosophical Studies, 176, 45–64.

94

Section 3: The scope of responsibility

7 ARISTOTLE ON LEGAL AND MORAL RESPONSIBILITY Interpretation and reform Terence Irwin

1  Legal responsibility Aristotle discusses questions about responsibility in his three major ethical works (cited hereinafter as as MM (= Magna Moralia), EE (= Ethica Eudemia), and EN (= Ethica Nicomachea)).1 He considers some legal aspects of responsibility in the Rhetoric. He does not suppose that these questions are purely philosophical. He recognizes that they arise in interactions between people who praise or blame each other for what they have done or left undone, and that they arise in the more formal context of a legal system, in trials, sentences, and punishments. He considers responsibility in his discussion of voluntary action, in the course of his discussion of virtuous and vicious characters and actions. Virtuous actions are proper object of praise, and vicious actions of blame; the proper objects of praise and blame are voluntary actions; hence we need to know what makes an action voluntary. A discussion of this question is also useful to legislators (EN 1109b30–5). Aristotle does not mean that voluntary actions are the only ones that we can praise and blame. We might decide – for all that he says – to praise and blame people for involuntary actions that are not in their control.2 But if we carried out that decision we would be acting unjustly. It is unjust to punish people for something that they cannot avoid doing, and equally it is unjust to blame them.3 This is why a discussion of the voluntary belongs not only to the section of the Ethics that discusses virtue in general, but also to the section that discusses justice. (EN v 8) An account of the voluntary might be expected to benefit not only legislators, but also citizens in their role in the legal system. Many trials in Athens had juries of 500 or more; the relatively small Council of the Areopagus, which tried some homicide cases, had perhaps 200 members. Other juries were much larger.4 They were not guided by expert judges who interpreted the law for them, or told them what sorts of evidence or arguments were inadmissible. If a question arose about whether the accused was responsible for some action, the question might be answered in different ways by the advocates on each side, but the members of the jury would have to decide for themselves who was right. The Athenian juror swore to judge in accordance with the laws, but where the laws did not settle a question to use his best judgment (gnȏmȇ, Aristotle, Rhetoric 1375a29–30). 97

DOI: 10.4324/9781003282242-12

Terence Irwin

When we decide that someone who has done the wrong action is not to be blamed for it, or at least not to be blamed to the usual degree, we exercise sungnȏmȇ (EN 1109b32). This term suggests some kind of sympathetic judgment;5 by taking the other person’s point of view as well as our own, we sometimes reach a sounder judgment about the degree of blame that is appropriate. Hence sungnômê may refer to excuse, pardon, or forgiveness, in the appropriate contexts, but the term has a wider scope than any of these. Aristotle recognizes an intellectual virtue that reaches the right judgment in such questions. What is called judgment (gnȏmȇ), which we say makes people have good judgment (eugnȏmonas), is the correct discrimination of what is decent. Here is a sign: We say that the decent person most of all has sympathetic judgment (sungnȏmonikon), and that having sympathetic judgment (sungnȏmȇn) about certain things is decent; now sympathetic judgment is correct judgment that discriminates what is decent , and the correct judgment is the one that discriminates what is true . (EN 1143a19–24) The description of this virtue shows that we should emphasize the element of judgment (gnȏmȇ) that is implied in the exercise of sympathetic judgment (sungnȏmȇ) in the remission of blame. This is the virtue that jurors need when they use their judgment in cases where the letter of the law does not give the correct answer. The virtue of good judgment is said to reach a true view of what would be decent (epieikes) in a particular case. In speaking of decency Aristotle refers back to his earlier discussion of decency (epieikeia, sometimes rendered ‘equity’; EN v 10). In that discussion he argues that decency is not opposed to justice, but it is needed to correct the injustice that would result from the rigid observance of a law that is stated in general terms. This correction agrees with the intention of the legislator (assumed to be just), who recognizes that injustice will result unless we exercise the correct judgment in some particular cases. From these remarks on good judgment, sympathetic judgment, and decency, we reasonably infer that questions about responsibility, praise, and blame need to be answered by the discovery of the decent course of action for the judge (or juror). Though the defendant’s action may appear to deserve punishment (he killed someone), further consideration shows that he had good reason to think the victim was an enemy soldier (because the victim was wearing the uniform of the enemy army), and he should not be punished. We exercise decency in cases in which one ought to exercise sympathetic judgment (sungnȏmȇ), and in distinguishing errors from acts of injustice, and from misfortunes (Aristotle, Rhetoric 1374b2–6). Wounding (trauma) involves the use of a weapon in assaulting a victim; but if every metal object counts as a weapon, anyone who gets into a fight while wearing a ring will be liable to a charge of wounding. The right judgment in this case is that merely wearing a ring (e.g., not using it to scratch) is not carrying a weapon (1374a33-b1), and therefore someone who uses his fists while wearing a ring is not guilty of using a weapon to wound the victim. When Aristotle says that his discussion will be useful to legislators, he might mean that we ought to understand the criteria that legislators actually use for drawing the relevant distinctions, or that we ought to set out the criteria that they ought to use, or he might mean both at once. If actual legislators have been partly right and partly wrong, the right judgments about responsibility will emerge partly from understanding the distinctions that are drawn, and partly from seeing where distinctions ought to be drawn. Aristotle and his 98

Aristotle on legal and moral responsibility

school studied the legal systems of different states,6 and in the course of their study they might expect to find the different criteria of responsibility that are maintained or presupposed by different legal systems. If we try to compare Aristotle’s discussion with the assumptions about responsibility that underlie contemporary Greek law, it is difficult to say how far he describes Greek assumptions, and how far he criticizes them. Athenian law is the only contemporary Greek legal system about which we know any details; unlike Aristotle, we cannot compare it in detail with other contemporary Greek laws, and hence we do not always know what specific laws he has in mind. But a comparison of Aristotle’s views with contemporary Athenian views will at least help us to see where Aristotle may be criticizing contemporary views, and not simply analysing them. The task of understanding Aristotle’s views is further complicated by the fact that we have to compare the three treatments of the questions in his three major ethical works. These different discussions are useful for us because, as we will see, they respond to contemporary legal assumptions in different ways. These different responses suggest different ways of explaining and of revising the views that we can discover from the study of Athenian law.

2  Voluntary and involuntary A preliminary survey of the distinctions that Aristotle takes to be relevant to responsibility will help us to identify some puzzles about the relation of his views to contemporary legal practice. In his view, praiseworthy actions are voluntary (hekousia). The appropriate attitude to involuntary (akousia) actions is sympathetic judgment (sungnȏmȇ) (EN 1109b30–2).7 Some, but not all, voluntary actions are actions from election (prohairesis, EN 1111b6– 10, EE 1226b30–1227a2). Some voluntary actions are ‘from forethought’ (pronoia, EE 1226b38); these appear, sometimes at least, to be identified with actions from election. These actions from election and forethought are contrasted with actions from non-rational desires (appetite (epithumia) and spirit (or anger; thumos)). Involuntary actions include those we do because of force (bia) and ignorance (agnoia) (EN 1109b35). The choice of ‘voluntary’ and ‘involuntary’ to render hekousion and akousion may be questioned.8 These renderings follow the Latin tradition, which uses ‘voluntarium’, and thereby suggests that they involve some relation to the will (voluntas). This suggestion may be questioned on the ground that Aristotle’s account of action and motivation has no place for the will, as it is conceived in the Latin philosophical tradition. The force of this objection depends on how the will is conceived in this tradition. But even if we stick to a nontechnical use of ‘voluntary’ (as in ‘voluntary contribution’ or ‘volunteer’), we may think it is unsuitable for Aristotle. In the non-technical English use, what we do voluntarily is sometimes opposed to what we are required to do, or what it compulsory for us to. But Aristotle does not believe that if we serve in the army because military service is compulsory, we serve akousiȏs. Similarly, actions performed under duress are often regarded as involuntary, but Aristotle, sometimes at least, takes them, or some of them, to be hekousia (EN 1110a4-b5). If, however, we render hekousion in the light of its use in earlier Greek, ‘voluntary’ is not a bad choice.9 Acting hekousiȏs is closely connected to acting hekȏn, and in many contexts ‘willing’ is the apt rendering of hekȏn, and ‘unwilling’ is apt for its opposite, akȏn. Aristotle chooses a term that suggests that an agent acts willingly, and not reluctantly. Why he chooses this term may be clearer when we have seen what distinctions he intends to draw. 99

Terence Irwin

These observations about the voluntary state Aristotle’s view about responsibility, since one is responsible (aitios) for one’s voluntary actions (e.g., EN 1111a29), but nor for one’s involuntary actions.10 When we say that A is aitios for breaking a window, we hold A responsible (aitiasthai) for it. If A is an inanimate object, being aitios implies only causal responsibility (1110b15), but if A is a human agent, being aitios normally implies the sort of responsibility that makes A open to blame for breaking the window. Hence it is often difficult to decide whether we should render aitios by ‘cause of’ or ‘responsible for’ or ‘to blame for’, since all three features are often combined.

3  Voluntary action and action from forethought A comparison of Aristotle’s views with Athenian law is not as easy at it would be if we could compare his views with contemporary legal textbooks or with other authoritative guides to the interpretation of the law. Since there were no textbooks, we have to derive our evidence from the speeches of the Athenian orators. Each speech advocates one side or the other in a given case, and we usually do not know what the other side said (except as far as we can infer it from the surviving speech),11 or how the case turned out. An advocate may not present a clear or impartial statement of what the law says, or of how it is normally understood. If it is in his client’s interest to confuse the jury about what the law says, or to apply a far-fetched interpretation, or to adduce irrelevant considerations that will arouse sympathy for his client, he may do any or all of these things. We should take an advocate’s statement as evidence of what someone might try to induce a jury to believe about the law and its application to a particular case; and this may be only indirect evidence of what the law says or of how it is normally applied. Keeping this qualification in mind, we can compare Aristotle’s remarks with the evidence provided by orators. It is easy to confirm Aristotle’s claim that voluntary actions are subject to praise and blame. As Antiphon says, those who do or undergo something voluntary are responsible (aitioi) for what happens (Tetralogies ii 2.6). Demosthenes describes the various Athenian courts that deal with voluntary and involuntary homicide, mentioning the different penalties. The law of homicide distinguishes voluntary (hekousios phonos, Dem. 23.77; 21.43) from involuntary killing. It is recognized as just that involuntary offenses should receive a lighter punishment than voluntary offenses receive (Dem. 23.73). The relevant laws about homicide and ‘wounding’ (i.e., violent assault with a potentially fatal weapon), say that agents who act involuntarily are exempt from punishment. But those who are liable to punishment are said to be those who act from forethought (pronoia); the Areopagus, for instance, is said to take account of killing from forethought and of wounding (Dem. 23.24). Similarly, other sources say that the Areopagus is competent to try charges of killing from forethought in contrast to involuntary killing (e.g., [Aristotle], Constitution of Athens, 57.3). They do not contrast voluntary with involuntary, or action from forethought with action without forethought. Nothing in the relevant laws, as far as we know, distinguishes action from forethought from voluntary action. Similarly, no surviving defence against a charge of killing or wounding from forethought argues that the accused acted voluntarily but not from forethought. None of these remarks suggests that action from forethought is ever involuntary. They show that ‘involuntary’ is a legal term, but ‘voluntary’ is not.12 To explain these facts, we might say that Athenian law recognized no difference between voluntary action and action from forethought, so that a homicide, for instance might be described indifferently as from 100

Aristotle on legal and moral responsibility

forethought or as voluntary.13 If the law had recognized voluntary homicide as a wider class than homicide from forethought, it would have allowed a defendant to escape punishment by showing that he acted voluntarily, but not from forethought. But the relevant laws remit punishment only for someone who has killed involuntarily or who has killed justifiably (e.g., in self-defence). Someone who had killed voluntarily but not from forethought would be neither liable to punishment nor exempt from punishment.14 If this is an accurate statement of the legal situation, we should treat statements that the law prohibits voluntary killing, and that it treats voluntary actions differently from involuntary, as comments on the law, not as attempts to quote its very words. The law speaks only of killing from forethought and of involuntary killing, but the speaker comments that this amounts to a distinction between voluntary and involuntary actions. The comment is informative to the extent that ‘voluntary’ and ‘from forethought’ are different concepts that we grasp well enough to see that the law prohibiting killing from forethought prohibits voluntary killing.15 Such comments are true and informative if voluntary actions are those that engage the will of the agents, because the agents want to do them, and if forethought involves previous planning and premeditation; premeditated actions involve the deliberation and choice of the agents. If this is the relation between voluntary action and action from forethought, the speaker who says that the law punishes voluntary actions and remits punishment for involuntary actions would be correct in one respect, since actions from forethought are all voluntary. But the speaker would not be justified in asserting that the law punishes all voluntary killing; it is not obvious that there are no voluntary killings without premeditation.

4  Conditions for forethought To answer these questions about voluntary action and action from forethought, we need to ask what is needed for forethought.16 The term readily suggests premeditation and planning in advance. In that case the ‘fore’ in ‘forethought’ refers to some time before the occasion on which A killed B, and the ‘thought’ refers to some planning by A about what A will do when A meets B. If, for instance, A brought a dagger to the meeting with B, and if A asked to meet B in a deserted place after dark, A’s actions support the charge that A  killed or wounded B from forethought. We might also, however, understand ‘forethought’ more broadly, so that it includes not only the sort of planning just described, but also the thought that precedes the action whenever we know what we are doing. If I know that my fist is about to make contact with your nose, and I do not withdraw it, I have had the relevant thought before the action of punching your nose, and in this respect I act from forethought. Actions that do not result from this sort of forethought would be actions in which I am not aware of what I am doing, or my awareness makes no difference to what I do (so that the action would not be ‘from’ forethought). Action from this sort of forethought may not involve any of the planning and anticipation that would constitute premeditation and planning. Even if I act on the spur of the moment, and even if a sudden impulse causes me to do the opposite of what I had planned to do, I act on forethought of this minimal sort. These two types of forethought are extremes on a scale that allows intermediate cases. Perhaps, for instance, A did not lay any specific plan to kill B, but A had a previous grudge against B and for this reason recognized the opportunity that arose when they met in a deserted place. Or perhaps A unintentionally got into a fight with B, because B provoked A; then A chose to retaliate by stabbing B with a dagger. In such cases the opportunity to kill 101

Terence Irwin

B might arise suddenly and without previous planning by A, but none the less A might be justly punished for killing B from forethought. With this continuum of degrees of forethought in mind, we can ask what Athenian law and Athenian juries required for forethought. Did they require maximum forethought (deliberate planning in advance), or only minimal forethought (knowing what one is about to do), or some degree of forethought between the minimum and below the maximum. One answer to our question is easy: Athenian law did not explain ‘forethought’, and therefore did not choose among these different possibilities. The answer about Athenian juries is more difficult, and our sources, not surprisingly, suggest different answers.

5  Arguments about forethought Some of the arguments that might be offered for and against a charge of killing or wounding from forethought can be gathered from Lysias’ speech ‘On wounding from forethought’ (Lys. 4). B has accused A of wounding B from forethought because A has wounded B by hitting B in the face with a potsherd. A admits the fact, but argues that it is no proof of forethought. A offers different arguments to show there was no forethought: 1. A and B were previously on good terms, having settled a previous dispute (1–3). 2. Even if A  were to concede B’s claim that they were enemies, there would still be no forethought (5). A did not know B would be at home. A arrived with his friends, having taken drink, and looking for boys and flute-girls; how could A’s wounding result from forethought in these circumstances? (7). If A had exercised forethought about attacking B, he would have brought a dagger; he would not have counted on finding a potsherd ready to hand in B’s house (7). 3. In any case it was B who struck the first blow, and A only retaliated (11). Despite all this B claims that A has plotted (epibouleuthȇnai, 10) against B. The arguments for each side in this case suggest how a case for forethought might be supported or rebutted. Since B prosecuted A  while knowing that A  had not brought a weapon with him and that B had struck the first blow, B apparently believed that a case could be made for forethought without any evidence of plotting in advance by A, and that A was guilty of wounding from forethought even if B struck the first blow. B would presumably argue that, even if both A and B had both had a bit to drink and got into a quarrel, B’s hitting A did not justify A’s assault on B with a potsherd, which A was sober enough to pick up and to use on B’s face. To show that it was not just a drunken brawl, B (as we gather from A’s defence) alleged that they had in fact been on bad terms, and that A had seriously wounded B, so that B had to be carried around in a litter. A replied that B had only got a black eye, and was only pretending to be seriously wounded. Antiphon’s Third Tetralogy raises some similar questions. It concerns a trial in which A was accused of murdering B, but A pleaded that it was a case of self-defence, and hence of justifiable homicide. The prosecutor admits that those who kill involuntarily deserve to be excused (sungnômȇs tuchein, 1.6). But in this case no plea of involuntariness is open to A; for A grabbed B when B was drunk, and struck and choked B (1.6). This assault shows that A did not kill B involuntarily. In his defence A  argues that B started it (arxanta, 2.2), and is therefore responsible (aitios, 2.1). B struck A in a drunken rage, and A simply fought back, as he was entitled to 102

Aristotle on legal and moral responsibility

do (2.2). In any case A did not kill B, even justifiably, because B’s death was the fault of B’s incompetent doctor, who gave B bad treatment (2.4). The law requires punishment of the one who plots (epibouleusanta, 2.6) the death of the victim, but A did not plot B’s death any more than B plotted A’s death by striking the first blow in a drunken rage. The prosecution reply that A  plotted B’s death, because B died as a result of what A did after having thought about what to do (dianoȇtheis, 3.4). They agree that A did not intend to kill B, and that B’s misfortune resulted from A’s mistake in killing B intending to. But since A struck the fatal blow, and struck after having thought, A is responsible for the misfortune that happened to B. B was a weak old man, and A was a strong young man (2.2) who, typically of young men, was stimulated by drink to strike back in anger. A’s disproportionate response to B’s feeble efforts at hitting A make A responsible for B’s death. In reply to this claim about what A had done after thinking about it, the defence insists that A’s thought (dianoȇtheis, 4.4) was to strike, not to kill, and that therefore A  is not guilty of murder. The prosecution answer A’s claim that it was the doctor’s fault for giving B the wrong treatment, by observing that B needed medical treatment because A had hit so hard (4.3). The defence answers that if this reasoning is accepted, then B is responsible; for it was because B hit A that A hit back. Hence A is not responsible. In these speeches the prosecution and defence speak of ‘plotting’ (epibouleuein), but they do not discuss (as the opponents in Lysias’ speech do) whether the two parties were on bad terms, so that one might have had a motive to plan an attack on the other in advance. To show that A plotted B’s death, the prosecution argue only that A responded to B’s provocation by thinking about what to do and choosing to respond disproportionately to B. In reply A claims to have simply struck B in anger in response to B’s provocation. We do not know how effective these arguments were, or how plausible the advocates for each side took them to be. But if we assume that the advocates did not think it was futile to try the arguments on a jury, we can draw some conclusions about how forethought might be understood. The arguments of the prosecution in each case presuppose that forethought has a wide scope, requiring only minimum forethought: 1. Earlier planning and deliberation are not taken to be necessary for forethought. In both of these cases it is agreed that A was not carrying out some design that he had planned in advance. A had not come with a weapon to a pre-arranged meeting in a deserted place. 2. The fact that some degree of provocation by B precedes A’s action does not exculpate A. In both cases it is agreed that B struck the first blow. 3. Though some right of retaliation is recognized, retaliation may be disproportionate. The prosecution do not suggest that A was required to do nothing, but they imply that A went too far. The arguments of the defence, however, presuppose that forethought has a narrow scope, requiring maximum forethought: 4. The defence try to show that A and B were not on bad terms at the time of the assault, and that A had not taken any steps to make it easier to kill or wound B (e.g., bringing a concealed weapon). 103

Terence Irwin

5. They try to show that the assault was not cold-blooded, but resulted from a quarrel in which both parties were drunk and angry. These arguments on each side suggest that jurors were not all firmly disposed to apply either the maximal or the minimal conditions for forethought. Apparently they might be persuaded to go in either direction, by an appropriate presentation of the case. The prosecution hope for a conviction even if they admit that A’s assault was a response to provocation by B in a drunken quarrel; they maintain that A recognized what he was doing, and that A’s response was disproportionate, going beyond justified retaliation. The defence hope for an acquittal even if they admit the points on which the prosecution insist, as long as they can show that A had not planned anything in advance, and was provoked into striking B back in the heat of the moment. If this is a fair summary of Athenian law and of jurors’ attitudes to it, jurors who wanted to exercise the appropriate judgment (gnȏmȇ, as described above) had a difficult task. They were expected to look for evidence of forethought or its absence, but what were they supposed to look for? Some jurors might vote to convict if the minimal conditions for forethought were satisfied. Others jurors might object that it would be too harsh to impose the same penalty on a defendant who satisfied only the minimal conditions and on one who satisfied the maximal conditions. Still, it might seem unduly lenient to exercise sympathetic judgment (sungnȏmȇ) and hence to acquit a defendant who chose to respond disproportionately to a provocation. The law rejects punishment for those who act involuntarily (akousiȏs), but the defendants in these trials for killing and wounding from forethought do not seem to have acted involuntarily.

6  Aristotle’s first response Aristotle recognizes that something is wrong with the options that Athenian law presents to jurors. He looks for options that will make it easier to reach just verdicts. He presents different options in different works. In MM he argues that the voluntary is what results from thought (dianoia). In support of his view he appeals to what we commonly say about voluntary and involuntary action. He takes it to be a common view that those who have acted without any forethought have done what they did unwillingly (akȏn). the remaining possibility is that it (sc. the voluntary) is what results from thought. . .. This is clear from actual examples. For whenever one person strikes, or kills someone, or does something of that sort, without any forethought, we say that he has done it unwillingly, on the assumption that the voluntary consists in having thought about the action. They say, for instance, that a woman once gave someone a love-potion to drink, and then he died from the potion. She was tried before the Areopagus. In her trial they acquitted her, precisely because she had not acted from forethought. For she gave the potion out of love, but she failed in her aim; that is why [her killing him] seemed not to be voluntary, because she was giving him the gift of the love-potion not with the thought of killing him. Here, then, the voluntary comes under the head of what is done with thought. (MM 1188b25–38)

104

Aristotle on legal and moral responsibility

Since the woman was tried before the Areopagus, she was charged with killing from forethought. She was acquitted because she had not acted from forethought, but had given the victim a love-potion that killed him, in the belief that it would arouse him, not in the belief that it would kill him. Whether she killed voluntarily or not depends on what thought she acted on. In this case the questions that arose in the more complex examples from Lysias and Antiphon do not arise. The woman’s advocate would not have tried to show that she killed him on an impulse after being provoked, but had not planned in advance to kill him. If she had thought the potion would be lethal, then she would have killed from forethought, but since she did not have this thought, she killed the victim involuntarily. Though Aristotle endorses the view that without forethought there is no voluntary action, the view that he really endorses, as he makes clear, is the view that without the relevant thought there is no voluntary action. The forethought that he takes to be necessary for voluntary action is the minimal forethought that simply involves the relevant belief at the time of the action (e.g., ‘I am giving him a love-potion’). We may be puzzled by what Aristotle says about forethought when he returns to it in his discussion of election (prohairesis). Election involves desire for an end and deliberation about the means to the end,17 resulting in an action that the agent takes to be a means to the end. Hence acting on election is not simply acting on thought. Hence not all voluntary action, as Aristotle has described it, is action on election. Since some voluntary actions are undertaken without thought (dianoia) and deliberation, they are not elected. If, then, election is a type of deliberative desire involving thought, what is voluntary is not [the same as] what we elect. For we do many things willingly before thinking and deliberating about them  – as, for instance, we sit down, stand up, and many other such things, willingly but without having thought about them – but all action on election involves thought, as we saw. The voluntary, then, is not [the same as] what we elect. But what we elect is voluntary  – for if we elect to do something after deliberation, we do it voluntarily. And indeed some few among legislators also appear to distinguish voluntary action from action on election as being different, since they make the penalties for [merely] voluntary actions smaller than those for actions on election. (MM 1189a31-b6) When Aristotle says that actions that are voluntary but not elected are undertaken without thought, he does not contradict his previous claim that all voluntary action is from thought. The thought that he takes to be unnecessary for merely voluntary action is the thought that belongs to prior deliberation. The thought that is necessary for all voluntary action is minimal forethought. Aristotle adds that a few legislators appear to distinguish voluntary action from action on election (1189b3–6). As far as we can tell from the evidence we have discussed, these few legislators do not include Athenian legislators, who do not distinguish killing from forethought from merely voluntary killing. Aristotle does not contrast action from forethought with merely voluntary action, since he has used Athenian law on killing from forethought to explain his conception of the voluntary. MM has consistently understood ‘forethought’ as minimal forethought.

105

Terence Irwin

7  Aristotle’s second response EE agrees with MM in identifying voluntary action with action on thought (1225b1, kata tȇn dianoian). The relevant kind of thought involves knowledge of the relevant features of the action (1225b1–10). In contrast to MM, however, EE does not appeal to common views about forethought as support for its views about the voluntary. In particular it does not mention the case in which the woman who administered the love-potion was found not guilty of murder from forethought. We can see why Aristotle does not mention forethought in this context, when we come to his discussion of the difference between acting on election and acting voluntarily. He gives the same reasons as in MM for believing that actions on election are a proper subset of voluntary actions, but he appeals for support to the distinction that legislators draw between voluntary actions, involuntary actions, and actions on forethought. Their distinction is inexact, but it has some grasp of the truth. We have found that if someone does or refrains from doing what it is up to him to do or not to do, and does it because of himself and not because of ignorance, he willingly does it or refrains from it. Moreover, we do many such things without previous deliberation or forethought. It necessarily follows, then, that whatever is elected is voluntary, but not everything that is voluntary is elected, and that all the actions based on election are voluntary, but not all voluntary actions are based on election. At the same time it is clear from this that those who distinguish voluntary, involuntary, and premeditated actions in their legislation, are drawing the right distinctions – for even if they are not completely exact, nonetheless they have some hold on the truth. But we will discuss these questions in the examination of justice. (EE 1226b30–1227a3) MM does not mention forethought at this point, but mentions the few legislators who distinguish voluntary action from action on election. Neither MM nor EE tells us which legislators Aristotle has in mind; hence it is not clear whether he has the same legislators in mind, but describes the distinctions that they draw differently, or he has different legislators in mind. Since MM says there are only a few who draw the distinction Aristotle endorses, and EE does not say how many there are, it is not obvious that he has the same ones in mind in each context. EE may appear to contradict MM. EE distinguishes voluntary action from action on forethought, whereas MM identifies them, Moreover, EE identifies action on forethought with action on election, whereas MM distinguishes them. The appearance of contradiction, however, is misleading. We saw that MM assumes minimal conditions for forethought. EE assumes maximal conditions; if it did not, its claim about the difference between voluntary action and action on forethought would be false. Given the two sorts of conditions for forethought, neither work contradicts the other. We have no reason to believe that when Aristotle introduces his threefold divisions (involuntary, voluntary, from election (from forethought)), he is trying to describe Athenian law. We have noticed that Athenian laws and forensic rhetoric uses all these four terms (involuntary, voluntary, forethought, election), but they recognize only two classes of actions (involuntary and from forethought). Aristotle appeals to common views only in MM, when he takes lack of forethought – which he understands as minimal forethought – to imply that 106

Aristotle on legal and moral responsibility

an action is involuntary. We learn from both ethical works that Aristotle is familiar with non-Athenian legal systems that capture the appropriate distinctions better than Athenian laws capture them. EE refers forward to a further treatment of the distinction between voluntary action and action on forethought, which he says (1227a2–3) he will treat in the section on justice. We know what Aristotle says in this later treatment, if and only if the relevant section of EN v = EE iv belongs to EE.18 In this section Aristotle distinguishes voluntary action from action on election, and maintains that only the latter type of action results from a virtue or vice (1135b19–25). This is why it is just to distinguish actions from anger from actions from forethought; for the one who starts it is not the one who responds in anger, but the one who provokes the anger. But whenever one does the harm from election, one is unjust and vicious. That is why actions on spirit are judged not to be from forethought; for the one who originates it is not the one who acts on spirit, but the one who provoked him to anger. (EN 1135a25–7) The fact that someone commits an injustice as a result of provocation does not show that he is unjust. To condemn him as unjust, we also have to know that he does this sort of thing not because of provocation but because he has deliberately chosen this unjust action as a means to the unjust end he has set for himself. Aristotle does not say that this division between acting from anger and acting from forethought is drawn in Athenian law, and our review of the evidence suggests that it is not drawn there. Prosecutors in cases about killing and wounding from forethought are not deterred by evidence that the defendant struck the victim in anger, or even by evidence that the victim started it. They do not try to undermine such evidence, because they assume that, even if it is admitted, the accused is guilty of acting from forethought. In this respect Aristotle advocates the exercise of sympathetic judgment (sungnȏmȇ) in cases where Athenian law is less lenient. His leniency, however, is not wholly alien to Athenian forensic argument. We have seen that the defence in such cases, while admitting that the assault took place, deny that there was any advance planning or any reason for the accused to have wanted to harm the victim; and they emphasize that the accused struck the blow only when he was provoked to anger by a previous attack by the victim. These arguments assume that the jury might be moved by the thought that the victim started it, and that it was a momentary lapse, rather than the result of a plot, on the side of the defendant. If such argument moves a jury, the jury recognizes the justice of Aristotle’s arguments. Aristotle claims to make it easier for the jury in such cases, because he does not require them to agree that if the accused killed from anger and not from forethought, he did nothing wrong and does not deserve any punishment. Since he has voluntarily committed an injustice, he is not to be excused on the ground that he acted involuntarily. But he deserves less punishment than he would deserve if he had acted on previous deliberation and election, and had thereby shown himself to be an unjust person. This discussion of forethought fits the earlier discussion that refers forward to a later treatment. In both passages Aristotle’s argument relies on maximal conditions for forethought. Such conditions are neither required nor rejected by Athenian law. Aristotle prefers a law that distinguishes voluntary action from action on forethought. Such a law would 107

Terence Irwin

avoid both the injustice of equal punishment for vicious and merely impulsive people, and the injustice of excusing those who act voluntarily though not from forethought. Aristotle’s remarks on anger do not settle all the questions that might arise. We might agree that an angry response provoked by a previous assault does not necessarily express someone’s character. But not all responses caused by anger are actions on a momentary impulse. Medea was angry with Jason, but she took her time to plot her revenge. She was different from someone who hits back in the heat of the moment.19 Aristotle may reply that in such cases the agents do not merely act from anger, but the anger affects their deliberation and election. Such cases suggest that the division between voluntary actions and actions on forethought still leaves certain questions open that are not directly answered by Aristotle’s distinction between doing injustice (a voluntary action) and acting unjustly (the result of one’s character). But his division at least introduces some clarity into an area where Athenian law creates avoidable difficulties for a conscientious juror who tries to exercise sound judgment about responsibility.20

Notes 1 I assume that this is their order of composition. The mss of EE say that three books (EE iv– vi = EN v–vii) are common to EE and EN. Many students believe that these three books, as they appear in our mss, were originally composed for EE. I agree with those who believe that the books were revised for EN. Translations of Aristotle and other Greek texts are my own. 2 A satisfactory account of the voluntary should apply to omissions as well as to actions. For this purpose we might appeal to Aristotle’s remarks on culpable ignorance and on negligence and carelessness in EN 1113b30–1114a7. 3 These remarks raise questions about the connections that Aristotle sees between being voluntary, being up the agent (EN 1110a14–18, 1113b7–14, EE 1225b8) and being avoidable (EN 1113b26–30). 4 On the size of Athenian courts see MacDowell 1963: 41–42. The jury that narrowly found Socrates guilty probably had 501 members. 5 Hence the cognate verb sungignȏskein often refers to excusing or pardoning. 6 A work on laws is mentioned in the catalogue given by Diogenes Laertius, v 26. 7 Some voluntary actions are also subject to sungnȏmȇ. I will not discuss them further. 8 See Charles 2012; Gauthier & Jolif 1970, vol. 2:1, 170. 9 See Maschke 1926: 54. 10 This claim about involuntary actions may need to be modified in the light of, e.g., EN 1113b30. 11 An exception is Antiphon’s Tetralogies, which include speeches on each side. These speeches are probably exercises, rather than speeches in actual cases. 12 Strictly speaking, the evidence just presented only shows that ‘voluntary’ does not appear in the laws about killing and wounding. But I am not aware of any counter-instances in other laws. 13 See MacDowell 1978: 115; Loomis 1972; Phillips 2007. Contrast Carawan 1998, ch. 1. 14 Stroud (1968: 41), observes that lack of premeditation does not imply involuntariness. He suggests that a reference to voluntary but unpremeditated homicide may have fallen out of the surviving fragments of the inscription containing Dracon’s law. MacDowell (1963: 124–125) considers the possibility that relatives of the victim could pardon someone who had committed an involuntary but unpremeditated homicide. Against this possibility he cites Dem 21.45, which mentions only homicide from forethought (punished) and involuntary homicide (treated with much sympathy). 15 If we are to agree that no action from forethought is involuntary, we have to take ‘from’ to imply causation. If someone is using irresistible force to push me through a window, I may be well aware in advance that I will break the window, but since my awareness makes no causal difference, I am not breaking the window from forethought. 16 The following remarks are indebted to the enlightening essay Harris 2013.

108

Aristotle on legal and moral responsibility 17 The deliberative element in prohairesis shows that ‘choice’ would be an unsuitable translation. Merely voluntary action is chosen, though not elected. Aristotle does not confuse choice (hairesis) with prohairesis. 18 On the books common to EE and EN see n.1. In contrast to EE, EN has no earlier discussion of forethought, and has not suggested that action from forethought is the same as action on election. 19 Aristotle draws an unsatisfactory contrast between anger and appetite on this point, at 1149b13–18. 20 I am indebted to helpful comments by the editor, and by participants in a workshop on Ancient Law and Ancient Philosophy in January 2023, and especially to written comments by Susan Sauvé Meyer. I have not been able to take proper account of them all in this short essay.

Further reading A general discussion of Aristotle’s views on the voluntary and on responsibility: Meyer, S. S. (2011). Aristotle on Moral Responsibility: Character and Cause (2nd ed.). Oxford: Oxford University Press. Notes on some of the relevant texts in Aristotle can be found in Aristotle. (1992). Eudemian Ethics I, II, VIII, trans. M. J. Woods (2nd ed.). Oxford: Oxford University Press. Aristotle. (2019). Nicomachean Ethics, trans. T. H. Irwin (3rd ed.). Indianapolis: Hackett. The most helpful introductions to Athenian views on the relevant questions are MacDowell, Law in Classical Athens, and Harris, ‘Is Oedipus guilty? (both cited previously).

References Antiphon. (1953). “Tetralogies,” in K. J. Maidment (trans.), Minor Attic Orators, vol. 1. Cambridge, MA: Harvard University Press. Aristotle. (1984). “Nicomachean Ethics, Eudemian Ethics, Magna Moralia, Rhetoric, and Constitution of Athens,” in J. Barnes (ed.), The Complete Works of Aristotle, vol. 2. Princeton: Princeton University Press. Carawan, E. M. (1998). Rhetoric and the Law of Draco. Oxford University Press. Charles, D. (2012). “The Eudemian Ethics on the ‘voluntary’,” in F. Leigh & B. Inwood (eds.), The Eudemian Ethics on the Voluntary, Friendship, and Luck (pp. 1–28). Lanham, MD: Brill. Demosthenes. (1949–62). Orations. 7 vols. Cambridge, MA: Harvard University Press. Diogenes Laertius. (1931–8). Lives of the Philosophers, trans. R. D. Hicks. 2 vols. Cambridge MA: Harvard University Press. Gauthier, R.-A., & Jolif, J.-Y. (eds.). (1970). Aristote: L’Ethique à Nicomaque (2nd ed.), 4 vols. Louvain: Publications universitaires de Louvain. Harris, E. M. (2013). “Is Oedipus Guilty? Sophocles and Athenian Homicide Law,” in Law and Drama in Ancient Greece (pp. 122–146). London: Bloomsbury. Loomis, W. T. (1972). “The Nature of Premeditation in Athenian Homicide Law,” Journal of Hellenic Studies, 92, 86–95. Lysias. (1930). Orations, trans. W. R. M. Lamb. Cambridge, MA: Harvard University Press. MacDowell, D. M. (1963). Athenian Homicide Law. Manchester: Manchester University Press. ———. (1978). Law in Classical Athens. London: Thames & Hudson. Maschke, R. (1926). Die Willenslehre im griechischen Recht. Berlin: Stilke. Phillips, D. D. (2007). “Trauma ek pronoias in Athenian law,” Journal of Hellenic Studies, 127, 74–105. Stroud, R. S. (1968). Drakon’s Law on Homicide. Berkeley: University of California Press.

109

8 KANT ON ABSOLUTE RESPONSIBILITY AND TRANSCENDENTAL FREEDOM David Sussman

Each of us is guilty before everyone for everyone, and I more than the others. – Fyodor Dostoeveysky, The Brothers Karamazov

Immanuel Kant offers a conception of moral responsibility so extreme that it is often taken to be incoherent.1 For Kant, an agent is morally responsible only for deeds that are “imputable” to her in a quasi-legal sense: Imputation (imputatio) in the moral sense is the judgment by which someone is regard as the author (causa libera) of an action, which is then called a deed (factum) and stands under laws. If the judgment also carries with it the rightful consequences of this deed, it is an imputation having rightful force. . . . The (natural or moral) person that is authorized to impute with rightful force is called a judge or a court (iudex s. forum). (MM 6:227) When I am responsible for wrongdoing, my misdeed is not merely attributable to me in a way that merits some negative evaluation of me or my character. To be responsible for a wrong is in part to deserve to suffer for so acting: “the rightful effect of what is culpable is punishment” (MM 6:227). Admittedly, Kant is not the simple retributivist about criminal punishment that he is often taken to be.2 However, he does take it to be an a priori truth that moral wrongdoing immediately calls for the infliction of proportionate harm, regardless of any further effects such punishment may or may not produce: But if someone who likes to disturb peace-loving people finally gets a sound thrashing  .  .  . everyone would approve of it and take it as good in itself even if nothing further resulted from it; indeed, even the one who received it must in his reason recognize that justice was done to him, because he sees the proportion between well-being and acting well, which reason unavoidably holds before him, here put into practice exactly. (CPrR 5:61) DOI: 10.4324/9781003282242-13 110

Kant on absolute responsibility and transcendental freedom

Kant is not here advocating vigilantism; later in the Doctrine of Right, he makes clear that deserved punishment may only be meted out by the state, and even then the state’s authority to punish is importantly conditioned by non-retributive concerns. For Kant, no temporal power could ever be fully entitled to give people their just deserts; God alone may (and will) render such “divine justice” (MM 6:489, R 6:74). Unfortunately, Kant gives no explicit argument for his moral retributivism beyond the assertion that such desert is contained in very concept of morality itself: “there is in the idea of our practical reason something further that accompanies the transgression of a moral law, namely its deserving punishment” (CPrR 5:37, my emphasis). For Kant, to be a subject of imputation is just what it is to be a “person” rather than a mere “thing.” Unlike things, persons possess a special “dignity” that demands respect, precluding being treated as a mere means, resource, or obstacle to other people’s plans (G 4:428). Kant takes it to be a conceptual truth that a person is morally responsible just insofar as she is a free agent. Kant also argues that if we are properly governed by any valid moral principles at all, then we must be ultimately bound by the Moral Law (as he understands it). The Moral Law (or “Categorical Imperative” as it is addressed to imperfect beings like us) is the unconditional and authoritative principle that specifies the basic form of moral justification and action. To recognize this law is to see ourselves first and foremost as fellow citizen–legislators in a universal “Kingdom of Ends.” As such citizen–legislators, we relate to one another simply as free and equal persons engaged in the joint enterprise of defining and inhabiting a social world together. For Kant, this self-understanding involves a commitment to only act in ways that could at least be consistent with such norms for their common life that equal, rational persons might be able to collectively establish for themselves (G 4:443).

1  Morality and the “intelligible world” For Kant, the Moral Law determines our fundamental obligations independently of any of the ways in which we might differ from one another as individual agents, with respect to our particular desires, attachments, or social positions. Kant does not deny that such facts about our personal differences can be morally important. He holds only that the significance of such differences is itself to be derived from a starting point where we confront one another simply as persons, accountable to each other simply as persons, in abstraction from any features of our particular circumstances or individual identities. Kant claims that we must recognize ourselves as morally bound in this way if we are to be able to see ourselves as being free or autonomous in the sense needed to support any proper attributions of responsibility. Supposedly, we must see ourselves as accountable without qualification to this ultimate law if we are to make sense of ourselves as having the power to act in a way distinct from the casually-determined order of nature. For Kant, a self-conscious agent must attribute such a power to himself if he is to lay claim to “a will of his own” (G 4:448). Such a will defines an agent’s “moral personality,” which serves as the ground of personhood in general. For Kant, an agent must see herself as unconditionally responsible in this way if she is to be able to recognize herself as the self-same agent in different times, situations, or psychological conditions. Such open-ended responsibility is a precondition of my not just living a life (as animals do), but of being able to lead that life through an understanding of it as my life. For Kant, our responsibility to one another is not merely an important aspect of our interpersonal or political relations, bearing on the enforcement of basic social and legal 111

David Sussman

norms needed to make cooperation possible. For philosophers such as Hobbes and Locke, we can form coherent notions of an agent outside of any practices of mutual accountability (i.e., in a “state of nature”), and we can use that conception as a conceptual baseline against which to assess such practices. For Hobbes, even though human life in a state of nature would necessarily be “nasty, brutish, and short” (Leviathan I.13.9, (p. 96)), that sort of life would still be intelligible enough to serve as a contrast case in our theorizing about morality and justice. For Kant, however, responsibility is a defining feature of the human essence from the start. Responsibility is the “formal cause” that allows human persons to be something interestingly different from animals and other objects in nature. As such a cause, the idea of responsibility provides a distinctive standard of what it is for a human being to be in good shape as the kind of thing she is, defining what is to count as defect or disorder in creatures like us. For Kant, responsibility is not just an ethical, political, or legal concept (like property) that applies to some limited sphere of human life. Rather, responsibility is a fundamental metaphysical notion, inextricably tied up with the very possibility of practical reason and self-conscious action. On this view, no sense can be given to the idea of who or what I am as a person without presupposing the idea of a kind of unrestricted responsibility to other persons. Such responsibility cannot be qualified by anything like the concrete identities I might embrace or the social roles I occupy, simply because there can be no “me” outside of such relations of accountability for such conditions to apply to. In the first instance, we are each wholly responsible, simply as persons, to all other persons, simply as such. Any further responsibilities (relative to particular social positions or relationships) can only be filled in within the space delineated by this most basic, absolute conception. Our most fundamental identity is to be found, not in any particular social or even biological relations, but in membership in the “intelligible world,” where we understand ourselves as mere persons in abstraction from any distinctive qualities that might distinguish us. Here we do not merely understand ourselves as we appear (as “phenomena”), to be made sense of scientifically as material objects in space and time bound by causal laws to the rest of the natural world. Rather, as “intelligences” (or “noumena”) we understand ourselves as nodes in a web of ethical relationships: constituted not by causal forces, but by distinctively normative powers and liabilities (rights, duties, responsibilities, liberties, etc.).

2  Morality and the “empirical world” Kant allows that even if we are not really of nature, we are still very much in it. Human agents, although fundamentally residents of the intelligible world, can still only act through their physical bodies that are ineluctably parts of the phenomenal (or “sensible”) world, governed by the general causal laws of that world that make possible a determinate spatiotemporal order. Similarly, Kant allows that human action must in some way issue from our psychological states and character, which he takes to similarly depend on the deterministic causal structure of the world. Kant believes that if our deeds are not integrated into the natural order in this way, we would be unable to understand actions as datable events unfolding in a definite temporal sequence, with some stages of our activities (breaking the eggs) coming before and being made sense of in terms of the next ones (beating the yolks, adding butter, etc.). To act is not merely to launch a decision and then sit back and observe its trajectory in the world. Rather, actions of any complexity require that we continuously monitor and adjust our performance to our changing circumstances. Kant realizes that 112

Kant on absolute responsibility and transcendental freedom

human beings can have knowledge of such circumstances only through empirical intuition (or sense-perception); to act we must be able to understand our acts as parts of the natural world so known. Kant’s notion of unconditional, absolute responsibility may be starting to look very strange. On the one hand, if our acts are to be morally assessable at all, they must be expressions of a profoundly free, and so non-natural, power of rational self-determination. On the other hand, if we are to have any hope of understanding our actions in temporal terms, those acts must also be parts of the natural order of causally determined events. Kant emphatically rejects any sort of psychologistic compatibilism, where the requisite freedom would be just a matter of our doings having the right kinds of mental causes (what Kant derides as a “wretched subterfuge” that secures only “the freedom of the turnspit” (CPrR 5:96–7)). Instead, Kant insists that moral responsibility presupposes “transcendental freedom,” i.e., the power be a true “first cause,” and so to “spontaneously” start a causal chain in a way that is not necessitated by any features of the state of the world prior to that beginning. Such freedom must be thought as independence from everything empirical and so from nature generally, whether it is regarded as an object of inner sense in time only [our psychological states] or also of outer sense in both space and time; without this freedom (in the latter and proper sense), which alone is practical a priori, no moral law is possible and no imputation in accordance with it. (CPrR 5:97)3 Kant seems to have painted himself into a corner; we can make sense of our acts morally or temporally, but not both simultaneously. Yet a person can be held responsible for any particular act only if she can be understood as both an intelligible and empirical being in the very same judgment.

3  The intelligible and empirical united: responsibility for the whole world Kant does offer a solution to this dilemma, but it may be worse than the problem. Kant holds that if X causally necessitates Y, then we can be morally responsible for Y only insofar as we are morally responsible for X as well. Supposedly, for our acts to be determinately located in the spatiotemporal order, they must be causally necessitated by our prior physical and mental states, states that have to themselves be necessitated by even earlier such states, etc. Such a causal chain cannot have any beginning in time. For Kant, the connexion between causation and time entails that every efficient cause must be the effect of an earlier efficient cause. Kant has to conclude that, in order for a person to be truly responsible for any act she performs, she must be responsible for all the causal laws and conditions that figure in the explanation of that act, no matter how remotely. I must have freely chosen the causal laws governing my entire earthly life, even those laws that were operating long before my birth: [T]o remove the apparent contradiction between the mechanism of nature and freedom in one and the same action . . . a rational being can now rightly say of every unlawful action he performed that he could have omitted it even though as appearance it is sufficiently determined in the past and, so far, is inevitably necessary; for 113

David Sussman

this action, with all the past which determines it, belongs to a single phenomenon of his character, which he gives to himself and in accordance with which he imputes to himself, as a cause independent of all sensibility, the causality of those appearances. (CPrR 5:98, my emphasis) Kant believes that the laws of nature must be systematically connected to one another if they are to define a coherent spatiotemporal order (the laws must make up a “totality” if they are to determinately constitute a “world”). If so, then to be responsible for any of my deeds, I must have freely established not just the natural laws immediately involved in my act; I must also be responsible for the system of natural laws in its entirety. Because the totality of such laws is what defines the temporal order, any free choice of this system would itself have to somehow occur completely “outside” of time. As usual, bringing in other people just makes matters worse. Not only does my responsibility for my acts presuppose that I am responsible for the entire order of nature: such must be equally true of any other responsible agent with whom I share that world. Kant is driven to the conclusion that, for anyone to ever be responsible for anything, everyone must always be responsible for everything. To be a person, one must be God, grounding and continuously sustaining the world from outside of time and space.

4  Some rescue attempts Kant’s defenders have tried to avoid this seemingly absurd conclusion in a variety of ways. Allen Wood allows that Kant’s picture of our agency would indeed require “staggering revisions” in our conception of rational agency, but argues that Kant is not really committed to this view. Instead, Kant is offering these claims only speculatively, merely to show that the ideas of transcendental freedom and empirical determinism are not logically inconsistent. Wood argues that Kant is completely agnostic about how intelligible agency could really work; the idea of freedom is only of use in practical thought, and for that the mere logically possibility of transcendental freedom is enough.4 Other commentators have been less confident that the theoretical and practical dimensions of freedom can be so neatly separated from one another. Derk Pereboom argues that without a plausible account of the real (and not merely logical) possibility of freedom, we will have to jettison the strong retributive components of Kant’s understanding of moral responsibility.5 However, Perebloom believes that much of Kantian morality can be salvaged despite this revision  – in particular, the fundamentally imperatival character of morality along with the less punitive aspects of our practices of holding one another responsible. Joe Saunders, in contrast, argues that we should just give up Kant’s insistence that an exercise of transcendental freedom must be outside of time.6 Admittedly, doing so is tantamount to abandoning transcendental idealism altogether; at the very least, such a retreat would expose claims about free agency to empirical challenge, something Kant understandably very much wanted to avoid. Other apologists have taken Kant’s claims about intelligible agency more seriously but have argued that the way of conceptualizing ourselves as natural beings is so different from how we understand ourselves as free agents that seemingly contradictory judgements can be true of the same act (e.g., S was causally determined to lie; S lied of his own free will). These approaches draw heavily on Henry Allison’s “Two-aspect” reading of transcendental idealism, where the difference between the intelligible and sensible worlds is not that 114

Kant on absolute responsibility and transcendental freedom

between two ontologically distinct realms, but rather that of two different ways of making sense of the one world, human agents included.7 The central thought here is that every act can rightfully be seen as an event in the world involving causal relations between bodies. So understood, every act is causally determined. However, our acts can also be understood as intentional doings, chosen in the light of reasons rather than impelled by causal forces. This way of understanding ourselves, which is neither more or less real than the scientific, cannot be translated into or reduced to causal explanations; here the idea of determinism is not even denied, since it cannot find any purchase at all.8 Even Kant’s most stalwart defenders tend to dismiss Kant’s claim that as intelligible beings, we freely establish the natural laws that govern our behavior in time. However, in recent work Ben Vilhauer has argued that we can take this position seriously if we attend to later developments in Kant’s Critique of Judgment.9 Vilhauer contends that Kant comes to hold that the sense in which the world must be causally determined is only formal. That is, for there to be a determinate, knowable spatiotemporal order, the events of that order must stand under some causal laws or other. Supposedly, this does not mean that those laws have to always have the same content. Vilhauer contends that when I freely choose to do something, I “fill in” the content of the causal laws governing my behavior in a particular way. Yet these new laws do not dictate my future behavior in any worrisome way; with each new free choice, the content of the laws covering the entirety of my behavior changes accordingly, like a continually updated cover-story. Vilhauer’s strategy is ingenious, but it but it comes at a high price. Vilhauer admits that on this view, the laws of nature cannot constitute a “totality”; that is, a determinately unified body of explanations that makes sense as a whole. Moreover, the natural world I inhabit would have to be different from the natural world inhabited by other agents; indeed, it’s not clear that I could really see myself as living in the same physical world as myself throughout the course of my life. Vilhauer admits that his reading requires that there be “single-shot” causal laws – that is, laws that are in principle unrepeatable, applying only to one particular event at one particular time and place. One does not have to be wedded to a Humean “constant-conjunction” model of causation to have a hard time recognizing this as any sort of natural law at all. Even if Kant’s metaphysics can countenance such a conception of causal law, we would have to abandon any hope of gaining any empirical knowledge of the world, insofar as human beings are acting within it. The idea of a totality would have to be abandoned not just as a constitutive principle, but as a regulative one as well.

5  Choosing ourselves: the predispositions to goodness Like Vilhauer, I take seriously Kant’s claim that we freely establish the laws of our empirical character. But to make sense of this I look not to the Critique of Judgment, but instead Religion within the Limits of Reason Alone, in which considerations of freedom and culpability take center stage. In the Religion, Kant explicitly argues that as moral agents, each human being makes a unique, unchangeable, and timeless choice of her entire psychological character and biography, a choice that then manifests itself to us finite beings as an unfolding causal process in time. Kant tells us that true moral goodness (“virtus noumenon”) “cannot be effected through gradual reform [“virtus phenomenon”] but must be effected through a revolution in the disposition of the human being” (R 6:47). Nevertheless, this unique intelligible choice is manifested in time as just such a gradual process, a process that can only be fully appreciated by a being that grasps the whole story in one intuition: “for 115

David Sussman

him to whom this endless progress is a unity, i.e., for God, this is the same as actually being a good human being (pleasing to him); and to this extent can be considered a revolution” (R 6:48a). On this picture, human beings do not really make free choices within their lives in any ordinary way (at one time rather than another, some before or after others, etc.). Instead, what we freely choose is our biography and psychology as a whole (“a single phenomenon of character” (CPrR 5:98)), a totality that cannot “be thought as divisible into temporal segments but rather only as an absolute unity” (R 6:69n). This choice must include the whole of the natural world in which we are embedded, along with the datable actions and psychological character of everybody else. Kant’s account of responsible agency seems to end up depriving us of anything recognizable as human action at all. However, there may be a way of reading Kant’s final position that is not quite so crazy. To see this, we need to attend to the ontological role that moral responsibility plays in Kant’s thought. For Kant, the free and responsible subject is not just a new being to be somehow inserted into the rest of reality, shoe-horned between the natural ones. Rather, for Kant the idea of moral responsibility inaugurates a whole new order of reality; that is, a logically distinct way of making sense of things as a context of meaningful human action and interaction. Kant may not be introducing new beings into the world, but only calling our attention to a fundamentally different way of being in general. The ontological dimension of responsibility comes out most clearly in the Religion’s discussion of human nature. There, Kant explains that the distinctively human form of being is constituted by three “predispositions to good” that are parts of the very “possibility of human nature” (R 6:28). These predispositions specify the essential features that make human beings something importantly distinct from other natural or living things. Kant tells us that the predispositions to goodness are “animality . . . as a living being”; “humanity . . . as a living and at the same time rational being”; and “personality, as a rational and at the same time responsible being” (R 6:26). For Kant, the predispositions form an ascending series, with the human being only stepping forth with the realization of personality, which is nothing less than “the idea of humanity considered wholly intellectually” (R 6:28). These predispositions apparently represent three aspects of the “formal cause” or distinctive “life-form” of the human being. Animality, humanity, and personality do not merely name different properties or abilities that human beings typically have but that other things lack. Rather, the predispositions set forth different rational standards (Kant’s technical sense of “idea”) for making sense of the human being as a unified agent, as something that has a distinctive power to do something rather than just be acted upon. All of these ideas are partially normative; they specify what the human being is by introducing some background notion of what it is for such an agent to be acting well (or being well disposed to action). Each predisposition brings with it a distinctive scheme of excellences and defects, as ways of being or failing to be in good shape as the kind of agent one is. The predisposition to animality involves the sort of teleological assumptions needed to recognize something as a sentient form of life in the first place, capable of purposive, selfdirected movement and voluntary behavior. To have an animal nature, a being’s doings must be intelligible in terms of such activities as growing, healing, eating, hunting, and reproducing as conducted through forms of perception, desire, and feeling more generally. Kant explains that while this aspect of our life-form does not involve reasoning proper, animality nevertheless has a kind of rationality about it as a whole. To count as an animal, a creature’s behavior has to generally make sense in terms of such basic life-activities. 116

Kant on absolute responsibility and transcendental freedom

Moreover, the animal’s basic capacities of perception and desire must be able to do a minimally decent job of realizing such activities in that being’s normal environment. Of course, a “minimally decent job” is consistent with significant failures and defects, which Kant here identifies as the “natural vices” of gluttony, lust, and “wild lawlessness.” The predisposition to humanity takes these basic life-activities (and so, structures of perception and desire) and translates them into a new form, with a new standard of success and failure. For Kant, the level of humanity involves reflective pursuit of a conception of happiness, some articulate notion of what a good or desirable life involves. This predisposition is a predisposition to humanity because it is only here that a distinctively human form of life begins to emerge, something with no real analogue in the rest of nature. Animals may in some sense sees things in the world as good or bad for them (and pursue them as such), but they do not deliberate and act from any abstract representation of what they value overall. In order to be guided by such a general idea of my own good, some sort of conceptual language of desirability must be available to me. Such concepts in turn presuppose that I am embedded in some culture and history that supplies the background roles and expectations from which I can start reasoning about what is really important to me in life. With the predisposition to humanity, a new kind of active power or agency becomes available to us. With general evaluative concepts, we are now capable of acting not just from motives (as do animals), but from reasons: that is, from considerations that can be offered by way of explanation and justification of our actions to others (thereby making it possible for us to explain and justify our actions to ourselves.) At this level, we are not beings that can be held accountable to standards merely in the sense of being properly assessed as physically or psychologically healthy or disordered. Rather, here we can evaluate ourselves and act with an eye to such standards, inhabiting a form of life governed by pride and shame, honor and disgrace. Although this is not responsibility in the full-blooded sense that Kant associates with moral personality, it is nevertheless a necessary stage along the way. With such new standards, and new ways of being held accountable to those standards, come new forms of being badly off with respect to action (manifested as the “cultural vices” of jealously, envy, and ingratitude). However, we have not yet attained humanity even with the predisposition to humanity. What emerges at this level is only the “proximate matter” characteristic of the human: that is, the basic human powers and capacities that still need to be given the distinctively human form or structure. The human being finally shows up only with the attainment of “personality” where we reinterpret ourselves as not just rational but as essentially responsible beings. Kant makes it clear that this kind of accountability is not just that of holding oneself to shared evaluative standards. Rather, the accountability of personality is that of interpersonal justification by appeal to common laws, laws that define not just evaluative standards, but interpersonal demands (and so obligations, rights, prerogatives, etc.). It is at the level of personality that we are introduced to the distinctive way of making sense of ourselves characteristic of the intelligible world. For Kant, such responsibility has an essentially political or juridical form. Here we relate to others (and so, to ourselves) not just as beings competing to excel in the same way (in terms of pride, shame, esteem, contempt, etc.). Rather, at this level we relate to one another as equal members of some joint activity, as something like fellow citizens of a republic without any borders (the universal “Kingdom of Ends”). Once again, with a new logic of accountability comes new ways of making (or failing to make) sense of ourselves as agents. Relations of pride and shame give way to more basic forms of guilt, blame, and resentment. 117

David Sussman

The conceptions of the good that we seek for ourselves become secondary in our own eyes to the obligations we owe others. Although the predispositions to humanity and personality cannot be really separated, neither are they just the same thing. Kant claims that at the level of humanity, an agent “might apply the most reflection to [its conception of happiness] without thereby even suspecting the possibility of such a thing as the absolutely imperative moral law” (R 6:26n). Kant continues Were this law not given to us from within, no amount of subtle reasoning on our part would produce it. . . . Yet this law is the only law that makes us conscious of the independence of our power of choice from determination by all other incentives (of our freedom) and thereby also of the accountability of all our actions. (R 26n) Here Kant is alluding to the “Fact of Reason” that he introduced in the second Critique. In the Critique of Practical Reason, Kant gives up any aspiration to ground moral obligation and accountability in a non-moralized sense of freedom that we could attribute to ourselves without presupposing any moral claims or commitments. Instead of trying to derive the Moral Law from any prior notion of freedom of the will, Kant instead takes as primitive the recognition that we have unconditional obligations to one another simply as persons to persons. This recognition is now offered as an underived and underivable “Fact of Reason” (Faktum der Vernunft), a claim that we supposedly cannot deny once we hear it addressed to us.

6  Forced to be free Kant may appear to be just dogmatically appealing to intuition, as little more than philosophical table-pounding. Yet although he no longer pretends to be giving us any justification of the Moral Law, Kant does hope to show that no such justification is really needed (providing what he calls “defense” rather than “justification” (G 4:459)). To confront the Moral Law this way is to see oneself as now accountable for all of one’s actions, and so, to be free in a profound way. Although freedom remains the “ratio essendi” (essence) of the Moral Law, the experience of seeing oneself as accountable to that law is the “ratio cognoscendi” (the grounds of knowing) of that very freedom (CPrR 5:4n). For Kant, must first see ourselves as accountable to each other simply as persons if we are to be able to claim or even comprehend the kind of freedom that is itself the ultimate ground of just such moral responsibility. Credo ut intelligam (“I believe in order to understand”). For Kant, the freedom presupposed by morality need not be anything intelligible outside of the moral outlook: certainly not the “contra-causal freedom” of the metaphysical libertarian or the psychological harmonies of the naturalistic compatibilist. Instead, freedom is just another word for having satisfied the (prior) conditions of moral responsibility, of acting in a way consistent with the idea of unlimited interpersonal accountability. Since freedom is only supposed to be comprehensible from within an ethical outlook, such freedom can neither be supported nor refuted by any appeal to independent empirical or metaphysical considerations external to that perspective.

118

Kant on absolute responsibility and transcendental freedom

Kant’s argument here is not really theoretical, but existential. He is not merely arguing that the Moral Law is a presupposition of a kind of freedom that we believe we already have, that the law follows from a truth that has held all along. Rather, the Fact of Reason involves not so much a discovery of the Moral Law, but a basic confrontation with its demands that transforms us into the very sort of agents to whom it applies. That is, in being accosted by this radical notion of responsibility, we are made to conceive of ourselves in a new way, and so to become something new (Faktum der Vernunft could equally be translated as the “doing” of “making” of reason). Confronted by this new notion of accountability, I  see myself as being able to succeed or fail in a new sense, and so recognize a new way I can count as doing something. Once I have so reconceived myself, there’s no going back. The Moral Law does not show us that we are free; it makes us free, whether we like it or not.10 Kant’s conclusions are less outrageous once we realize that freedom is only supposed to make sense within an autonomous ethical outlook, and so bears on only moral questions and contexts. Of course, my deeds still must in some way involve matter in motion in space and time within the causally determined order of nature. The problem with this deterministic picture is not that it makes our acts unfree. The real problem is that, interpreted in terms of natural events, our deeds can be neither free nor unfree. It’s not just that freedom cannot be located in the causal order of nature; no meaning can be given to such freedom as being either present or absent at all. Although it is not the case that the number 7 has any weight, neither is 7 weightless (the way a photon is). There is no number of grams that 7 weighs, not even zero. Attributing weight to 7 would simply be a category mistake. Since this assertion would be nonsense, so too would be the corresponding denial. Something similar may lie behind Kant’s worry about the determinism of nature. Given that Kant defines natural objects, relations, and explanations by reference to causal laws, it doesn’t make sense to even raise questions of freedom about anything understood as a natural event. Here we quite literally don’t know what we could be talking about, what we’re really trying to say. Our problem then is not finding a way to “break into” the laws of nature, as if to perform a minor miracle (or major magic trick). Instead, our problem is conceptual: of finding some other way of understanding our bodies and motions so that to say an act was or wasn’t free is actually to assert something, rather than just to confess a lack of any explanation or understanding. Such a reconceptualization can’t be a matter of just working up more elaborate scientific or natural stories, since the problem is with the very logic of such accounts, not their content. Of course, we can only act in the world, a world that must be available to be made sense of in terms of deterministic laws that leave no space for any kind of freedom (other than mere randomness). In order to make judgments of responsibility, there must be some definite way to reinterpret that world in a morally significant way, and so see our acts as being done intentionally and so in response to reasons. In turn, such an interpretation would need to bring in some notion of what can minimally count as an intelligible reason for action. Science cannot answer or even really address these questions, because they concern how and how far we can make sense to one another not as physical objects or living organisms, but as fellow participants in ethically meaningful relationships. The basic terms for such ethical self-understanding cannot be found in nature as it might be dispassionately known. Instead, these forms of interpersonal intelligibility could only come from human ethical life itself, as that life immediately presents itself not to Martian anthropologists, but to its native inhabitants, those who know this life from the inside.

119

David Sussman

There is indeed a sense in which it is true that, if anyone is to be morally responsible for anything, we all must be jointly responsible for the laws of nature in their entirety. But this claim does not entail that humanity has somehow set up those laws in a way that might be relevant to a scientist trying to explain the origin or causal structure of the universe. Instead, humanity may be responsible for the laws of nature only in the sense that it is relative to the engaged “internal” experience of our forms of ethical life that anything in the causal order can have moral significance at all. English-speakers do not choose the principles of acoustics that govern sound waves; nevertheless, they are jointly responsible for when and how those sounds count as speech, and for what that speech can mean. Such meanings are up to us collectively, and they may be constituted in a better or worse manner. In this way we can be morally responsible to ourselves for our ways of assigning ethical significance to the world. Kant is right that there must indeed be something more than nature insofar as nature can be scientifically known, but there need not be anything supernatural. All that really follows is that there must be more than one fundamental way for people to make sense of themselves.

7  Transcendental freedom regained What then of Kant’s insistence that moral responsibility presupposes transcendental freedom, the power to spontaneously initiate a new causal chain, to be a first cause? This certainly sounds like Kant is describing the sort of “spooky” freedom of the metaphysical libertarian, and he is almost always read as such. But again, Kant’s position can be rescued if “first cause” is read not materially but logically. In the Religion, Kant tells us that “Origin (the first origin) is the descent of an effect from its first cause, i.e., from that cause which is not in turn the effect of another cause of the same kind ” (R 6:39, my emphasis). Kant does not have to claim that moral responsibility requires the agent to originate their actions as a first efficient cause, rupturing the natural order with a Fiat! that tips over the first causal domino in a way that disrupts our scientific explanations. Instead, an agent could count as transcendentally free if she serves as the first cause of a particular logical type, of so as being the start of a new story that is intelligible in a distinctive way. On this more anodyne reading of transcendental freedom, I am the spontaneous origin of my actions not in the sense of producing them by creating or violating the laws of nature. Rather, I am the first cause of my actions insofar as they can be understood as parts a narrative that has ethical meaning. As such, my doings are not just physical movements but meaningful behavior that can have ethical significance for other moral agents and my relations to them. My actions are now to be made sense of not in terms of natural forces, but with respect to the possibility of blame, excuse, gratitude, challenge, affront, protest, advice, apology, forgiveness, etc. Such ethical interpretations reach their ultimate ground out when they get to me, as a unified first-personal perspective choosing light of reasons, holding myself accountable to other such agents in these terms. My acts would be “independent of the mechanism of nature” not in the sense of occurring in supernatural realm ontologically distinct from the phenomenal world (like the objects casting the shadows in Plato’s cave). Rather, my deeds would be independent of natural mechanism in that, having been reinterpreted in terms of interpersonal ethical significance, they would make sense in a way orthogonal to the order of concepts through which determinism can be either asserted or denied (just the meaning of the words on these pages differs deeply from, but does not contradict, an understanding of the chemical properties of the ink in which they’re printed). 120

Kant on absolute responsibility and transcendental freedom

Scientific explanations will still be available for our behavior, at least insofar as we always have the option of understanding our doings in terms of impersonal causal forces or sub-personal psychological processes. Yet this view doesn’t collapse into the compatibilism Kant scorns, because here we would no longer be talking about me or my actions anymore. Once we start understanding behavior in terms of physical or psychological forces, the ‘I’ as rational agent drops out of the story altogether, and so ethical concepts simply no longer have purchase. As a morally responsible agent, I need be transcendentally free only in the sense that I serve as the ultimate ground of a particular way of seeing my doings as being comprehensibly connected to one another, of being the first cause relative to a distinctive logic of intelligibility, a cause that is neither natural nor supernatural. True deterministic explanations can still be given, but these accounts need not compete with the ethical ones. Instead, such naturalistic approaches would just be ways of changing the subject (in more ways than one). Note: All references to Kant’s works are by the Prussian Academy of Sciences numbering in the Cambridge Edition of the Works of Immanuel Kant: CPR: Critique of Pure Reason CPrR: Critique of Practical Reason G: Groundwork of the Metaphysics of Morals MM: Metaphysics of Morals R: Religion Within the Boundaries of Mere Reason

Notes 1 See, e.g., Bennett (1984) and Williams, Ethics and the Limits of Philosophy (1986), ch.10. 2 See Scheid (1982) and Hill (2000). 3 See CPR A533/B561: “[Transcendental freedom] is the faculty of beginning a state from itself, the causality of which does not turn stand under a stand under another cause determining it in time in accordance with the law of nature”; CPR A446/B474: “[Transcendental freedom is] an absolute causal spontaneity beginning from itself a series of appearances that runs according to natural law.” 4 See Wood (1984). 5 Pereboom (2006). 6 Saunders (2022). 7 Allison (1990). 8 However, it is unclear whether moral responsibility requires only that we have some way to understand our agency that makes no reference to determinism, or if we must presuppose her to be causally undetermined, whether or not we can articulate that supposition the way we are understanding her. As Terrence Irwin (1984) observes, “if an event is determined, it true of it under all descriptions that it is determined, even though only some true descriptions . . . show why it is determined” (p. 38). 9 Vilhauer (2014). 10 For a similar view to which I am greatly indebted, see Korsgaard (1996).

Further reading Allison, Henry. (2012). “Kant on the Freedom of the Will,” in Essays on Kant (pp. 137–167). Oxford: Oxford University Press gives a helpful summary of Kant’s understanding of freedom in light of both its antecedents and its subsequent reception in German philosophy; Insole’s, Christopher J. (2013). Kant and the Creation of Freedom. Oxford: Oxford University Press reconstructs Kant’s

121

David Sussman understanding of freedom against its broader theological background; Korsgaard, Christine. (1996). “Morality as Freedom,” in Creating the Kingdom of Ends. Cambridge: Cambridge University Press shows how the Moral Law can both expresses and establishes freedom of the will; Wood’s, Allen W. (2020). Kant and Religion. Cambridge: Cambridge University Press offers a full reconstruction and defense of the most puzzling parts of Religion within the Boundaries of Mere Reason.

References Allison, Henry. (1990). Kant’s Theory of Freedom. Cambridge: Cambridge University Press. ———. (2012). “Kant on Freedom of the Will,” in Essays on Kant (pp. 137–167). Oxford: Oxford University Press. Bennet, Jonathan. (1984). “Kant’s Theory of Freedom,” in Allen W. Wood (ed.), Self and Nature in Kant’s Philosophy. Ithaca, NY: Cornell University Press. Hill, Thomas E., Jr. (2000). “Kant on Punishment: A Coherent Mix of Deterrence and Retribution?” in Respect, Pluralism and Justice: Kantian Perspectives (pp. 173–199). Oxford: Oxford University Press. Insole, Christopher. (2013). Kant and the Creation of Freedom: A  Theological Problem. Oxford: Oxford University Press. Irwin, Terence. (1984). “Morality and Personality: Kant and Green,” in Allen W. Wood (ed.), Self and Nature in Kant’s Philosophy (pp. 73–101). Ithaca, NY: Cornell University Press. Kant, Immanuel. (1996a). “Critique of Practical Reason,” trans. Mary J. Gregor, in Mary J. Gregor (ed.), Practical Philosophy (pp. 137–271). Cambridge: Cambridge University Press. ———. (1996b). “The Groundwork of the Metaphysics of Morals,” trans. Mary J. Gregor, in Mary J. Gregor (ed.), Practical Philosophy (pp. 37–108). Cambridge: Cambridge University Press. ———. (1996c). “The Metaphysics of Morals,” trans. Mary J. Gregor, in Mary J. Gregor (ed.), Practical Philosophy (pp. 353–603). Cambridge: Cambridge University Press. ———. (1996d). “Religion Within the Boundaries of Mere Reason,” trans. George di Giovanni, in Allen W. Wood (ed.), Religion and Rational Theology. Cambridge: Cambridge University Press. ———. (1998). Critique of Pure Reason, trans. Paul Guyer  & Allen W. Wood. Cambridge: Cambridge University Press. Korsgaard, Christine M. (1996). “Morality as Freedom,” in Creating the Kingdom of Ends (pp. 159– 187). Cambridge: Cambridge University Press. Pereboom, Derk. (2006). “Kant on Transcendental Freedom,” Philosophy and Phenomenological Research, 73(3), 537–567. Saunders, Joe. (2022). “Timeless Freedom in Kant: Transcendental Freedom and Things-in-Themselves,” History of Philosophy Quarterly, 39(3), 275–292. Scheid, Don E. (1982). “Kant’s Retributivism,” Ethics, 93(2), 262–282. Vilhauer, Benjamin. (2014). “Kant and the Possibility of Transcendental Freedom,” in The Palgrave Handbook of German Idealism (pp. 105–124). New York: Palgrave Macmill. Williams, Bernard. (1986). Ethics and the Limits of Philosophy. Cambridge MA: Harvard University Press. Wood, Allen W. (1984). “Kant’s Compatibilism,” in Allen W. Wood (ed.), Self and Nature in Kant’s Philosophy (pp. 73–101). Ithaca, NY: Cornell University Press. ———. (2020). Kant and Religion. Cambridge: Cambridge University Press.

122

Section 4: Individuals and society

9 RESPONSIBILITY IN CONFUCIAN THOUGHT David B. Wong

Introduction: responsibility as a bridge concept This chapter will use responsibility as a “bridge” concept to facilitate comparison of some major figures in the Confucian and Western moral and political traditions.1 There is no major concept used in Classical Chinese that is fully equivalent to the word “responsibility” in English. The relevant Chinese and English concepts overlap but also diverge in meaning. An important point of convergence lies in the root of responsibility being “to respond.” In many respects the ethics of these major Confucians (Kongzi or Confucius, Mengzi or Mencius, and Xunzi) is about the character of moral and political response, especially as related to reliability and trustworthiness. The idea of response comprehends important relationships between the responder and the one to whom there is response. The responder is entrusted and relied upon to respond, to answer to, the needs and to the vulnerability of the one to whom response is due. This is why Chinese concepts overlapping in meaning with responsibility include ze 責 duty and ren 任 responsibility (the latter especially related to holding a position or office). Another overlapping meaning of responsibility with Chinese concepts is that of “being held to answer for what one has done or is required to do.” This meaning is connected to being subject to blame or praise for what one has done, or for fulfilling or failing to fulfill what one is required to do. To be subject to blame or praise does not necessarily entail that one ultimately deserves either. Upon further investigation, one’s connection to the action in question may not reflect sufficient agency to merit blame or praise; or the degree to which one is subject to blame or praise may be mitigated or increased by discovering how the agency of others is involved. This second meaning of being held to answer for is related to the first meaning of responding to. One might be faulted or blamed, or credited and praised, for responding appropriately or not to the needs and vulnerabilities of others.

Responsibility as answering to the welfare of others Both answering to and answering for are present in the text named after the classical Confucian philosopher Mengzi (孟子; 4th c. B.C.E.), Master Meng, or as his name was latinized by Jesuit scholars, “Mencius.” In the eponymous text Mengzi, 4B292 presents the sage kings

125

DOI: 10.4324/9781003282242-15

David B. Wong

Yu 禹and Ji 稷 as exemplars of fulfilling duty and responsibility. Yu, legendary for working tirelessly to control the flooding of the great rivers of China, regarded those drowning as if he had drowned them himself. Ji, who ruled in a time of famine, regarded those starving as if he had starved them himself. It is not literally meant that Yu and Ji believed they themselves had done the drowning or starving (for a translation, see Bloom 2009). The “as if” wording rather conveys the supreme degree of responsiveness that Confucians required of rulers to the needs and suffering of their people. They had to do everything in their power to eliminate the harm and suffering. Xunzi (荀子, 4th–3rd. c. B.C.E), the last great Confucian thinker of the classical era, conveys this degree of responsibility eloquently when he says, “Tian [the supreme ordering force immanent within the world] gave birth to the people not for the sake of their lords, but established the lords for the sake of the people” (Xunzi: chapter 27: 66; for a translation, see Hutton 2014).3 Perhaps the implicit principle behind these passages in the Mengzi and Xunzi is that the greatest responsibility goes to those who have the greatest power to affect the welfare of others. A pre-eminent virtue of Confucian ethics is ren 仁, not the ren 任 of responsibility in office, but the ren often translated as benevolence, humaneness, or caring. I shall hereafter translate this ren as humaneness. The root of this kind of ren is openness and receptivity to others (see Wang 2012).

The responsibility relation The Chinese Confucian tradition and some dominant strands of the Western tradition have in common these concepts of duty to respond to the most central needs and vulnerabilities of others and of being held to answer for living up to that duty. Call this the responsibility relation. The two traditions diverge in the aspects of the responsibility relation to which the most attention is paid. The Western tradition quite often pays the most attention to the meaning of “being held to answer for” (which will be more familiar to Western theorists as “answerability”) as might be confirmed by the content of the papers for this very volume. This is confirmed by the vast literature concerning the appropriate conditions for blaming or praising an agent for what they have done. The literature extends from the familiar Aristotelian conditions such as knowing what one is doing to the vexed question of what constitutes having freely acted. Much of the discussion around the latter question has focused on the extent to which human agents are acted upon by factors external to themselves, putatively compromising their freedom and hence responsibility. Nothing in this focus on answering for logically prevents one from also paying attention to questions of responsiveness to the needs and vulnerabilities of others. One reason for the far greater attention paid to questions about blame and praise of the agent is the oft-held view that if one cannot be held responsible for being or failing to be responsive to others, then one has no duty to do so or at least excused from such duty. Another possible reason is a greater tendency to focus on the individual rather than the relation. Though the logic of the concept of responsibility is relational, one can choose to focus on the agent end of the relationship and in particular on the conditions for which the agent can be held to answer for their actions. The larger relationship recedes into the background.

The holistic focus on the larger responsibility relation and the continuity of the individual with the social environment In the classical Confucians, the larger relationship stays in the foreground of attention. One reason may be that their thinking does not frame as inherently problematic the idea that 126

Responsibility in Confucian thought

the individual human agent is deeply influenced by “external” factors, because there is held to be a continuity between individuals and their (particularly social) environment that is at least as important as whatever distinction can be drawn between them. Moreover, other people are parts of this social environment and also exercise varying degrees of control over that environment. Those who have more control often have, in effect, control over the choices of those who have less control. This understanding of the interdependence between people and their social environments, and their interdependence on each other that may come in asymmetrical degrees, is entirely in accord with the supreme degree of responsibility that Mengzi and Xunzi accord to rulers for the welfare of their people. People can be responsible for themselves and their own welfare, but in ways and in degrees that are compatible with other people, the ones with more control over the social environment, having a greater share of that responsibility. Though the classical Confucians did not much discuss metaphysics, a conception of causation emerged later in prominent strands of the tradition that is not linear and mechanical but closer to that of reciprocal, interactive influence. This kind of causation got connected to a holistic and organic world view that emphasizes the interdependence of all things. The causality operative in this interdependence is that of stimulus and response, and the effects of a stimulus are never purely passive or inert in the thing stimulated. They rather depend on the receptive and responsive capacities of the thing stimulated (Perkins 2019).

Mengzi’s view that responsibility is shared Mengzi is famous for his theory that human nature contains incipient dispositions (duan 端) to respond to situations in morally appropriate ways, but these dispositions are incomplete in the sense that they manifest themselves only in some but not all of the situations in which they would be appropriate. When a duan is fully extended to all the situations it should be, it grows into a moral virtue. Some conditions inhibit this growth, as material deprivation does. Other conditions nurture this growth. To illustrate this point, Mengzi uses the metaphor of barley being sown into the ground. Differences in the fertility of soil, in watering through rain and dew, and in the efforts of the farmer inhibit or encourage growth (6A7). One of the duan is the disposition to feel compassion (chuti ceyin zhi xin 怵惕惻隱之心), which, when fully developed, becomes the virtue of ren or humaneness. Another duan is the disposition to feel shame and dislike of what is contrary to rightness (Xiuwu zhi xin羞惡之心). Tian (天), or the ordering force within the world, implants these duan in human nature such that people are enabled to know the way for them to live. Corresponding to the metaphorical condition of the farmer putting sufficient effort into growing barley sprouts is the effort people make to extend the duan within once they realize they have them and that extending them is crucial for the realization of their full humanity. In 1A7, Mengzi attempts to prompt King Xuan (齊宣王) to realize that he possesses natural compassion by asking whether he had once spared an ox from ritual slaughter. The king affirms this fact, but in remembering the ox’s fear, he also once again feels compassion. Mengzi suggests that the king could extend the compassion he is feeling at that moment to his people. Mengzi is not just showing the king that compassion toward his people is justified or required, but also drawing an emotional analogy between the ox’s suffering and the people’s suffering, so that it feels emotionally right to the king to feel compassion for them. Mengzi then turns to what the king must do to help his people extend their own compassion: to set an example of treating one’s elders as elders and extend that (ji 及, or making 127

David B. Wong

it reach) to the elders of others, treating one’s young as young and extend it to the young of others. By modeling compassion and ensuring that his people receive an education in the virtues, the king can help them to extend their own natural compassion and thereby to become humane toward each other. Setting an example, however, is not enough. If the people lack a constant means of livelihood, Mengzi says, they will lack a constant mind that they can reliably set on the task of cultivating the beginnings of goodness in themselves (1A7). Lacking those constant means, they will be inclined to do whatever they can, including crime, to avoid suffering, and to feed and clothe themselves and their families. If the king punishes them for the crimes they commit, he in effect entraps them. He entraps them because their lack of constant means is much more under his control than theirs, and Mengzi goes on to propose concrete policies to the king to avoid entrapping the people: not overtaxing them to fund aggressive wars against neighboring states, prohibiting close-mesh fishing nets so as not to catch young fish and turtles, limiting the harvesting of trees to certain seasons, and planting mulberry trees so that the people will have silk clothing to wear (1A3, 1A7). Such policies make for sufficiently fertile soil from which his people’s duan might grow. It might be objected that Mengzi is too certain as to what the people would do when lacking a constant means of livelihood, and in particular, his seeming certainty that they would resort to crime. That is a fair criticism, but the point that remains is that the task of realizing their potential for goodness is much harder given their circumstances, and by enacting the wrong policies, the king has made it much harder for them. Here, Mengzi might be construed as invoking what has been called “circumstantial luck” by Thomas Nagel in his classic “Moral Luck” essay (1976). While we do have control over some of our circumstances, we have none or very little control over other circumstances that greatly affect what can we make of ourselves. This is the source of Nagel’s well-taken intuition that an arbitrariness of evaluation infects our judgments of what we tend to hold people to answer for. Some of us have very unfavorable circumstances in which we choose, yet others with more favorable circumstances get more credit for their choices. Interestingly, the only exception Mengzi makes to the rule that those without a constant means will be unable to have a constant mind is the class of shi 士, the gentry-scholar class of which Mengzi himself was a member. In an era of great political turbulence and disintegrating social order, members of this class were dislodged from their comfortable niches in the lower aristocracy, but their knowledge of traditions made them valuable to competing warlord–kings who wished to replicate the glory of the old kings. This distinct set of circumstances for the shi made genteel poverty and itinerancy more favorable, or at least less adverse, circumstances for maintaining a constant mind than others lower in the social order. Though most of us have no control over basic circumstances in which we make our choices, others do have greater control over them. This is the point of Mengzi holding the king much more responsible for the impoverished circumstances in which his people make their choices. Though this is not usually an observation that accompanies the discussion of circumstantial moral luck, it seems undeniable, and perhaps can be seen more clearly if one keeps the full set of responsibility relationships in view. People normally have responsibility to refrain from crime, which involves their responsibility to respond to the needs of their potential victims, but that responsibility can be mitigated by the lack of control they have to avoid severe suffering if they refrain from crime. Their responsibility is mitigated by the way desperation focuses the mind on self and those for whom one cares the most. When 128

Responsibility in Confucian thought

rulers or governments fail to realize the material conditions under which their people could devote adequate attention to the development of their virtue, they share a great deal of the responsibility for the wrongdoing and harm that results. That is, the responsibility of people, in the sense of what they can be held to answer for, is mitigated, and a significant portion of portion of the responsibility they might have had is subject to distribution to the ruler or government. That distribution is grounded in the responsibilities of the ruler or government to respond to the circumstances that profoundly affect the conditions under which people can make good choices.

Xunzi’s view on shared responsibility Whereas Mengzi holds that people share responsibility for growing one another’s moral sprouts, the last great Confucian of the classical era, Xunzi, starts from an ostensibly different view of human nature. Tian, while imparting order to the world such as the rotation of the seasons, does not implant dispositions in human nature enabling people to live the way they should. Rather, human nature contains dispositions toward seeking profit, to feel hate and dislike of others, and to seek sensual gratification. If they follow along with such dispositions, they will struggle, contend and become violent with each other. The sage kings of old, Xunzi recounts, established standards of rightness and rituals to foster good social order in which people would cooperate to produce and to distribute equitably goods sufficient to equitably satisfy the reasonable desires of all (Hutton 2014: 248; Xunzi 23.1–2). Rituals are those customary, socially established patterns of conduct that govern everyday interactions such as greeting and addressing others and taking meals with them, as well as major ceremonies marking the passages of human life, such as birth, attainment of young adulthood, marriage, death, and mourning. Xunzi held them to be especially important for moral cultivation of the self because they constitute enactments and physical embodiments of the sorts of attitudes and stances one should take toward others: respect and care in particular. Xunzi exemplifies a shared belief among the classical Confucians in the great influence of officials and rulers as models of virtue or the lack of it. Xunzi tells a story about the founding figure of the Confucian tradition, Kongzi (孔子, 551–479 BCE, better known in the West by his latinized name of “Confucius”), when he was Minister of Justice in his home state of Lu. A father and son had become embroiled in a legal dispute. Kongzi detained them both. Kongzi’s superior, Ji Sun (季孫), asked why he had released them when an unfilial son should be executed. Kongzi questions the permissibility of killing subordinates when their superiors are at fault: To fail to teach the people and then judge their crimes is to kill the innocent. . . . To fail to teach the people and yet hold them responsible for achieving meritorious works is being cruel. It is only when one has put a stop to these three practices that it is permissible to apply punishments. (translation by Hutton 2014: 319; Xunzi 28.3)

How modeling and teaching from others help develop one’s moral virtue Another common belief among Confucians is that moral agents always share some degree of responsibility with the others who influence the development of their agency. Rulers (and governments in general) share a great deal of responsibility when they shape the conditions 129

David B. Wong

under which people can devote proper attention to the development of their own virtue. As the previous story about Kongzi implies, parents can also share great responsibility because they similarly influence their children’s moral development by shaping conditions for their paying proper attention to developing their virtue and by providing them models of virtue (or the lack of it). Responsibility is shared with others who come to assume the role of one’s teachers and models of moral virtue. It is crucial to have models who inspire and show what good judgement looks like, partly through providing guidelines, and partly through revealing what features of the situation are morally salient in applying such guidelines. Further, it is crucial for people to have someone who knows them and who can point out their moral weaknesses, given that these features of the self often fall into one’s blind spots. Xunzi also was cognizant of the influence of peers. Quoting the saying, “If you do not know your son, observe his friends. If you do not know your lord, observe his companions,” Xunzi observes, “Everything depends on what you rub up against! Everything depends on what you rub up against!” (translation from Hutton 2014: 257; Xunzi 23.24). It is the responsibility of those who have advanced along the path of moral cultivation to show the less advanced how much more enduring satisfaction is achieved in following rightness and ritual propriety. Everyone begins in a state of ignorance and boorish preoccupation with gain and sensual satisfaction. “Without a teacher or the proper model, people’s hearts are nothing beyond their mouths and bellies” (Hutton 2014: 28; Xunzi 4.10). Xunzi then presents an analogy to a person having never seen meats and fine rice but only acquainted with vegetables and rice dregs. Such a person would think that the latter suffices. If presented with the former, they would not fail to choose them. If presented with the way of the ancient sage kings and the ordering influence of ren and yi, and how these make for “communal life, mutual support, mutual adornment, and mutual security,” no one would fail to choose them (Hutton 2014: 28; Xunzi 4.10). A teacher or model guides the student to the experiences that show them a better way to live. Consider that one of the main moral guidelines offered by Kongzi in the Analects would not suffice on its own because its proper implementation requires a teacher and a model. The guideline is called shu (恕), which is subject to several translations, but here I render it as “sympathetic understanding.” Kongzi’s student Zi Gong (子貢) asked whether there is a single word that can serve as a guide throughout one’s life. Kongzi replies, “It is sympathetic understanding: what you do not want, do not do to others” (Analects, Chinese Text Project Analects 15.24; for a translation of the Analects, see Ames  & Rosemont 1998). Analects 12.2 associates this “negative Golden Rule,” as it is often called, with ren or humaneness. In 6.30, Kongzi says that a humane person, wishing to establish themselves also seeks to establish others and in wishing to be prominent also helps others to be prominent (this is often called the “positive Golden Rule”). The passage goes on to say that to be able to draw from what is close as an analogy is called the method of ren (Analects 6.30). Another classical Confucian text, the Doctrine of the Mean (Zhongyong 中庸; for a translation see Eno 2016) says that zhong (what I would translate as “Doing one’s best in serving others”) and shu are not far from the Way (dao 道). It goes on to say that there are four things in the way of the exemplary person: to serve my father as I would expect my son to serve me, to serve my ruler as I would expect my minister to serve me, to serve elder brothers as I expect younger brothers to treat me, and to treat my friends as I expect my friends to serve me (Zhongyong, Chinese Text Project: 13). Here, what one would want or not want is envisioned by putting oneself in the place of others.

130

Responsibility in Confucian thought

A common criticism of the Golden Rule, whether in its Confucian embodiment or in Western traditions, is that it is peculiarly lacking as a guide to conduct.4 What one wants or would not want is not a reliable guide to how one should treat others. One’s wants may be undeserving of satisfaction, after all. And what even might be appropriate to want in one’s own situation may inappropriate for another to want in their situation. All this criticism is apt if shu is treated as a standalone moral rule, divorced from all values, virtues and other forms of moral guidance provided by Confucianism. But shu might instead play an organizing role in bringing other values, virtues and form of moral guidance into more coherent relationship with each other. To begin to see how it might play such a role, first consider how one should apply the analogy from what is close to others. The sense in which we are called upon to reflect on what we wish or want (yu 欲) for ourselves is the sense with which the Analects begins (1.1), where we are asked whether it is indeed not pleasant to learn and to practise repeatedly what one has learned, whether it is not delightful to have friends come to visit from afar, and to be untroubled when not recognized by others. Is this not the mark of an exemplary person? When Kongzi asks Yan Hui (顏回) and Zilu (子路) to speak of their aspirations, Zilu speaks of sharing his chariots, horses, clothing and furs with his friends, while Yan Hui wishes never to brag of his excellence or to make a display of his accomplishments, but Kongzi himself answers that he wishes to comfort the aged, to engender trust in his friends, and to nurture the young (Analects 5.26). Kongzi’s statement reveals the sense in which one is bidden to reflect on what is close to hand. It is to reflect on the aspirations one has to make the most of one’s life as a human being. In applying shu, we are bidden to reflect on what we most truly want for ourselves and to help others achieve it. The students’ answers are not so much wrong, but rather seem too specific and too small in scope compared to the largeness, the completeness but also the profound simplicity of Kongzi’s vision of a life of relationship with others. A teacher and a model can help one come closer to that vision. Often, we act misguidedly out of our most intense impulses or obsessions, not for the sake of a well-founded concern for our own well-being. Mengzi (6A10) says that to sell one’s life for a beautiful mansion or many concubines is to settle for far too low a price. Xunzi points out that valuing material goods to the extent that one is willing to violate the good order that makes cooperation possible will endanger one on the outside and will thereby make one fearful on the inside: “If one’s heart is worried and fearful, then even if one’s mouth is stuffed with grassfed and grain-fed meats, one will not know their flavor” (Xunzi 22.18; translation from Hutton 2014: 246). To extend to others from what is close to oneself, then, is to extend from what most matters to the self. Moreover, one must take care to recognize differences between people in life circumstances, in formative experiences, in social roles, and in temperament. One can wrongly extend or mis-analogize from self in countless ways. One might be tempted to dismiss what others want and don’t want as misguided because they differ from one’s specific aspirations and aversions. One must not be too quick to judge such differences as signs of others’ inferiority. One’s similarities with them help one to identify those aspects of life-fulfillment that one shares with others. One’s differences with them can help one to see how the shared aspects might be realized differently for the other. To make this kind of judgment, one must do one’s best to try to see the world and oneself as the other sees them. At one point, Kongzi comments with irritation on his student Zigong’s habit of judging

131

David B. Wong

others negatively, as if looking down upon them (Analects 14.29). Zigong is not good at sympathetically seeing how others see the world. Yet Zigong is highly adept in the performance of rituals. Kongzi emphasized their importance to moral cultivation of the self because they constitute enactments and physical embodiments of the sorts of attitudes and stances one should take toward others: respect and care in particular. Ritual performance is a kind of training of oneself to have the right sorts of attitudes and conduct oneself in the right way toward others. However, Zigong’s insufficient sensitivity to others shows that ritual performance might be technically proficient in smoothly following the prescribed patterns of conduct but can at the same time lack the feeling toward others that the conduct should convey. It can lack inflection by awareness of the way others’ situations legitimately differ from one’s own. As a result, his ritual performances may reveal a kind of unappealing glibness. Olberding (2012) provides an instructive analysis of the possible nature of Zigong’s most serious fault. Zigong might be too insecure about being genuinely good and tries to cover it up with his judgmental attitude toward others and his technically proficient ritual performances. As Olberding says, He thus tries to simulate a better version of himself, manipulating the technical skills he can command to create an image he hopes will better secure the approval and appreciation of others. Thus we see Zigong opining on the virtue of shu but doing so in a way that rather unsettlingly rings hollow and seems, as Confucius remarks, beyond his reach. (Olberding 2012: 170) Whether Zigong was ever able to overcome this fault is not known, but his best chance for doing so came from Kongzi’s insight into him. In general, we are fortunate to have teachers and models who can guide us with such insight.

Responsibility to self That teachers and models are crucial one’s moral development does not mean that one has lesser responsibility for one’s own development. If the material conditions sufficient for the focusing of one’s attention on moral virtue are present, and if others capable of pointing the way toward learning are available to be followed, then it is up to the individual (see Jiang 2002). Mengzi (6A9) makes this point vivid through his example of two students of Chess Qiu, the finest chess player in the state. One concentrates his mind and will, listening only to Qiu, while the other daydreams of shooting a swan. If the right teachers and models are available, it is up to the individual to choose to follow them. The development of agency is interactive with that of others but is consistent with having responsibility once the minimal conditions for exercising choice are fulfilled. A supreme master at teaching and modeling such as Kongzi can bring one to the threshold of knowing one’s faults, but one must ultimately cross that threshold and deal with those faults through one’s own effort. Xunzi is clear that there is no fundamental difference between the nature or intelligence of the sage kings of old and the typically boorish person who seeks sensual gratification and

132

Responsibility in Confucian thought

material goods. The latter must be shown what a truly gratifying life would be like. Having been shown this, it is up to the boorish person to put in the effort to accumulate goodness: They must seek it and only then will they obtain it. They must work at it and only then will they achieve it. They must accumulate it and only then will they be lofty. They must make it complete and only then are they sages. (Hutton 2014: 66; Xunzi 8.25)

Responsibility outruns control The extent to which the agent has control over conduct matters to the attribution of responsibility, yet in the end, the classical Confucian view does not make limit responsibility to what can be controlled by agents. There are two kinds of reasons for why responsibility outruns control: some responsibilities stem from the special relationships people have to each other; and the inherently relational nature of personal identity and flourishing. Consider the first kind of reason. The classical Confucians hold that one can have responsibilities to respond to the needs and well-being of those with whom one stands in special relationships such as family, friendship, and teacher–student. They are pluralistic in recognizing the different sources of responsibility. We have some responsibilities to people simply in virtue of their humanity. We have other responsibilities to certain others because we stand in special relationships to them. Parents have responsibilities to their children either because they gave birth to them or committed to their care at the time of adoption. Because of these relationships, parents have special duties to care for and nurture their children. And if they fulfill these duties to a reasonable degree, their children have special duties to respect them and to care for them when they become elderly. The nature of friendship and the teacher–student relationship (at least in the Chinese tradition) involve in somewhat different ways responsibilities that stem from initial commitments made, based on mutual regard and commitment to foster the well-being of the other. How the relationship evolves can strengthen or weaken the duties to respond to the needs and well-being of the other. Expectations that the other will be there for support, advice and guidance build up over time. Sometimes, one is required to try to prevent the moral failures of others with whom one stands in special relationship. Analects 16.1 reports on Kongzi’s criticism of his students Ranyou (冉有) and Zilu for failing to dissuade their employers, the warlord Ji clan (季氏), from planning to attack a neighboring, vassal state. He therefore is calling to account his students, who are failing to call to account their superiors for planning a wrongful action. It can be a necessary part of loyalty to others to try to prevent them from doing what is wrong. Another example of what loyalty can require comes from a particularly controversial stance that Kongzi took in Analects 13.18, where Kongzi says in response to praise for the son who turned in his father to the authorities for stealing a sheep that fathers and sons should rather cover up for each other. The Confucian tradition contains debate over Kongzi’s stance and how to interpret the reasoning behind it. A line of interpretation continuous with Kongzi’s chastising Zilu and Ranyou is based on the premise that filiality requires dedication to a parent’s moral well-being. The reason for the son not to turn in his father is

133

David B. Wong

to make the son’s remonstrance with the father more effective. The son’s duty is to persuade the father to get back on the right path, to return the sheep or to compensate the owner (see Huang 2013: 131–148 for a trenchant defense of this interpretation of 13.18). Now consider the other reason why responsibility can outrun the individual’s control: others can come to be so much a part of who one is that one may judge that one shares in the responsibility for what they have done. Recall Kongzi’s aspiration for his own life that he enjoy the trust of his friends. In the text Kongzi Jiayu (孔子家語 The School Sayings of Kongzi, in Chinese Text Project: 37.4), a story is told of the time Kongzi helps an old friend acquire an outer coffin for the burial of his mother. The friend engages in unseemly behavior that flaunts the prescribed mourning rituals when he climbs on top of the wood acquired to build the coffin and sings a song. Since he acquired the wood for his friend, Kongzi is implicated in the unseemly conduct, but pretends not to notice it and does not distance himself from it or his friend. In explaining this, Kongzi says that family members never lose that which makes them family and old friends never lose that which makes them old friends. Deep and lasting friendship can make us partly responsible for the misconduct of our friends (see Ing 2017 for an illuminating discussion of this passage). This is not to say that the responsibility has unlimited scope. Sometimes our friends may shock us, may turn out to be not the people we thought them to be. But at other times we commit ourselves to them as the flawed persons we know them to be; we accept that their flaws are in some sense our flaws, and our shared responsibility for what they do extends to a limit we cannot know in advance. It may seem that this view of responsibility as outrunning control is peculiar to classical Confucian ethics, but it is rather a view that is shared across culture. The chapter on reconciliation and responsibility in the African tradition by Thaddeus Metz in this volume articulates a similar theme. Further, the phrase “responsibility outruns control” is taken from the American philosopher Margaret Urban Walker, whose reference points are primarily from contemporary American ethical life (Walker 1993: 241). As Walker points out, to have child is a choice, but what is not chosen is for the child to be sickly or difficult. Nevertheless, one has responsibilities to this sickly or difficult child. Sage Kings Yu and Ji were what Walker calls “impure moral agents:” “saddled with weighty responsibilities and the open-ended possibility of acquiring more due to circumstances beyond their control” (Walker 1993: 25). Walker also gives a strong argument for this kind of impurity of agency: Yet agents who recognize their vulnerability to fortune are primed for dependability of humanly invaluable sorts. These are agents on whom we can depend, or at least to whom the presumption of dependability applies, and so whose undependability in many cases can be duly registered as failure. To the extent that these agents are people of integrity, they won’t fail us even under the blows of bad fortune or odd turns of fate which might otherwise prompt denial or opportunism; to the extent that we ourselves are such agents and possess integrity, we can depend, morally, on ourselves even in a bad spot. (Walker 1993: 25–26)

Conclusion This chapter has argued that classical Confucians can help reveal certain relational dimensions of responsibility that are obscured by disproportionate emphasis on what the

134

Responsibility in Confucian thought

individual is responsible for having done. These neglected dimensions involve the ways we are responsible to others to answer to their welfare, the responsibility we might have to them in virtue of the control we have over the circumstances in which they make their choices, the responsibility we have to ourselves to develop the moral potentials we have, the responsibilities we share with certain others with whom we have special relationships or who enter into our identities. Arguably, foregrounding the responsibility of the individual human beings for what they have done has led to a certain lopsidedness of moral attention to the detriment of human society and the entire globe. It may be time to explore an ethic that places the various relationships underlying responsibility in the foreground, especially given the urgent global challenges that arise from our interdependence on each other and on the entire planet.

Notes Aaron Stalnaker 2006 introduces bridge concepts in his comparative study of Xunzi and Augustine. 1 2 All references to the Chinese text Mengzi will be from the Chinese Text Project 2006–2022, where 4B29 appears as Li Lou II 離婁下. 3 All references to this text will be to Xunzi in the Chinese Text Project 2006–2022. 4 Kant criticizes the Golden Rule for vacuity in a footnote in his Groundwork for the Metaphysics of Morals Ak 4:430 (Kant 2002: 48).

Further reading Chan, Sin Yee. (1999). “Disputes on the One Thread of Chung-Shu,” Journal of Chinese Philosophy, 26(2), 165–186, provides an excellent survey of different interpretations of shu or the Confucian Golden Rule; Ivanhoe, Philip J. (2000). Confucian Moral Self Cultivation (2nd ed.). Indianapolis, IN: Hackett is a concise and accessible introduction to the evolution of moral self-cultivation in the Confucian tradition, from Kongzi, Mengzi, and Xunzi to later thinkers Zhu Xi, Wang Yangming, Yan Yuan, and Dai Zhen; Rosemont, Henry Jr. “State and Society in the Xunzi: A  Philosophical Commentary,” in T. C. Kline III & Philip J. Ivanhoe (es.), Virtue, Nature, and Agency in the Xunzi (pp. 1–38), presents a very useful overall map of Xunzi’s ideal state, including the economy, political institutions, social structure, rituals, and education; Tiwald, Justin. “Xunzi on Moral Expertise,” Dao: A Journal of Comparative Philosophy, 11(3), 275–293, provides an illuminating discussion of the question prompted by Xunzi, which is when one should follow the directives of someone with moral expertise when one does not know the correct course of action by oneself; Wong, David B. (2015). “Early Confucian Philosophy and the Development of Compassion,” Dao, 14(2), 157–194, discusses how Kongzi, Mengzi, and Xunzi differently conceived human nature to interact with the social environment in moral self-cultivation, and points out recent empirical work in the human sciences that supports various implications of their conceptions.

References Ames, Roger T., & Rosemont Jr., Henry (trans.). (1998). The Analects of Confucius: A Philosophical Translation. New York: Random House. Analects, in Chinese Text Project (2006–2022). Accessed October 7, 2022, from @ https://ctext.org/ analects. Bloom, Irene (trans.). (2009). Mencius, ed. and with an intro. Philip J. Ivanhoe. New York: Columbia University Press. Eno, Robert. (2016). The Great Learning and the Doctrine of the Mean: Translation, Commentary, and Notes. Online @http://hdl.handle.net/2022/23422. Huang, Yong. (2013). Confucius: A Guide for the Perplexed. London: Bloomsbury. Hutton, Eric L. (trans.). (2014). Xunzi: The Complete Text. Princeton, NJ: Princeton University Press.

135

David B. Wong Ing, Michael D. K. (2017). The Vulnerability of Integrity in Early Confucian Thought. New York: Oxford University Press. Jiang, Xinyan. (2002). “Mencius on Moral Responsibility,” in Xinyan Jiang & Robert Neville (eds.), The Examined Life: Chinese Perspectives: Essays on Chinese Ethical Traditions (pp. 141–159). Binghamton, NY: Global Publications. Kant, Immanuel. (2002). Groundwork for the Metaphysics of Morals, trans. and ed. Allen Wood. New Haven, CT: Yale University Press. Kongzi Jiayu, in Chinese Text Project (2006–2022). Accessed October 7, 2022, from @ https://ctext. org/kongzi-jiayu. Mengzi, in Chinese Text Project (2006–2022). Accessed October 7, 2022, from @https://ctext.org/ mengzi. Nagel, Thomas. (1976). “Moral Luck,” Proceedings of the Aristotelian Society, supplementary vol. 50, 137–155. Olberding, Amy. (2012). Moral Exemplars in the Analects: Goodness is Like That. New York: Routledge. Perkins, Franklin. (2019). “Metaphysics in Chinese Philosophy,” in Edward N. Zalta (ed.), The Stanford Encyclopedia of Philosophy. https://plato.stanford.edu/archives/sum2019/entries/ chinese-metaphysics/. Stalnaker, Aaron. (2006). Overcoming Our Evil: Human Nature and Spiritual Exercises in Xunzi and Augustine. Washington, DC: Georgetown University Press. Walker, Margaret Urban. (1993). “Moral Luck and the Virtues of Impure Agency,” in Daniel Statman (ed.), Moral Luck (pp. 235–250). Albany, NY: State University of New York Press. Wang, Huaiyu. (2012). “Ren and Gantong: Openness of Heart and the Root of Confucianism,” Philosophy East & West, 62(4), 463–504. Xunzi, in Chinese Text Project (2006–2022). Accessed December 4, 2022, from @ https://ctext.org/ xunzi. Zhongyong, in Chinese Text Project Liji (2006–2022). Accessed December 31, 2022, from @ https:// ctext.org/liji/zhong-yong.

136

10 AQUINAS ON HOLDING OTHERS TO BLAME Jeffrey Hause

Introduction Aquinas details his account of moral responsibility across multiple works. We find its fullest expression in the Second Part of the Summa theologiae: its foundations in the Treatise on Human Action (QQ 6–21) and its details in the Treatise on Sin (QQ 71–89).1 In brief, we are responsible for all and only those acts and omissions that are voluntary in the full sense. In contrast to non-human animals, whose acts are voluntary in a more limited sense, humans, who cognize not just particulars but also universals, are able to develop a conception of their ultimate end, which they can tweak, reformulate, or entirely revise. As a result, they can understand the ways in which various possible courses of action contribute to or fall short of their end and, in light of this understanding, adjust their ends. As a result, when they make their choices, those choices are not due to nature, but to a person’s deliberate will. Therefore, responsibility for those choices and the actions they form a part of rests entirely with the person who made them, not with nature (I–II 6.1, 6.2). These actions are voluntary in the fullest sense, free, and therefore subject to praise, blame, merit, and punishment (I 83.1, I–II 21.2c, DM 6). We are praiseworthy for morally good voluntary actions and blameworthy for bad ones.2 Aquinas’s theological aim of guiding human beings away from sinning and toward meriting the eternal happiness of the next life results in his rich and detailed accounts of praiseworthiness and blameworthiness. Assessing our worthiness of praise and blame helps us to guide our lives correctly. However, in these discussions he says very little about the acts of praising or blaming, such as what forms they should take or the conditions under which we may engage in them. When he does, he is often citing a prior authority, such as Aristotle. We might therefore conclude that Aquinas’s interest is solely on praiseworthiness and blameworthiness; after all, knowing whether our acts place us on or off the path of righteousness is crucial for determining how effectively we are striving for our ultimate end of happiness. But this assumption would be a mistake. Like many contemporary theorists who have developed their ideas in the wake of Strawson’s (1993) pioneering work, Aquinas contends that blaming people is a distinctive response to their wrongdoing over and above finding them blameworthy.3 Aquinas details, across multiple works, two importantly 137

DOI: 10.4324/9781003282242-16

Jeffrey Hause

different practices that meet the generally accepted criteria for blaming: they are responses to someone’s wrongdoing, serve as rebuke or protest of that wrongdoing, and exert a distinctive force, in particular a force on the wrongdoer’s psychology meant to underscore the offensive nature of the action. Because these are practices, Aquinas discusses them under the heading of the virtue governing each of them. The first is vengeful anger, governed by the virtue of vengeance, affiliated with justice. The second is fraternal correction, governed by the virtue of mercy, affiliated with charity. These virtues set the conditions under which each blaming practice is obligatory or at least appropriate. The goal of this chapter is to explain what each of these practices consists of, the circumstances under which it is fitting to execute them, and why each counts as blaming. After all, Aquinas speaks of vengeful anger as seeking redress and causing someone who has harmed me to suffer. The practice may thus look decidedly unlike blame and more like bloodthirstiness. Likewise, fraternal correction, which requires me to correct my neighbor out of love and concern for her happiness, looks more like a form of compassionate pedagogy, a nudge or exhortation to reform. On investigation, however, we will see that both practices  – fraternal correction always, and vengeful anger often – are ways of blaming. I will conclude with some observations about why Aquinas is not interested in a more general discussion of blame that would answer questions many contemporary accounts take pains to discuss, such as private blame or blaming the dead.

Vengeful anger Before we see how Aquinas conceives of vengeful anger as a blaming practice, we need to explore both his understanding of vengeance and that of anger. Vengeance is a response to a prior injury that one takes to be unjust (I–II 47.2c). Here Aquinas conceives of injustice very broadly, to include not simply failures to render what is owed in the strict sense but also to render what virtue requires (II–II 80 un. c). For instance, if you promise to help me with a logic problem, that promise creates a debt in the strict sense. If you lie to me about your ability to solve logic problems, you fail to treat me the way the virtue of truthfulness requires. In addition, this prior injury must be voluntary, whether directly, as a foreseen side effect, through negligence, or through negligent omission (I–II 75–76). In return for this sort of harm, I may consider responding with vengeance, which is an act “through which violence or injury, and absolutely anything that is disgraceful, is repelled by defending and avenging” (II–II 80.1c).4 Vengeance seeks to inflict on the offender something painful.5 However, to keep vengeance from deteriorating into mere revenge or even hatred, the victim must not seek this harm because she finds the offender’s suffering attractive for its own sake. The vengeance must instead be grounded in moral principles: the victim must judge whatever pain she inflicts to be a morally fitting response for the offense. What makes it fitting is, first and foremost, that it preserves justice by repairing and “removing” the offender’s harm; it should also, if possible, deter the offender or others from committing similar offenses in future (II-II 108.1, 158.1, DM 12.2c, Super 1 Thes. 5 l.2).6 Although vengeance is in some way punitive because of the pain inflicted, this punishment is distinct from the commutative justice dispensed by a legal authority, which has coercive power. Vengeance dispenses punishment for offenses against “the liberties (immunitas) of an individual person who drives back the injury” (II–II 108.2 ad 1).7 The victim might of course seek the aid of the courts to see justice done, in which case vengeance will be achieved in part through coercive punishment. Other cases of vengeance will 138

Aquinas on holding others to blame

be purely private: a vengeance that is non-coercive, without the power to levy fines or imprison offenders and that limits itself to the means open to private citizens – that is, social pressures – to “drive back” the injury. It is within these bounds that Aquinas makes room for vengeful blaming: The vengeance a private citizen may exact by her own power and within her own authority will be an expression of blame. The will’s desire for vengeance normally results in the passion of anger. Following Aristotle, Aquinas has a narrow conception of anger as a passion dependent on a rational moral judgment. There is something akin to anger in non-humans (and in humans who are, if only temporarily, deprived of the use of reason), such as the furiously hostile passion of a bee whose hive has been disturbed. This bee feels an irascible passion, one that fights against impediments to well being and seeks to overcome them. Genuine anger is a similar irascible passion. It incorporates the passion of sadness, a disturbance to one’s well being or tranquillity, and the passion of hope, a committed desire to repel and overcome those disturbances. In the case of anger, however, reason determines that the obstacle to one’s well-being is a harm unjustly inflicted on one, and the hope is to repel and overcome that injustice. Anger, therefore, is conceptually connected to a very precise rational judgment. If, for instance, I learn that the harm another person inflicts on me is accidental, or if I have no hope of achieving vengeance (say, because the offender is beyond the reach of my power), I may still be bothered, irritated, and upset, but these are manifestations of tristitia or “sadness” rather than anger. Aquinas’s anger is therefore a variant of Butler’s famous “settled and deliberate” resentment. Each philosopher takes injury or moral wrong as this passion’s occasion, and each warns us against “hasty” (Butler) or “antecedent” (Aquinas) anger, that is, anger that does not wait for reason’s full assessment that vengeance is warranted and in this degree and by these methods.8 Thomistic anger’s likeness to Butler’s deliberate resentment might seem to decay when we learn that, according to Aquinas, the vengeance I take always responds to a harm done to me. In discussing anger, Aquinas explains that “an injury that A inflicts concerns B only if A has done something that is against B in some way. From this it follows that a person’s motive for anger is always something done against that person” (I–II 47.1c). He echoes this idea later, in his treatment of vengeance, when he asserts that it is the person whose liberties are violated who takes vengeance (II–II 108.2 ad 1). While many contemporary theorists hold that this sort of personal anger (or the propensity for it) in response to injustice or a bad “quality of will” is a paradigm case of blaming, they also recognize similar reactions of anger toward those who have unjustly harmed others, and these reactions too constitute blaming.9 If Aquinas restricts vengeful anger to personal cases, then his account will not be able to explain the large number of cases in which we blame people for their treatment of others. However, Aquinas does not actually restrict vengeful anger to offenses against me personally. We also experience vengeful anger for those who “are connected with us in some way, whether through kinship, through friendship, or at the very least by sharing the same nature” (I-II 47.1 ad 2). As Aquinas points out, we think of our kin as extensions of ourselves, and we have habitual compassion for our friends and therefore in imagination link their welfare with ours, so it is easy to see how we consider their injuries ours.10 Likewise, Aquinas allows that we can practice vengeance against those who offend against God, against institutions we belong to, and against our fellow human beings (II–II 108.1 ad 4). He does not offer his reason for including God and our institutions, but presumably it is that we identify with each. This identification is clear in the case of our institutions, 139

Jeffrey Hause

and it becomes clear in the case of God once we conceive of God as the source of our existence and the being whose image we are created in. Finally, it is possible to identify with other humans through the affections of mercy (misericordia). Sometimes that happens through pity, when we identify with others who are harmed because we can remember or imagine the same happening to us; and sometimes that happens through compassion, when through imagination we grasp the harms done to another from the perspective of the harmed person.11 In sum, vengeful anger requires the judgment that someone has unjustly harmed me (or a person or institution I identify with) and is blameworthy for this harm. In return, one seeks to “repel” and “remove” the harm by visiting a harm on the perpetrator. Private citizens may not use coercive force in exacting vengeance, but they may apply social pressures and, in particular, the passion of anger, which is a natural expression of vengeance since it too aims at repelling harms to one’s well-being. From this brief summary, we can now see that vengeful anger is recognizably a blaming reaction. First, it is directed only at those who are blameworthy. Since all my blameworthy acts are subject to my rational judgment and lie in my control, this account satisfies the so-called “depth” requirement. One isn’t blameworthy for “superficial” acts and characteristics such as how well one sees; one is responsible only for what is “deep,” namely, actions and omissions that are subject to one’s deliberative judgment.12 The nature of the vindictive judgment, the influence of the angry passion, and the message of repudiation they send constitute the distinctive force blame must have. Finally, even though vengeful anger may result in a court proceeding, most frequently the force of blame remains merely social force and is non-coercive, and so the blame is distinct from judicial penalty. Aquinas’s generous conception of what constitutes a harm to me allows us to determine more precisely the range of harms that vengeful anger responds to. Of course, harm done to me personally, or to someone I identify with, is one such object, as when I am angry with a thief for stealing my bicycle, and restitution or replacement resolves this harm. However, some philosophers typically hold that such remedies do not fully make up for the offense and that offenders need to do “something more.”13 Otherwise, anyone who takes another’s property, or even injures her body, may make up for this fully by returning the stolen goods or doing chores free of charge for the victim. Aquinas agrees with the view that offenders must do more than restore what was lost. His reason is that, in addition to the personal harm an offender causes to her neighbor, the offender also “subtracts from” or creates a “defect in” the commonality (universus) or the state. Despite the mysterious ring of these expressions, Aquinas’s point is easy to appreciate. The offender has asserted an unwarranted dominance that the principles of justice, which everyone is obligated to subscribe to, forbid. This arrogation threatens to elicit reactions of disappointment, fear, and anxiety; it erodes trust in the community; and it potentially “scandalizes” others, that is, promotes like arrogance in others. Even if these dismal consequences do not actually follow, harming others is the sort of act that is apt to have these consequences and so is an affront to everyone in the community and not simply to the victim. Therefore, offenders harm not just the victim but also the community or state the victim is a part of. Since vengeful anger seeks to rectify harm to oneself, under Aquinas’s generous conception of “oneself,” vengeance will continue even after restitution is made. Otherwise, it would simply express a concern with disruption to my narrow self-interest and not a virtuous passion. Private citizens may not employ coercive force to punish offenders for harm to the community, but they may employ social pressures that guide the offender’s conscience to awareness of guilt and feelings of 140

Aquinas on holding others to blame

remorse. In these ways, vengeful anger responds both to the harm the victim personally sustains and the harm the wider community sustains. Aquinas frequently describes the nature of this response as “repelling” or “pushing back against” injury or harm done to oneself.14 In the case of offenses already committed, that repelling takes the form of inflicting suffering on the offender in response to the offense.15 However, the goal of vengeful anger is not simply that the offender suffer. In order to serve justice, the expression of anger must also send a message to the offender: I repudiate the harm done, I call for the offender to acknowledge these harms, and I declare that recompense or redress is required. In short, to borrow Angela Smith’s illuminating expression, the anger protests the offense.16 As Aquinas explains: The angry person aims at what is bad for another insofar as it has the character of vengeful justice. But vengeance is accomplished by inflicting punishment; and it belongs to the nature of punishment that it is contrary to the will, painful, and inflicted in return for some fault. Therefore, the angry person aims at this: that the one on whom he inflicts harm perceive it, and grieve over it, and recognize that he is facing it because of the injury he inflicted on him. (I-II 46.6 ad 2) For the protest to be effective, the person seeking vengeance must communicate to the offender and any bystanders that any applied social pressures (such as upbraiding, exclusion from social activities, loss of privileges) are a fitting response to the harm inflicted by the offender and not simply further unjust injuries. As we have seen, the passion of anger is ideally suited to convey this message. Because of anger’s conceptual connection with vengeance, the expression of anger always conveys both the judgment that the offender has inflicted an unjust harm and the vengeful person’s commitment to respond to it through appropriate punishment. Aquinas says very little about particular practices that vengeful blaming might take, but that is not unexpected. He says very little about the particular practices of judicial punishments as well, since his work is focused on moral principles rather than particular practices. A vengeful blamer may stay within the bounds of justice and not intrude on judicial authority if she blames in any of the following ways: by expression of anger alone; by rebuking the offender verbally; by excluding the offender from social activities; by no longer treating the offender warmly; by casting looks at the offender meant to induce shame. Aquinas will appeal to the virtues as a way of ensuring that whatever expression one chooses is both appropriate for the character of the offense and apt to communicate one’s protest to the offender. The primary goal of vengeful blaming is, therefore, to inflict non-coercive pain on an offender that is apt to register a protest against the harm done by the offender. Sometimes the offender will receive the message and seek atonement, but in other cases the offender will not accept the message or will accept it but not care to respond. Aquinas recognizes a secondary value of vengeful blaming, which he calls “medicinal” (Super sent. 2.42.1.2c, I-II 87.3 ad 2, 87.7c, 87.8c and ad 2). For instance, suppose students notice one of their classmates cheating on an exam and call this to the professor’s attention. If the professor fails to engage in vengeful blaming, the other students may take this as a sign that the professor is lax and they are free to cheat as well.17 Instead, by public vengeful blaming (say, removing the student’s exam and telling the student to leave the classroom), the professor publicly 141

Jeffrey Hause

protests the cheater’s action and communicates the values of honesty and trustworthiness, which promote justice in the classroom. So, even in cases in which there is little hope that the offender will reform, there may be good reason to respond with vengeful blaming.

Fraternal correction Aquinas offers vengeful anger as a blaming practice appropriate when one needs to communicate a protest against an offense to the wider community or when the offender is unlikely to respond to argument and persuasion.18 However, he advocates a different sort of blaming practice that is appropriate precisely when argument and persuasion are likely to succeed. Following a long tradition, he calls this “fraternal correction.” Unlike vengeful anger, which is governed by the virtue of justice, fraternal correction is governed by mercy, a virtue affiliated with the virtue of charity, in particular, charity’s aspect as love of neighbor. “Mercy” translates the Latin misericordia, or “wretched-heartedness.” This virtue directs us to help others who have lost or are in danger of losing their ultimate end of happiness, whether this-worldly or other-worldly. When we see our neighbor under just such a threat, our neighborly love inclines us to identify with her both cognitively and affectively. If we grasp the threat to her happiness by imagining ourselves facing a similar obstacle, then we join with our neighbor in a “real union”: We are or were vulnerable to the same impediments to happiness and can either remember or anticipate enduring them. If, by contrast, we know our neighbor well enough to imagine what it is like to be her and to see her troubles from her perspective, we enter into an “affective union” and feel compassion.19 Both forms of mercy are ways of identifying with our neighbor and feeling pain at the threat to her happiness. This identification allows us to see and care about our neighbor’s needs as if they were our own. In Aquinas’s ethical system, one major impediment to happiness is serious wrongdoing, or “mortal sin.” One who commits mortal sin rejects divine grace (and is thereby shut out from the happiness of the next life); in addition, sin tends to corrupt what moral virtues she may have or bolsters her vices, thereby endangering her opportunity for happiness in this life, since Aquinas maintains that happiness requires virtue. Therefore, when our neighbor sins, we should (when it is fitting) respond with mercy and try to help her to atone so as to regain a chance at happiness. He advocates a procedure outlined in the Gospel of Matthew: If your brother offends against you, go, and rebuke him between the two of you alone. If he hears you, you have gained your brother. 16. And if he will not hear you, take with you one or two more, that in the mouth of two or three witnesses every word may stand. 17. And if he will not hear them, tell the Church. And if he will not hear the Church, let him be to you as the heathen and the publican. (Matthew 18:15–17, Douay version) When your neighbor sins “against you” – that is, commits not a public sin but one that harms just you or that just you are aware of – you have reason to rebuke your neighbor. In fact, one may even have an obligation to engage in fraternal correction. After all, if we are well placed to do so, we are obligated to save those whose lives are in danger from injury or starvation, since otherwise they face wretchedness. However, those who commit serious sins likewise face wretchedness, and so if we are well placed to correct them, we are obligated to do so. This obligation stems from mercy and neighborly love, which must be 142

Aquinas on holding others to blame

our motive for correcting. If correcting counters the aims of love and mercy – namely, the happiness and flourishing of the sinner – then we must refrain. For instance, if correcting is likely to elicit anger, or to entrench the offender in the offense, or to be destructive of social bonds, then we may not proceed. Likewise, if correcting gives us feelings of superiority or inclines us to enjoy our own virtue posturing, then our self-aggrandizing aim is incompatible with genuine love of neighbor and we may not proceed. We must correct in ways designed to help the sinner to reform, and so they cannot be self-promoting, callous, insolent, or mocking.20 At first glance, fraternal correction might look very unlike blaming. It is, after all, a form of correcting, a sort of pedagogy, a forward-looking practice leading to reform and atonement, and thus very unlike vengeful anger, which looks backward to the offender’s harm and treats her accordingly. In addition, it appears to lack the characteristic force of blame. There are forms of correction that exert coercive power, but those are exercised by authorities, such as judges, princes, or bishops. By contrast, rebuke of one’s neighbor consists in a word of warning (Super Sent. 4.19.2.1). In fact, Aquinas tells correctors to treat sinners with mercy and, at least in the initial stages of correction, to avoid such social pressures as shaming. Correctors will therefore approach offenders having identified with them, out of love for them and concern for their happiness. Nevertheless, we can see that fraternal correction is a blaming practice from its other central features. First, the corrector rebukes the offender by conveying the seriousness of the act’s blameworthy character. This communication of the act’s character is not mere grading. What makes it rebuke is the corrector’s opprobrium toward the act in order to convince the offender that she is threatening her own happiness. Hence, the corrector’s message may be accompanied by whatever passions are needed to convey the gravity of the offense and to express moral solidarity with the offender.21 Second, while correctors cannot exert the sort of coercive force characteristic of governmental or ecclesiastical authorities, they can exert social pressures to hold the offender to account. In the first stage of correction, the social pressure is simply the corrector’s negative judgment and affective response. Should the early stages of correction fail but the correctors still have hope of helping the offender to reform, they may escalate the social pressures by making the offense public. The rebuke, the demand, the threat of shame and exclusion comprise a force to rival that of vengeful anger. In addition, the very fact that fraternal correction stems from love introduces yet another social force, since the corrector’s loving identification with the offender creates a greater expectation of reform.22 Finally, correctors hold offenders to account in that the correction includes a demand: the offender must face her offense, give an account to the corrector, and try to reform. Although fraternal correction is recognizably a blaming practice by the standards of contemporary philosophy, the differences between its aims and procedures and those of most contemporary accounts are striking. For instance, most contemporary work on blaming, echoing Aquinas’s account of vengeful anger, holds that blame serves to defend the value of the offended party and to voice support for moral norms within the community. Hence, obligations to expressing overt blame stem from the fact that a failure to do so would amount to condonation of the offense and thus would implicate the would-be blamer in the misdeed.23 In addition, because moral offenses rupture relationships with others, one appropriate response is to distance or alienate the offender from oneself or from the wider community. This distancing, which consists in a diminishment or withdrawal of concern and goodwill for the offender, might simply be an acknowledgement that one’s relationship 143

Jeffrey Hause

is not as close as one had thought, or it might be a retributive response that symbolizes the rupture caused by the offender.24 Since fraternal correction is an act of mercy, by contrast, it begins with the corrector’s recognition of the offender’s needs and through an increase of goodwill strives to help the offender to repent. A failure to engage in this form of blaming would itself constitute an offense against the offender. Nevertheless, it is important to remember that fraternal correction does not neglect the victim: to atone, the offender must do what is possible to repair damage to the victim, and the corrector insists on this repair. Hence, in fratenal correction the good of both offender and victim is affirmed.

Conclusion The differences between vengeful anger and fraternal correction are obvious, but it might be useful to point out their similarities so that we can see at a glance how they function as blaming practices. Both seek to “repel” the offense, the one by insistent rebuke together with loving expectation, the other by protesting the infliction of unjust harm through an expression of anger. Of course, a person who engages in vengeful anger must take up the standpoint of the victim; Aquinas also holds that vengeful anger should, if possible, serve as a corrective for the offender. Hence, he encourages concern for the offender as well in executing this practice. Likewise, fraternal correction requires one to identify with the offender, but one must also persuade the offender to see the nature and seriousness of the harm done to any victims. In each case, through very different sorts of social pressures, the blamer aims at humbling the offender’s will (to borrow Fingarette’s [1977] felicitous phrase) with the further goal of restitution, reformation, and the seeking of forgiveness. Because blaming responses are made fitting by the virtues that govern them, a virtuous person will not simply react to harmful wrongdoing without considering which sort of blaming practice – if any – to engage in. While this feature of Aquinas’s account might seem to over-intellectualize the blaming response, anyone who has responsibly parented a child has experienced the considerable control parents have over the variety of blaming responses to their children’s misbehavior: loving parental rebuke and correction, or sterner anger. Likewise, those who have developed the virtues of mercy and justice can exercise similar judgment and control over how to respond to their neighbors’ wrongdoing. Since Aquinas ranks mercy as the higher virtue – it is an expression of love for neighbor and in his view especially Godlike – it is no surprise that he finds we are obligated to engage in fraternal correction as our first response, unless we have a moral reason not to (DQFC 1, II-II 33.2). While the obligatory character of this blaming practice puts Aquinas at odds with most contemporary thinkers, he is typical of medieval philosophers in this regard. After all, as they reasoned, our neighbor’s very happiness is at stake. Reasons that exempt one from fraternal correction are dictated by the nature of the practice itself: We should refrain from correction if our motive is not love of neighbor, if we are unlikely to succeed, if someone else is more likely to succeed, if we are guilty of a similar offense (and thus risk hypocrisy), if we have learned of our neighbor’s offense by snooping, or if we will create a scandal by doing so (DQFC 1, II–II 33.2, 33.5, 33.6). If we cannot employ fraternal correction, we may be in a position to engage in vengeful anger but, again, only if in doing so we will attain the ends of justice. If we are unlikely to succeed in our protest, or if our real aim is to see the offender suffer (out of revenge or cruelty), or if we risk exceeding the reasonable bounds of anger, we must refrain. This last is presumably Aquinas’s reason for cautioning us against vengeance for harms done against us personally, as opposed to harms against 144

Aquinas on holding others to blame

God or neighbor. Aquinas is sensitive to the risk of anger from the writings of the Roman stoics. In such cases, Aquinas counsels patient endurance.25 We are now in a better position to explain why Aquinas’s discussion of blaming is constricted to overt practices whose goal is to communicate one’s moral concern or moral protest to the offender. Many other accounts make room for private blame, that is, blame one expresses or feels inwardly without outward expression; blaming those not present, whether because they are dead or their whereabouts are unknown; blame that consists in modifying one’s understanding of one’s relationship with a person so that, for instance, ceasing to wish that a person who has wronged us meets with success in her career counts as blaming. The reason Aquinas has no interest in widening his account of blame along any of these lines is that doing so will not serve the ends of the virtues regulating interpersonal relationships. Private blaming or blaming those absent can neither serve to correct their faults nor communicate the fact of their harm, so it serves neither mercy nor justice. And while Aquinas does think that details of our relationships with others do respond in obvious ways to the way they treat us – we may cease trusting those who have proven themselves untrustworthy – virtue requires us always to wish that things go well for people, that is, that they succeed in attaining what is genuinely good, whether they have wronged us or not. If we wish for people to meet with some misfortune, that must be because we believe it will help them to develop virtue as a result, not because they have wronged us.26

Notes 1 I use the following abbreviations for Aquinas’s works: DM (De malo/On Evil), In Ethic. (Commentary on Aristotle’s Ethics), Super 1 Thes. (Super 1 epistolam ad Thessalonicenses/Commentary on 1 Thessalonians), Super sent. (Scriptum super libros sententiarum/Commentary on the Sentences). Abbreviations for the Summa theologiae’s divisions: I = First Part, I–II = First Part of the Second Part, II–II = Second Part of the Second Part, III = Third Part. Abbreviations for the units of Aquinas’s texts: Q = Question, c = corpus or body of the article, ad = reply to objection, l = lecture. All translations are my own. 2 For an extended discussion of Aquinas’s account of responsibility, see McCluskey (2017). 3 See McKenna (2011) for discussions of both the relationship of blameworthiness to blame and the distinctive role of blaming. 4 Aquinas quotes from Cicero, De inventione ii, 161, here (as at II-II 108.2c) reading obscurum (“disgraceful,” “shamefully reprehensible”) rather than Cicero’s obfuturum (“injurious,” “prejudicial,” which he cites at Super sent. 3.33.3.4 qc.1c). Cicero is talking about the virtue rather than the act, but it makes no difference to the point at issue. 5 Pain consists in the opposition of the concupiscible power – part of the sensory appetite – to a present perceived to be bad. By analogy, rational pain consists in the will’s opposition to a bad one faces. Hence, what is painful ranges from bodily injury to embarrassment to social ostracism. 6 Aquinas gives no explicit account of how pain removes or negates the original harm, but his account is open to multiple possible interpretations (e.g., suffering pain counters the expression of disordered value in the orginal harm). 7 As will become clear in what follows, in the case of blaming, the metaphor of driving back (propulsatur) serves to illustrate a moral protest against the offense. 8 Joseph Butler (2017), Sermon VIII. 9 Strawson (1993), esp. pp. 49 and 56; Wallace (1994: 118–153); McKenna (2011: 56–68). 10 On kin, see II–II 30.1 ad 2. Aquinas bases his account of vengeance loosely on Aristotle’s. However, his acceptance of the idea – inherited not from Aristotle but from stoic and Judeo–Christian sources – that all humans are kin to each other results in a breakdown of the distinction between personal and impersonal offense. 11 See II–II 30.2. Cf. Butler (2017), who holds that properly regulated resentment is “a fellow-feeling, which each individual has in behalf of the whole species, as well as of himself” (Sermon VIII).

145

Jeffrey Hause 12 On the depth requirement, see Wallace (1994), in particular Chapter  3 (“Responsibility”), and Smith (2008: 384–387). 13 Swinburne (1989: 83–84); Bennett (2002: 159–161). 14 He describes both vengeance and the irascible passions in this way, for instance: deprimere (I-II 87.1c); repellere (I-II 48.2c, II-II 108.2c, III 15.9c; propulsare (Super sent. 3.33.3.4 qc1c, II-II 80.1c, 108.2c, 108.2 ad 1, In 4 Ethicorum l.13). 15 Vindication sought through the courts might require suffering a coercive “pressing back” (e.g., I–II 87.1c). Here I am concerned only with suffering non-coercive forms of “pressing back” that private citizens may engage in, such as having to face another’s anger. In keeping with the definition of vengeance, Aquinas also uses these same terms to describe self-defense by “deflecting” harm, as when Jesus counters (repellit) the accusation that he has a demon (Commentary on John 8 l.7). 16 See Smith (2012). 17 See Franklin (2012) for a kindred view of blaming as protecting norms. 18 Bernard Boxill has argued that protest is appropriate “when the time for argument and persuasion is past” (1976: 63). 19 Aquinas adumbrates the distinction between these two sorts of identification at II–II 30.2. 20 See his On Fraternal Correction 1. 21 Aquinas argues that for an act to attain the full level of moral goodness possible, it must include the expression of an appropriate passion (see, e.g., I–II 24.3). In the case of fraternal correction, passion can effectively communicate the corrector’s concern for the offender as well as opprobrium toward the offense. 22 I detail the nature of the demand in fraternal correction in multiple 13th-century thinkers in Hause (2018). 23 Hampton’s (1988) seminal work on blame and punishment inspired much subsequent discussion on this topic. 24 See, e.g., chapter 4 of Scanlon (2008) and Bennett’s response (2012: 76 and 80). 25 Aquinas characterizes patience (II-II 136.1, 136.4) not as failing to show self-respect or as condonation, but rather as a show of strength against troubles one cannot change and an affirmation of one’s own value in the face of those troubles. Aquinas speaks of forgoing vengeance at II–II 108.1 ad 2 and ad 4. 26 I would like to thank Maximilian Kiener, Chris Pliatska, and Angela Smith for their helpful comments.

Further reading McCluskey, C. (2017). Thomas Aquinas on Moral Wrongdoing. Cambridge: Cambridge University Press is an accessible and illuminating study of Aquinas’s general account of moral responsibility. Moule, G. (2016). Corporate Jurisdiction, Academic Heresy, and Fraternal Correction at the University of Paris, 1200–1400. Leiden: Brill, details two centuries’ discussion of the medieval practice of fraternal correction. A. Smith’s “Moral Blame and Moral Protest,” pp. 27–48 in Coates and Tognazzini’s Blame: Its Nature and Norms (New York: Oxford, 2012) offers a clear and wellargued contemporary discussion of the protest account of blaming.

References Bennett, C. (2002). “The Varieties of Retributive Experience,” The Philosophical Quarterly, 52(207), 145–163. ———. (2012). “The Expressive Function of Blame,” in D. J. Coates  & N. A. Tognazzini (eds.), Blame: Its Nature and Norms (pp. 66–83). New York: Oxford University Press. Boxill, B. R. (1976). “Self-Respect and Protest,” Philosophy & Public Affairs, 6(1), 58–69. Butler, J. (2017). Fifteen Sermons Preached at the Rolls Chapel, ed. D. McNaughton. Oxford: Oxford University Press. Cicero. (1968). Cicero II: De Inventione; De Optimo Genere; Oratorium Topica, trans. H. M. Hubbell), Loeb Classical Library. Cambridge, MA: Harvard University Press.

146

Aquinas on holding others to blame Fingarette, H. (1977). “Punishment and Suffering,” Proceedings and Addresses of the American Philosophical Association, 50(6), 499–525. Franklin, C. E. (2012). “Valuing Blame,” in D. J. Coates  & Tognazzini, N. A. (eds.), Blame: Its Nature and Norms (pp. 207–223). New York: Oxford University Press. Hampton, J. (1988). “Forgiveness, Resentment, and Hatred,” in J. G. Murphy & J. Hampton (eds.), Forgiveness and Mercy (pp. 35–87). Cambridge: Cambridge University Press. Hause, J. (2018). “Merciful Demand: Fraternal Correction as a Form of Blame,” Oxford Studies in Medieval Philosophy, 6, 144–167. McCluskey, C. (2017). Thomas Aquinas on Moral Wrongdoing. Cambridge: Cambridge University Press. McKenna, M. (2011). Conversation and Responsibility. Oxford: Oxford University Press. Scanlon, T. M. (2008). Moral Dimensions: Permissibility, Meaning, Blame. Cambridge: Harvard University Press. Smith, A. (2008). “Control, Responsibility, and Moral Assessment,” Philosophical Studies, 138(3), 367–392. ———. (2012). “Moral Blame and Moral Protest,” in J. G. Murphy & J. Hampton (eds.), Forgiveness and Mercy (pp. 27–48). Cambridge: Cambridge University Press. Strawson, P. F. (1993). “Freedom and Resentment,” in J. M. Fischer & M. Ravizza (eds.), Perspectives on Moral Responsibility (pp. 45–66). Ithaca: Cornell University Press. Swinburne, R. (1989). Responsibility and Atonement. Oxford: Oxford University Press. Thomas Aquinas. (1882). Opera omnia, iussu Leonis XIII edita cura et studio Fratrum Praedicatorum. Rome: ex typographia polyglotta S. C. de Propaganda Fide. ———. (1947). Summa Theologica [Summa theologiae], trans. Fathers of the English Dominican Province. 3 vols. New York: Benziger Brothers. ———. (1964). Commentary on Aristotle’s Nicomachean Ethics, trans. C. I. Litzinger, O.P. Chicago: Henry Regnery Co. ———. (2001). The De malo [On Evil] of Thomas Aquinas, ed. Brian Davies, trans. Richard Regan. Oxford: Oxford University Press. ———. (2010). “On Fraternal Correction,” in Jeffrey Hause & Claudia Eisen Murphy (trans.), Disputed Questions on Virtue. Indianapolis: Hackett. Wallace, R. J. (1994). Responsibility and the Moral Sentiments. Cambridge: Harvard University Press.

147

PART 2

The theory of responsibility

Section 5: The concepts of responsibility

11 RESPONSIBILITY AND AGENCY Maria Alvarez

0 Introduction There was somewhere (the New Yorker?) a cartoon of a couple at a restaurant, the man energetically admonishing the woman, who looked at him with a mixture of perplexity and dismay. The caption was something like: “Look, don’t judge me just by the way I look, the clothes I wear, the job I do, the books I read, the friends I keep, the way I vote, the things I say, the way I act – OK?!” We do judge each other by all those things and the joke is, of course, that the man’s list leaves little else of significance on which she might judge him. But not everything we judge each other by involves an imputation of responsibility for that thing, and that’s so whether the judgement is one of admiration or reprobation. To acknowledge someone’s good looks or talents, or to note their failures and shortcomings, is not yet to impute responsibility to them for those things, or their consequences. Imputing responsibility to someone for doings or not doings, for thoughts, feelings, sentiments and emotions, or for occurrences or states of affairs we ascribe to them, requires and implies more than just viewing those things as theirs, or somehow connected to them. And so, one may ask: What kind of connection does an imputation of responsibility require? Let me clarify the sense of “responsibility” at issue. Recently, philosophers have distinguished various kinds of responsibility: responsibility as “attributability”, as “answerability”, and (the most stringent) as “accountability.”1 The coherence and significance of these distinctions are disputed;2 but the sense I am concerned with is the last, “accountability”. However, I  will sometimes talk about “attributions” or “imputations” of responsibility without thereby meaning to engage in those debates. Angela Smith, who defends a single notion of responsibility, writes: I interpret the fundamental question of responsibility as a question about the conditions of moral attributability. . . . To say that a person is responsible for something, in this sense, is only to say that she is open to moral appraisal on account of it (where nothing is implied about what that appraisal, if any, should be). (Smith 2005: 238. See also Smith 2012: 578) 153

DOI: 10.4324/9781003282242-19

Maria Alvarez

I agree with Smith that imputing moral responsibility to someone for something makes them open to moral appraisal, and that this may be morally positive (“generous”, “kind”), negative (“mendacious”, “selfish”) or neutral (“efficient”, “punctual”). But, if there is a unified concept of moral responsibility, it surely involves, in addition to liability to moral appraisal and “demands for justification” (Smith 2012: 578), the notions of moral merit and fault and, in the case of responsibility for wrongdoing, prima facie liability to sanctions, such as blame and punishment, the requirement to apologise and to make amends, etc.3 A plausible answer to our previous question, one that underpins much of our talk and thought about responsibility, is that an imputation of responsibility to someone for something requires a certain kind of control over that thing – specifically, I shall suggest, the control that comes with our agential capacities. Roughly, the idea is that you are responsible for something you do or fail to do, for your thoughts, beliefs, desires, emotions, etc., and their consequences only if, and insofar as, you have certain agential capacities that confer control over those things. (Moral responsibility requires, in addition, the capacity for moral understanding). The range of things we are responsible for, and the grounds for responsibility for those things, are both matters of controversy. Here I shall focus on the question of responsibility concerning actions and omissions, and their consequences.

1  Agency as the basis for responsibility To say that responsibility depends on agency is not to say that we are responsible only for our actions (and their consequences). Rather, the suggestion is that we are responsible only for things that depend on our agential capacities because, and to the extent that, those capacities give us (some) control over those things; that is to say, whether those things happen or obtain depends on whether, and sometimes how, we exercise those capacities. That is because something may depend on someone’s agential capacities because of what she does or, by contrast, because of what she fails to do. Thus, we have control, and so can be responsible not only for our actions and their consequences but also for our omissions, or “refrainings”, and their consequences.4 Some question whether omissions, i.e., “notdoings”, can, strictly speaking, have consequences because, they say, this implies the possibility of causation by absence.5 But, while it may be true that not-doings, absences, etc., are not things or events that can cause anything, precisely because they are not things or events, it doesn’t follow that omissions (absences, etc.) don’t belong in legitimate (although sometimes normativity-dependent) causal explanations of occurrences or states. For example, a baby’s state of malnutrition may be causally explained by her parents’ failure to feed her properly – and that is because their role brings with it the duty to feed her. In that sense at least, omissions can have consequences. The idea is, then, that responsibility requires control and control stems from our agential capacities, which may be “implicated” both in our actions and in our omissions. But can I really control anything, given that my actions and omissions, and their consequences, depend on many factors other than me?6 A positive answer is supported by the consideration that, for something to depend on me, it is not necessary that it should depend wholly on me – and, therefore, total dependence on me is not necessary for something to be (at least partly) attributable to me, in the responsibility-relevant sense. What is necessary is that I should have the capacity to make a difference to whether something is the case; a 154

Responsibility and agency

difference that is sufficiently significant (because my positive or negative causal contribution is sufficiently salient) to warrant an attribution of control, and so of responsibility for that thing to me. There are different ways of conceptualising the agential capacities required for responsibility, which may or may not include a control factor.7 The approach I take here is one according to which an agent’s responsibility involves control, and that control depends on a distinctive type of abilities, namely “two-way” abilities. These are abilities whose exercise is up to the agent, as explained shortly.8 More specifically, on this view, for an action or omission and any consequences, to be imputable to an agent in the responsibility-relevant sense, the (non) occurrence or (non) obtaining of those things must depend on whether an agent exercises those abilities, and whether the agent exercises those abilities is up to the agent, in so far as she has a choice whether to exercise them.9 This does not mean that we are responsible for something only if its (non) occurrence depends on a choice to exercise an ability of this kind. Rather, we are responsible only if, and because, we are in a position to choose to exercise such an ability. Some philosophers reject the control condition altogether, on the ground that, they say, we can be morally responsible for things over which we have no control  – for instance, beliefs, desires, emotions, omissions, etc.10 I disagree: while those “attitudes” may sometimes warrant moral criticism or admiration regardless of one’s control, I believe that morally responsibility for them requires some control over them.

2  Agency, control, and alternatives The claim, then, is that the abilities that give us the control required for responsibility are “two-way” powers or abilities.11 The term may be unfamiliar but the kind of power it picks out is not, as should be clear when it is contrasted with two other kinds, “one-way” and “probabilistic” powers. Many of the powers of inanimate things, such as the fragility of glass or the capacity of some acids to corrode metals, are one-way powers. Others, such as radioactivity, are probabilistic powers. By contrast, many human abilities, such as our abilities to speak, to walk, grasp, push, pull and generally to move our bodies in goal-directed ways, as well as a range of acquired skills that depend on these more basic abilities, such as the ability to play musical instruments, or to drive a car, are two-way powers. The crucial difference between these three types of power is best brought out by reference to the relation between three factors, namely, (i) the (unimpaired) possession of the power; (ii) the conditions for its manifestation; and (iii) the manifestation of the power. Thus, when a thing possesses an (unimpaired) power and is in its manifestation conditions, (a) in the case of one-way powers, the power is manifested; (b) in the case of probabilistic powers, the power may or may not be manifested; and (c) in the case of two-way powers, the power may or may not be manifested, and moreover, whether the power is manifested, depends on the power’s possessor. The difference between (a) and (c) can be illustrated by considering the abilities to understand and to speak a language, say Spanish. A  crucial difference between them is that, typically, given certain circumstances, it is up to me whether I exercise my ability to speak but not my ability to understand Spanish. Given unimpaired possession of the ability (I’m not unconscious, drunk, or otherwise indisposed, etc.), when the conditions in which I can 155

Maria Alvarez

exercise my ability to understand Spanish obtain (I hear or read familiar Spanish words), I have no control over whether I exercise that ability: it is not up to me whether or not I understand the Spanish words I hear or read. If I hear with sufficient clarity a typical Spanish sentence, so long as my ability is unimpaired, I shall understand the sentence. However, even if the conditions for me to exercise my (unimpaired) ability to speak Spanish obtain, it is up to me whether I speak or not. Another way of putting the difference is this: if the conditions for the exercise of each of these abilities obtain, I can (typically) choose whether I speak but cannot choose whether I understand a Spanish sentence. Having a two-way ability, an ability over whose exercise one has control, involves having a choice over whether one exercises the ability when the opportunity to do so obtains (and in many cases, also over various aspects of how the ability is exercised). As I am using the term, a two-way power does not have “opposed” kinds of manifestation, say speaking and not speaking. Rather, speaking is the manifestation of the power, while not speaking is the absence of its manifestation.12 Second, and relatedly, these abilities are two-way powers because the (unimpaired) possession of the power, plus its manifestation conditions (opportunity), create the possibility of two opposed outcomes: the manifestation (speaking) and the non-manifestation (not speaking) of the power – though not both at the same time. Third, as already noted, what characterises these powers is that which of those outcomes occurs is up to the power’s possessor: when the conditions for their exercise obtain, the agent has control: she can choose to exercise the ability or refrain from exercising it. As it is sometimes put, given the right circumstances, these powers give us “alternatives”: to do or to refrain from doing something. In general, I can only refrain from doing what I have a two-way ability to do, or what I can do by exercising a two-way ability I have. And, on a particular occasion, I can only refrain from doing something if I  have the relevant unimpaired two-way ability and the opportunity to do that thing. By “the relevant two-way ability” I mean a two-way ability in or by exercising which, I would do the thing I can, in the circumstances, refrain from doing. For instance, I can refrain from leaving a room only if I can exercise a two-way ability I have by exercising which I can leave the room (e.g., my ability to walk), and I have the opportunity to leave: my exit is not barred. Thus, although refraining from A-ing is not an exercise of the two-way ability to A (or of some relevant two-way ability), it is something one can do only if one has the relevant two-way ability. In some circumstances, we may lack control over the exercise of these abilities. For example, I may speak while asleep or in a trance-like state. Or, contrariwise, I may become paralysed with extreme fear or fatigue and be unable to run or shout: I have not lost those abilities but my control over their exercise is lost or at least impaired. There is, of course, much more to say about two-way abilities, for example, whether all two-way abilities are agential (and vice-versa), what constitutes an impaired ability, where to draw the line between ability and opportunity,13 or, notably, whether the control these agential abilities confer requires certain epistemic conditions. But the upshot of the section is this. The abilities that give us responsibility-relevant control over our (not) doings and their consequences are these two-way abilities, because these are abilities that we can choose to exercise or not. When we are in circumstances propitious for their exercise, we can exercise them or refrain from doing so and, because of that, we can do or refrain from doing a range of things in those circumstances. We are responsible only for things that depend, in some way and to some extent, on whether and how we choose to exercise those 156

Responsibility and agency

abilities. Thus, we can be responsible for (not) doing something (and for its consequences), not only when we choose (not) to do that thing but also when (i) we do that thing by choosing to do something else; (ii) when we choose not to do something else by doing which we would have done that thing; and (iii) when we fail to choose to do something that was open to us to do because we had some relevant two-way ability, and the opportunity to do that thing.14 The thought that having a choice, having alternatives, when acting is necessary for moral responsibility is an old view. Something like it has been at the root of the debate about the compatibility of freedom and moral responsibility with a deterministic view of the world across the centuries, whether that view resulted from belief in divine providence or predestination; in a metaphysical principle according to which everything has “sufficient reason”; or, since modern philosophy, in a conception of “laws of nature” that “govern and explain the behavior of all objects, irrespective of their kind”.15 The abiding question has been whether determinism, in any of those forms, allows for the kind of freedom of will and action thought to underpin moral responsibility. The answer divided philosophers into incompatibilists (“no”), compatibilitists (“yes”); and “free-will skeptics” (“no, but neither does indeterminism”). I do not propose to take a position on that debate but would like instead to turn our attention to a significant change in intellectual climate that occurred around the last quarter of the 20th century, namely, from the assumption that alternative possibilities, however construed, are necessary for moral responsibility, to the view that they are irrelevant to that issue. Thus, in 1975 Peter Van Inwagen wrote: “almost all philosophers agree that a necessary condition for holding an agent responsible for an act is believing that the agent could have refrained from performing that act” (Van Inwagen 1975: 188). Regardless of whether such a consensus existed, it is significant that it was thought to exist and thought to have been soon afterwards shattered. The latter was largely due to the influence of Harry Frankfurt’s 1969 paper “Alternate Possibilities and Moral Responsibility”, where he claims to falsify the doctrine by counterexample. The first of what later became known as “Frankfurt-style” cases presents an agent who is, arguably, responsible for something he does, since he does it “for reasons of his own” but who, allegedly, “could not have acted otherwise” because an actually inactive but potential intervener would have made the agent act so anyway. The literature on this is too vast to summarise here.16 What matters for our purposes is that Frankfurt’s paper changed the intellectual climate, and the view mentioned in Van Inwagen’s passage is no longer the orthodoxy – it may even be the minority view. Incidentally, the passage is quoted in a 1984 paper by Daniel Dennett, provocatively entitled “I Could not have Done Otherwise – So What?” (Dennett 1984). So, the question is: Do alternatives matter for responsibility – and, if so, why?

3  The significance of alternatives – exculpation: exemptions, justifications and excuses Claims that someone couldn’t have done otherwise, or had no alternative, no choice, etc., are often advanced as a possible exculpation for (apparent) wrongdoing. (They may also be invoked in relation to morally good behavior but I will focus on exculpations). So, the significance of alternatives seems to be, precisely, that their absence can exculpate. Exculpations, however, work in various ways. Here I shall outline three ways of exculpating by reference to the presence or absence of alternatives, and argue that, while the first form of 157

Maria Alvarez

exculpation depends on whether the control condition for responsibility is met, the other two presuppose that it is. First, an exculpation can exempt someone from responsibility for something they did, or failed to do, or any consequences, by showing that the agent lacked the requisite control. Second, an exculpation can work by either justifying or excusing, partially or totally, because the lack of alternatives implies that, either a condition for blame worthiness is not (fully) met (excuse), or that there was no wrongdoing (justification).17 (These three forms of exculpation may be a matter of degree and, besides, there is controversy about the boundaries between them.) The fundamental difference between these forms of exculpation can be articulated also in terms of a distinction drawn by J.H.L. Hart and A.M. Honoré in between an agent’s having a “literal” and a “real” choice (Hart & Honoré 1985: 157). Consider the classic example of the choice between your wallet and your life with which a mugger presents their victim. An agent who, in these circumstances, reasonably hands over her wallet has a choice – a literal choice: she could refrain from handing over the wallet and risk her life. But, in another sense, she has no choice – no real choice because refusing to hand over the wallet is not an acceptable alternative, that is, an alternative that “meets an appropriate standard of eligibility, because it does not conflict with a binding obligation or because it is not too high a price to pay”.18 Before elaborating on each of these, I shall say something about the significance of this distinctions for recent debates about moral responsibility. Although the different forms of exculpation outlined above are familiar, the failure to pay proper attention to them has muddied debates about the relation between alternatives and moral responsibility – in particular, over the “Principle of Alternative Possibilities”, which says that an agent is responsible only if “she could have done otherwise”  – because these forms of exculpation are often implicitly conflated in these debates. Thus some philosophers, including Frankfurt, have claimed that the absence of alternatives is a condition for responsibility only when it plays a role in the agent’s motivation.19 So, they claim, a guard who chooses to, say, aid terrorists commit an attack motivated by a serious threat of substantial harm to his family may not be morally responsible for aiding the terrorists. In fact, I claim, the correct verdict in this sort of case is that the person is morally responsible (since he had a literal choice) but may be (partially) excused for blame, if the alternative is judged to be ineligible – in the example, because it is “a high a price to pay” for the guard (if it were judged to be too high, he may be justified). This is even more clear in cases of justification, as explained shortly. Thus, merit or culpability, and degrees therein, may depend on why people choose to act as they do, given the alternatives available to them. However, whether someone is morally responsible doesn’t depend on why they choose to act as they do – if they had a choice, they are responsible – but rather on whether they, literally, had a choice. (To be sure, if they had no choice, then they are likely to be free from blame, provided they are not responsible for their lack of choice.)

3.i Agential control and exemptions My suggestion has been that moral responsibility requires the sort of agential control that implies alternative possibilities in the literal choice sense and, where acting otherwise may be simply refraining, rather than doing something else. Thus, when “I couldn’t have done otherwise” works as an exemption, that is because the control condition for 158

Responsibility and agency

responsibility-relevant agency has not been met: what S did, or failed to do, was not a manifestation (in the broad sense of “manifestation”) of her agential capacities. This could be because what S did was not an exercise of a two-way ability they have, or something done by exercising such an ability, as when my arm moves because of a spasm. If I hurt you through that movement, I am exempted from responsibility since the control condition is not met. It may still be true that I ought to apologise to you, and feel bad about your pain, but not because I did anything wrong. Similarly, if I fail to meet you as promised because I am, through no fault of mine, locked in a room, I had no control over my failure to keep my promise: my failure to meet you is a not-doing but not a refraining since, although I had the ability to meet you (I had the ability to walk or drive to the appointed place), I had no opportunity to leave the room. Consider, by contrast, a case where I actually exercise a two-way ability I have and, say, slap someone. Here, the alternatives-control condition requires that I could have refrained. This condition is generally fulfilled because, since, ex hypothesi, I  exercise it, I  have the requisite two-way ability and the opportunity to refrain. In fact, opportunities to refrain from doing something, unlike opportunities to do it, come very cheap.20 To be sure, reasons and pressures of various kinds may on occasion be strong motivators not to refrain – but that is a different matter. The reader may think that “Frankfurt cases” show that this is not so. After all, these are presented as cases where an agent cannot do otherwise (refrain) because the presence of a potential intervener removes the opportunity of refraining. But the examples show that only if it is the case that the intervener could make the agent act (and not merely move her body), and make her act so that the agent had no literal choice. I have argued elsewhere that that the possibility is conceptually doubtful (Alvarez 2009). Things are more complex when the control condition concerns alternatives to refraining from doing something because that requires a positive ability and the opportunity to do what one refrains from doing – and abilities may be impaired, and the opportunity may be lacking. If so, it would seem that an alternative that the agent who refrains might in principle have had may be inaccessible to her. John Locke’s example of the man who willingly stays (i.e., does not leave) a room that unbeknownst to him is locked seems to the point (Locke 1689). This has led to the view that one may be responsible for refraining from doing something one could in fact not have done (i.e., leave the room) – that is, that one may responsible despite lacking literal alternatives.21 The challenge would require detailed discussion but, briefly, it is, at best, questionable whether the man in Locke’s example is responsible for refraining from leaving the room, as opposed to, say, for deciding to stay, or for not deciding (trying, etc.) to leave. If that is right, since he could have decided, tried, etc., to leave, it turns out he did have the requisite alternative: he could have done what he both fails to do and is responsible for. Compare the familiar example in the literature of someone who decides not to attempt to save a drowning child, or at least fails to decide to do so, in what, unbeknownst to the agent, are shark-infested waters. Arguably, in this case the agent is not responsible for not saving the child – since he could not have saved her, but he is responsible for not trying, or for not making a decision to save her, which he could have done. So these cases are not counterexamples to the alternatives–control condition for responsibility. A particularly thorny question is how certain psychological conditions affect control over the exercise of one’s abilities. For example, extreme fear, rage, or pain may paralyze someone, and some pathological conditions such as addictions, manias, compulsions, 159

Maria Alvarez

traumas, etc., may affect the control we have over the exercise of our volitional abilities (or, more widely our rational abilities). It is notoriously difficult, and to some extent an empirical matter, to establish when (and how) these conditions diminish that control, and when, if ever, they remove it altogether. The important point for our purposes is that, if these psychological conditions diminish or remove control over the exercise of our volitional abilities, they also diminish responsibility, or exempt the agent altogether of responsibility for the corresponding behavior and their consequences. I shall now move to the second way in which alternatives matter, which is not by exempting the agent from responsibility but by justifying or by excusing her for full blame for something for which she is responsible.

3.ii Eligible alternatives: justifications and excuses When the responsibility–control condition is met, and the agent had a literal choice, the question whether S could have acted other than she did may concern whether she had any real choice – where the absence of such choice, i.e., of an acceptable alternative, is a possible ground for justification or excuse. When “S could not have done otherwise” works as a justification, it does so by showing that the alternative to what the agent did, or failed to do, involved violating a more stringent obligation and, therefore, that what the agent did or failed to do which would otherwise have been a wrongdoing, or at least morally problematic, was in fact justified. Consider a couple of examples. I break my promise to have lunch with you because unexpectedly I have to accompany my mother to a hospital appointment. I am responsible for breaking my promise, but my doing so is justified and, I claim, involves no wrongdoing (or as some prefer to put it, a “justified wrongdoing”) because the alternative (not breaking it) would have violated a stronger obligation. A different kind of example: you decline to do me a favor because you have promised to take your child to the cinema. You are justified in declining to do what, let’s suppose, would have been an act of kindness (supererogatory) because the alternative would have involved breaking your promise to your child, which (or provided that it) was more important. When “S could not have done otherwise” works as an excuse, S is also responsible for the wrongdoing because S had control over the action or refraining that constituted or resulted in that wrongdoing.22 However, the absence of eligible alternatives, and the role that plays in S’s motivation, affect the degree of culpability for the wrongdoing. For example, a credible threat of harm may reduce someone’s blame for wrongdoing. Whether an alternative is eligible for someone depends on various factors, among others, their duties, commitments and values. For example, the risk of physical harm in a robbery may be “too high a price to pay” for a bank clerk who agrees to open a safe but not be so for a security guard. And the absence of an eligible alternative works as an excuse for wrongdoing provided that the agent’s choice is motivated by the lack of alternatives.

4 Conclusion The doctrine that moral responsibility  – at least for actions, omissions, and their consequences  – requires control can be defended, I  have suggested, by invoking abilities that confer control of one’s conduct because their exercise is up to us: in the right conditions, we have a choice over whether to exercise them. Some of the things we do or fail to do are 160

Responsibility and agency

not things we control in this way and, I have claimed, we are not typically responsible for them. That condition for moral responsibility should not be conflated, however, with the also correct idea that, when someone is responsible for their conduct, the lack of “eligible” alternatives can affect their culpability – either by excusing them or by justifying what they do. Why lack of alternatives can exculpate in these ways is a topic for another entry.

Notes 1 Gary Watson (1996) had added “attributability”, which is linked to virtue and vice judgments, to “answerability”. David Shoemaker (2011) distinguishes a third sense, which he call “attributability” (while using “accountability” for Watson’s “aretaic” sense). Mapping these alleged distinctions is rendered even harder by terminological variation. 2 Angela Smith, a main opponent of the need for these distinctions, has argued that there is a “single (and unified) concept” of responsibility (Smith 2015: 101). She initially (2005) termed this concept “attributability” but later switched to the term “answerability”, in order to mark her agreement with Shoemaker (2011) that her account aims to provide “conditions for forms of moral response that go well beyond mere aretaic appraisal” (Smith 2012: 576). 3 Smith holds that “moral criticism . . . can only be directed to a person with regard to things that involve her rational activity in some way” (Smith 2008: 382). Perhaps but, arguably, liability to moral criticism does not imply an imputation of moral responsibility. 4 “Omission” is normally a normative notion (Williams 1995): a failure to do something one ought to have done because of an obligation, convention, or some other normative ground. But I’ll sometimes use the term to mean simply a not-doing that is also a refraining. 5 For discussion see Armstrong (1999), Alvarez (2001), Beebee (2004) and Varzi (2007). 6 The question may resonate for those concerned about moral luck: the thought that factors wholly beyond one’s control should affect one’s moral responsibility and culpability/merit, whether absolutely or to some degree (See Nagel 1979; Williams 1976). A different issue is the sceptical challenge about the possibility of “ultimate” responsibility raised in contemporary debates by, for example, Galen Strawson (1994), based on the idea that we’re responsible for our choices only if we’re also responsible for the psychological features in us that lead us to make those choices. I have no space to discuss either of these important problems here. 7 For example, in terms of the capacity for intentional, voluntary or free action, for self-determination, for autonomous behaviour, for expressing one’s “true self” in one’s behaviour, for reason responsiveness, etc. 8 The use of the terms “powers”, “capacities”, “abilities”, “dispositions”, “tendencies”, etc., varies, and the distinctions that they are used to pick out is complex and contested (see Mumford & Anjum 2011). Here I use the first three interchangeably. 9 Some authors endorse a less demanding control condition for responsibility than suggested here. For example, John Fischer and Mark Ravizza (1998) developed a detailed account in terms of “reason-responsiveness” that they say requires “guidance” rather than “regulative” (i.e., alternatives) control for responsibility. 10 And some among them link responsibility to “rational evaluations”, e.g., T.M. Scanlon (1998), Smith (2005, et passim), or Matthew Talbert (2012). 11 The term “two-way powers” was introduced in contemporary literature, to my knowledge, by Peter Geach (1957) and taken up by Anthony Kenny (1975). See also Lowe (2013), Steward (2012), Alvarez (2013) and Pink (2016). For critical discussion see Clarke (2020). 12 The label “two-way” power is associated with Aristotelian “rational abilities” which do have opposed kinds of manifestation. For example, the ability to heal is manifested in healing and in causing of harm when that depends on the agent’s capacity to heal. See Frost (2020). 13 On this see Kenny (1975, 130ff.), Vihvelin (2013: ch.6), and Maier (2015). 14 See Alvarez (2013: 105–110). 15 Michael Frede emphasises that philosophers in antiquity did not possess the concept of a law of nature (Frede 2011: 15). The contemporary idea is perhaps that there is a body of laws that govern and explain the behaviour of the fundamental constituents (matter, force . . .) of everything in the Universe.

161

Maria Alvarez 6 For a helpful overview see Fischer 2010. 1 17 See Austin 1957 and Strawson 1962 for overviews of these distinctions. “Excuse” is sometimes used as a synonym for “exculpate”. Here I use it more narrowly. 18 Hyman (2015: 92, and Ch.4 for discussion). See also Olsaretti (2004). 19 “A person is not morally responsible for what he has done if he did it only because he could not have done otherwise” (Frankfurt 1969: 838). See also Chevarie-Cossette (2021) for the explicit claim that the Principle in fact concerns excuses. 20 The conditions for refraining are more demanding than those for an agent’s merely not doing something, in so far as refraining must be up to the agent and so requires that the agent could do the thing in question. That guarantees that the relevant alternative is “robust” and not a mere “flicker of freedom”, as John Fischer (1994) puts it. For further discussion of this point see Alvarez (2009). 21 Locke himself uses the example to make a claim about voluntariness, rather than about responsibility. 22 For different accounts of excuses see Gardner (1997) and Sliwa (2019).

Further reading Clarke, R. (2014). Omissions: Agency, Metaphysics, and Responsibility. New York: Oxford University Press. Systematic discussion of omissions covering relevant questions on metaphysics, philosophy of action, ethics, and law. Duff, R. A. (2009). Answering for Crime, Ch3. “Responsible for What?” examines the objects of responsibility and two familiar conditions: control and awareness, arguing that they are a condition for responsibility and for liability, respectively. Hart, H. L. A. (1961). “Negligence, Mens Rea, and Criminal Responsibility” and “Postscript. Responsibility and Retribution”, Reprinted in Punishment and Responsibility (2nd ed., pp. 136– 157). Oxford: Clarendon Press. First edition 1968. The first offers a defence of the legitimacy of holding people (criminally) responsible for harm caused through negligence; the second is Hart’s response to criticisms to his distinction between responsibility and retribution; it included a helpful discussion of different senses of responsibility (role, causal, legal-liability, moral blame and capacity-responsibility). Levy, Neil. (2014). Consciousness and Moral Responsibility. Oxford: Oxford University Press. Engaging with cognitive neuroscience and psychology, defends an account of consciousness as central to action and moral responsibility Robichaud, Philip, & Wieland, Jan Willem (eds.). (2017). Responsibility: The Epistemic Condition. Oxford: Oxford University Press. A collection of papers on the epistemic condition for responsibility, including when and why blameless ignorance excuses and the connection between this condition and issues about quality of will.

References Alvarez, M. (2001). “Letting Happen, Omissions and Causation,” Grazer Philosophische Studien, 61, 63–81. ———. (2009). “Actions, Thought-Experiments and the ‘Principle of Alternate Possibilities’,” Australasian Journal of Philosophy, 87, 61–81. ———. (2013). “Agency and Two-Way Powers,” Proceedings of the Aristotelian Society, 113, 101–121. Armstrong, D. (1999). “The Open Door: Counterfactual Versus Singularist Theories of Causation,” in H. Sankey (ed.), Causation and Laws of Nature (pp. 175–185). Dordrecht: Kluwer. Austin, J. L. (1957). “A Plea for Excuses,” The Presidential Address Proceedings of the Aristotelian Society, 57, 1–30. Beebee, H. (2004). “Causing and Nothingness,” in J. Collins, N. Hall, & L. A. Paul (eds.), Causation and Counterfactuals (pp. 291–308). Cambridge, MA: The MIT Press. Chevarie-Cossette, S-P. (2021). “Excuses and Alternatives,” Canadian Journal of Philosophy, 51, 1–16. Clarke, R. (2020). “It’s Up to You,” The Monist, 103(3), 328–341.

162

Responsibility and agency Dennett, D. (1984). “I  Could not have Done Otherwise  – So What?” The Journal of Philosophy, 81(10), 553–565. Fischer, J. M. (1994). The Metaphysics of Free Will: An Essay on Control, Oxford: Blackwell. ———. (2010). “The Frankfurt Cases: The Moral of the Stories,” The Philosophical Review, 119(3), 315–336. Fischer, J. M., & Ravizza, M. (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, 829–839. Frede, M. (2011). A Free Will. Origins of the Notion in Ancient Thought. Berkeley and Los Angeles, CA: University of California Press. Frost, K. (2020). “What Could a Two-Way Power Be?” Topoi, 39(5), 1141–1153. Gardner, J. (1997). “The Gist of Excuses,” Buffalo Criminal Law Review, 2, 575–598. Geach, P. (1957). Mental Acts. London: Routledge & Kegan Paul. Hart, H. L. A., & Honoré, A. M. (1985). Causation in the Law. Oxford: Oxford University Press. Hyman, J. (2015). Action, Knowledge and Will. Oxford: Oxford University Press. Kenny, A. (1975). Will, Freedom and Power. Oxford: Blackwell. Locke, J. (1689). An Essay Concerning Human Understanding, ed. P. H. Nidditch. Oxford: Oxford University Press, 1979. Lowe, E. J. (2013). “Substance Causation, Powers, and Human Agency,” in S. C. Gibb, E. J. Lowe, & R. D. Ingthorsson (eds.), Mental Causation and Ontology (pp. 153–172). Oxford: Oxford University Press. Maier, J. (2015). “The Agentive Modalities,” Philosophy and Phenomenological Research, 90, 113–134. Mumford, S. and Anjum, R. L. (2011). Getting Causes from Powers. Oxford: Oxford University Press. Nagel, T. (1979). Mortal Questions. New York: Cambridge University Press. Olsaretti, S. (2004). Liberty, Desert, and the Market: A Philosophical Study. Cambridge: Cambridge University Press. Pink, T. (2016). Self-Determination: The Ethics of Action, vol. 1. Oxford: Oxford University Press. Shoemaker, D. (2011). “Attributability, Answerability, and Accountability: Toward a Wider Theory of Moral Responsibility,” Ethics, 121, 602–632. Sliwa, P. (2019). “The Power of Excuses,” Philosophy & Public Affairs, 47(1), 37–71. Smith, A. (2005). “Responsibility for Attitudes: Activity and Passivity in Mental Life,” Ethics, 115, 236–271. ———. (2008). “Control, Responsibility, and Moral Assessment,” Philosophical Studies, 138, 367–392. ———. (2012). “Attributability, Answerability, and Accountability: In Defense of a Unified Account,” Ethics, 122, 575–589. ———. (2015). “Responsibility as Answerability,” Inquiry, 58, 99–126. Scanlon, T. M. (1998). What We Owe to Each Other. Harvard: Belknap Press of Harvard University Press. Steward, H. (2012). A Metaphysics for Freedom, Oxford: Oxford University Press. Strawson, G. (1994). “The Impossibility of Moral Responsibility.” Philosophical Studies, 75, 5–24. Strawson, P. (1962). “Freedom and Resentment,” Proceedings of the British Academy, 48, 187–211. Talbert, M. (2012). “Moral Competence, Moral Blame, and Protest,” The Journal of Ethics, 16, 89–109. Van Inwagen, P. (1975). “The Incompatibility of Free Will and Determinism,” Philosophical Studies, 27(3), 185–199. Varzi, A. C. (2007). “Omissions and Causal Explanations,” in F. Castellani  & J. Quitterer (eds.), Agency and Causation in the Human Sciences (pp. 155–156). Paderborn: Mentis Verlag. Vihvelin, K. (2013). Causes, Laws, and Free Will: Why Determinism Doesn’t Matter. Oxford: Oxford University Press. Watson, G. (1996). “Two Faces of Responsibility,” Philosophical Topics, 24, 227–248. Williams, B. A. O. (1976). “Moral Luck,” Proceedings of the Aristotelian Society Supplementary, 50, 115–135. ———. (1995). “Acts and Omissions, Doing and Not Doing,” in his Making Sense of Humanity (pp. 349–360). Cambridge: Cambridge University Press.

163

12 RESPONSIBILITY AND CAUSATION Alex Kaiserman

1 Introduction Consider the following case: Intersection: A and B both recklessly decide to run a red light. A crosses the intersection unscathed. B collides with V, killing her. A and B in Intersection are equally culpable. Indeed, it is in some important sense only a matter of luck that B struck a pedestrian and A did not. Nevertheless it is B, and not A, who is responsible for V’s death, because it was B’s action, and not A’s, which caused it. This seems to be an instance of a more general condition on responsibility for outcomes. However culpable one’s intentions, it seems inappropriate to ask one to answer for something one made no contribution to bringing about. We are responsible only for those traces we leave on the world, those events whose occurrence can be traced back to exercises of our agency (or failures thereof). As Aristotle puts it in the Eudemian Ethics, “Praise and blame are not given on account of what arises from necessity or chance or nature, but for what we are the causes of ourselves” (Aristotle, EE II.6.10). Recently, however, it has been argued that this intuitive picture cannot accommodate individual responsibility for the consequences of collective (in)action – things like climate change, structural injustice, war crimes, or industrial accidents. According to Kutz (2000: 164–165), for example, the claim that intentional participation in a collective venture is a basis for accountability for the harms and wrongs that result from this venture . . . conflicts with wellrooted convictions about the necessity of a link between individual accountability and individual causal contribution. Lawson (2013: 227) agrees that “individuals are blameworthy for collective harms insofar as they knowingly participate in those harms . . . regardless of whether they succeed in making a causal contribution to those harms”. And Young (2003: 11) insists that we should in DOI: 10.4324/9781003282242-20 164

Responsibility and causation

some sense “be held responsible for . . . structural injustice, as members of the collective that produces it, even though we cannot trace the outcome we regret to our own particular actions in a direct causal chain”. In this chapter, I will offer a partial defence of the idea that you are responsible only for what you cause. My focus will be on showing how a more nuanced understanding of the metaphysics of causation can help us better understand the nature and scope of responsibility for outcomes. After some ground-clearing in section 2, I turn in section 3 to consider various arguments to the effect that individual actions don’t cause collective harms because they make no difference to whether, or how, those harms occur. Then in section 4, I discuss to what extent the causal condition can be extended to cover responsibility for consequences of collective omission.

2 Preliminaries It’s helpful to begin by distinguishing two ways of being responsible for something – derivatively and directly. Derivative responsibility is responsibility we bear for things partly in virtue of being responsible for other things. For example, I’m responsible for the explosion partly in virtue of being responsible for planting the bomb. On pain of infinite regress, however, there must be some things – choices, perhaps, or basic actions – for which we’re responsible non-derivatively, i.e., not in virtue of being responsible for anything else. The concept of direct responsibility raises several interesting causal questions which I won’t have the space to discuss here. One foundational question, of course, is whether being directly responsible for one’s action/decision is compatible with its being causally determined by prior events. Incompatibilists say no. There are at least two challenges for this view. First, it’s plausible that we are directly responsible only for things over which we had control; but it’s not clear how to make sense of this idea, if not in causal terms. In response, some incompatibilists appeal to agent causation – agents control their actions by causing them, but this is consistent with the action not being caused by any prior event.1 Some metaphysicians claim to find the idea of irreducible substance causation hopelessly obscure;2 others find it perfectly coherent, but also so mundane that it couldn’t possibly be what distinguishes responsible from non-responsible conduct.3 The second challenge is that at least some of the things for which we’re directly responsible are things we do for reasons; but again, it’s not clear how to make sense of the idea of acting for a reason if not in causal terms. Some incompatibilists offer non-causal accounts of reason explanation.4 Others accept that reasons can cause free actions, as long as they do so without determining them;5 this, of course, requires an account of indeterministic causation.6 Compatibilists believe that being directly responsible for one’s action is compatible with its being caused by prior events. Many, taking inspiration from Frankfurt’s (1969) seminal work, argue further that whether one is directly responsible for one’s action is fully grounded in its causal history – acting freely, in other words, is a matter of one’s action being caused by the right sorts of things (perhaps also, in the right sort of way).7 Kaiserman (2021a) extends this idea to capture degrees of responsibility – one’s degree of (direct) responsibility is equal to the degree to which the right sorts of causes contributed to bringing one’s action about. This, of course, requires an account of degrees of causal contribution.8 We’ve seen that there are deep connections between direct responsibility and causation. But my focus in this chapter will be on derivative responsibility – specifically, the 165

Alex Kaiserman

conditions under which one is derivatively responsible for something in virtue of being directly responsible for something else. Here is a more careful statement of the view I want to defend: Causal Condition: S is derivatively responsible for o in virtue of being directly responsible for φ-ing only if S’s φ-ing was a cause of o. A few clarificatory remarks. First, it’s worth emphasising that Causal Condition only specifies a necessary condition. For one thing, there is plausibly an epistemic condition on derivative responsibility as well – if I flip a light switch which unbeknownst to me is wired up to a bomb, I am not responsible for the explosion, even though I was directly responsible for choosing to flip the switch and my choice caused the explosion.9 Second, it’s worth noting that o in Causal Condition can be any event, including, in particular, other actions and choices of S’s. For example, if a patient refuses to take their antipsychosis medication, then even if they aren’t directly responsible for the decisions made under the influence of their psychosis, they might still be derivatively responsible for them in virtue of being directly responsible for not taking the medication. Third, at least for now, I  will interpret φ in Causal Condition as restricted to action types, like pulling the trigger. It’s an interesting question whether Causal Condition is true when φ is allowed to refer to omission types as well, like failing to water the plants; since causation by omission raises its own distinctive challenges, I defer this question to section 4. Finally, it’s widely acknowledged that being responsible for some action or outcome can generate duties to do various things, or liabilities to be used in particular ways. Nevertheless, I don’t want to suggest that this is the only way that such duties or liabilities can arise. It’s compatible with Causal Condition that D has a duty to compensate V, or is liable to be used as a means to protect V or others in the future, even if they aren’t responsible for V’s harm. The connections between responsibility and responsibilities, in other words, are not straightforward.10

3  “But I made no difference!” Suppose Anna drives her gas-guzzler the short distance to work and back every day. Intuitively, Anna bears some responsibility for the consequences of climate change, like the melting of the polar ice caps, or the flooding in Pakistan in 2022. Yet had Anna cycled or walked to work instead, those events would most likely still have occurred, in more or less the same way they actually did. A similar point could be made about many cases of collective action – someone who knowingly buys the products of sweatshop labour, or votes for a fascist political party, or takes advantage of discriminatory zoning laws, intuitively bears some responsibility for the associated harms, even though, but for their actions, those harms would still have occurred. According to several writers on the topic, these intuitions are incompatible with Causal Condition. For example, Kutz criticizes the “individualistic conception of moral agency” (Kutz 2000: 4) according to which “the object of evaluation is solely what an individual has caused” (Kutz 2000: 5), which he takes to imply that “[i]f what I  do doesn’t make a difference  .  .  . to what we bring about together, then I  can’t be accountable for what we have done” (Kutz 2000: 6). Similarly, Sinnott-Armstrong claims that “my individual joyride does not cause global warming, climate change, or any of their resulting harms”

166

Responsibility and causation

(Sinnott-Armstrong 2005: 299) because “global warming will still occur even if I do not drive just for fun” (Sinnott-Armstrong 2005: 297). The presumption here appears to be that counterfactual dependence is necessary for causation – X is a cause of Y only if Y wouldn’t have occurred if X hadn’t occurred. This presumption is mistaken, however, as is widely recognised in the literature on causation on the basis of what are known as pre-emption and overdetermination cases. Here are paradigm examples of each: Backup Assassin: A  shoots V dead. If A  hadn’t shot, B would have shot and killed V instead. Two Assassins: A and B both shoot at V. Both bullets strike V at the same time. Either bullet would have been sufficient by itself to kill V. It is largely uncontroversial that V’s death was caused by A’s shot in Backup Assassin, and by both A’s shot and B’s shot in Two Assassins, even though V’s death would still have occurred had any one of the shots not occurred. Philosophers have of course developed more sophisticated accounts of causation to deal with such cases. Here is one example, drawing on ideas from Mackie (1965) and Wright (1985). Let’s say that X1, . . ., Xn were minimally jointly sufficient in the circumstances for Y if and only if they were jointly sufficient in the circumstances for Y, and no proper subplurality of X1, . . ., Xn were jointly sufficient in the circumstances for Y (for example, X1 and X2 are minimally jointly sufficient for Y if and only if they are jointly sufficient for Y and neither X1 nor X2 is individually sufficient for Y). Then: Minimal Sufficiency: X1, . . ., Xn collectively caused Y if and only if X1, . . ., Xn were minimally jointly sufficient in the circumstances for Y.11 X is a cause of Y if and only if X is one of a plurality which collectively caused it (just as someone is an author of a book if and only if they are one of a plurality of people who collectively authored it). Notice that ‘X caused Y’ and ‘X is a cause of Y’ are not synonymous, on this view. For example, consider the following case:12 Two Assassins (Hardy Victim): A and B both shoot at V. Both bullets strike V at the same time and V dies. But V was hardy – two bullets were required to kill him. In Two Assassins (Hardy Victim), neither A’s shot nor B’s shot caused V’s death, according to Minimal Sufficiency. But they were both causes of V’s death, in virtue of having collectively caused it, since they were minimally jointly sufficient in the circumstances for V’s death. Many cases of collective action can be straightforwardly handled by Minimal Sufficiency. For example, consider the following case: Election: A votes for B, an openly fascist candidate, in the presidential election. B wins and, as promised, enacts policies which harm V. A’s vote is neither necessary nor sufficient for the harm to V. But it was among a plurality of events – including other votes for B and decisions not to vote – which were minimally

167

Alex Kaiserman

jointly sufficient in the circumstances for B’s election, and hence for V’s harm. So A’s vote was a cause of V’s harm, according to Minimal Sufficiency (though it didn’t cause it). Insofar as we want to find A responsible for V’s harm, then, this is perfectly compatible with Causal Condition.13 Cases like Election are what Nefsky (2017) calls ‘threshold cases’. In threshold cases, “there is some precise number of acts of the relevant type needed to bring [the effect] about” (Nefsky 2017: 2746). But many cases of collective action are not like this. Consider the following familiar kind of case, for example:14 Torturers: A thousand people each have a switch in front of them. Each knows that V is hooked up to a device that will administer an electric shock of varying degrees of intensity, corresponding to how many of the thousand switches are flipped. If all one thousand switches are flipped, V will experience agonizing pain; if no switches are flipped, V won’t experience any pain at all. But the difference between the shock administered if n switches are flipped and the shock administered if n + 1 switches are flipped is indiscernible by V, for any n between 0 and 999. Every person flips their switch, and V experiences agonizing pain. Plausibly, there is no precise n such that flipping n switches is sufficient for V’s pain experience but flipping n−1 switches is not. According to some authors, this means that Minimal Sufficiency “will fail to deem any individual flip a cause of [V’s] pain” (Lee 2022: 84). To say that any one switch-flipping was necessary for the sufficiency of some plurality of switch-flippings for V’s pain “is like saying that adding a single grain of sand to a collection of grains of sand optimally arranged was necessary for the sufficiency of turning this collection into a heap” (Gunnemyr 2019: 409).15 I think this objection commits a familiar fallacy in the logic of vagueness. Suppose I have, say, seven grains of sand. Is there a way of arranging these grains so that they form a heap? Arguably, there is no determinate answer to that question – it’s just vague whether seven grains of sand, optimally arranged, form a heap. More generally: (1) There is no n such that, determinately, n grains of sand (optimally arranged) form a heap but n−1 grains of sand do not. The thing to recognise however is that (1) is compatible with (2):16 (2) Determinately, there is an n such that n grains of sand (optimally arranged) form a heap but n−1 grains of sand do not. Indeed (2) follows from the obvious fact that, determinately, one grain of sand does not form a heap and a trillion grains of sand, optimally arranged, do. In other words, it can be determinately true that there is a ‘cut-off’ point between heaps and non-heaps, even if it isn’t true of any n that it is the cut-off point between heaps and non-heaps. I think a similar thing is true of Torturers. The following claim may well be true: (3) There is no n such that, determinately, flipping n switches is sufficient for V’s pain but flipping n−1 switches is not.

168

Responsibility and causation

But (3) is compatible with (4): (4) Determinately, there is an n such that flipping n switches is sufficient for V’s pain but flipping n−1 switches is not. And indeed, (4) follows from the fact that, determinately, a single switch-flipping wasn’t sufficient for V’s pain but all one thousand jointly were. But (4) is all we need to reconcile Causal Condition with our judgements about responsibility in this case. For example, consider a particular switch-flipping, S – given (4), it follows that, determinately, S is one of a plurality of switch-flippings which were minimally jointly sufficient for V’s pain, and hence that, determinately, S is a cause of V’s pain. So-called ‘non-threshold cases’ thus pose no threat to Causal Condition. Here’s a slightly different version of the objection we’ve been considering. An event X can contribute to bringing about an event Y only by causing a part of Y. But in many cases of collective action, the relevant effect can’t be divided up into parts such that each individual action caused a different part of it. It’s not as if there is some part of the floods in Pakistan which was caused by Anna’s decision to drive her gas-guzzler to work, for example – climate change is much more complicated than that. Thus “we should not assume that one can find the morally relevant effect of one individual act by dividing the effect of the set of acts it belongs to into chunks that correspond to the effects of an individual act” (Kingston & Sinnott-Armstrong 2018: 173). My response here is simply to deny the premise – partially causing is not the same as causing a part. Consider again the analogy with authoring: someone can contribute to the authoring of a book without authoring a part of the book; perhaps they contributed some of the ideas, or did part of the research, or perhaps they and their co-authors simply co-wrote every single word. Similarly, one can contribute to the lifting of a table without lifting a part of the table, or contribute to the surrounding of a tree without surrounding a part of the tree (e.g., by joining hands with someone around the tree). The same is true of causation: X can contribute to a causing of Y without causing a part of Y. This is clearly captured by a view like Minimal Sufficiency. X contributes to a causing of Y, according to Minimal Sufficiency, if and only if X is one of a plurality of events which were minimally jointly sufficient for Y; whereas X causes a part of Y if and only if X was individually sufficient for that part of Y. But neither of these claims entails the other. For example, in Two Assassins (Hardy Victim) A’s shot contributed to a causing of V’s death, by being one of a plurality of events which collectively caused it. But there is no part of V’s death which A’s shot caused, since there is no part of V’s death for which A’s shot was individually sufficient (assuming it even makes sense to talk of ‘parts’ of deaths in the first place). In summary, an individual action can contribute to the causing of a harm, like the floods in Pakistan, without causing any part of those harms. Kaiserman (2016, 2017) argues that although causation is all-or-nothing, different events can contribute to a causing of some effect to different extents, by contributing more or less to making the plurality as a whole jointly sufficient for the effect. If that’s right, it seems natural to connect this idea to degrees of responsibility: S is fully responsible for some outcome o if S’s φ-ing caused o, S is partially responsible for o if S’s φ-ing contributed to a causing of o, and S is not at all responsible for o (at least not in virtue of having φ-ed) otherwise. Either way, S’s degree of responsibility for o is equal to the (maximum)17 degree

169

Alex Kaiserman

of contribution S’s φ-ing made to a causing of o. Depending on the details of one’s preferred account of degrees of causal contribution, it may well turn out that most people involved in big collective action problems, like climate change, structural injustice, and so on, bear only a tiny degree of responsibility for the relevant harms. Nevertheless, what they bear that degree of responsibility for is the whole harm, according to Causal Condition, not merely a part of it.

4  Omissions and overdetermination Here’s a reminder of the condition I have been arguing for: Causal Condition: S is derivatively responsible for o in virtue of being directly responsible for φ-ing only if S’s φ-ing was a cause of o. So far I have been interpreting φ in Causal Condition as restricted to action types. But on the face of it, one can also be derivatively responsible for an outcome in virtue of being directly responsible for failing to do something.18 Consider the following case, for example: One Button: There is a fault at a nuclear power plant. An explosion will happen unless a button in my office is pressed. It’s my last day, and I want to go out with a bang. So I deliberately don’t press the button, and the explosion happens. It seems that I could be responsible for the explosion in One Button, even if there is no action I performed which caused it. The question arises, then, whether Causal Condition remains true when we allow φ to refer to omission types, like failing to press the button, as well. Some philosophers deny that there is any such thing as causation by omission. Moore (2009: ch.18), for example, argues that omissions cannot enter into causal relations since (to quote The Sound of Music) ‘nothing comes from nothing, and nothing ever could’. He grants that one can be derivatively responsible for an outcome in virtue of being directly responsible for failing to φ, but insists that the grounds for this are counterfactual, not causal – I am responsible for the explosion in One Button, for example, not because my failure to press the button caused the explosion, but because had I pressed the button, the explosion wouldn’t have occurred. There are several ways of making sense of claims like ‘My failure to press the button caused the explosion’ which don’t require the existence of omissions considered as sui generis entities, however. According to some people, ‘My failure to press the button’ picks out an ordinary event  – presumably, whatever I  was doing instead of pressing the button – just under a negative description;19 others deny that the truth of sentences like ‘My failure to press the button caused the explosion’ need to be grounded in the holding of a relation between any two things in first place.20 I don’t propose to take a stand on this debate here. Hence I will simply assume, for the sake of argument, that sentences like ‘My failure to press the button caused the explosion’ can be true, which is all we need to reconcile Causal Condition with the intuition that we can be derivatively responsible for things in virtue of being directly responsible for failing to act a certain way.

170

Responsibility and causation

Interestingly, however, Sartorio (2004) argues that there are counterexamples to Causal Condition involving omissions even if we grant that causation by omission is generally possible. Her argument focuses on the following two cases: Two Buttons: As in One Button, except that the explosion will happen unless two buttons are pushed, one in my office and the other in yours. Neither of us press our buttons, and the explosion happens. Two Buttons – One Stuck: As in Two Buttons, except that one button is in my office and the other is hooked up to a safety mechanism designed to push the button if there’s a fault. The mechanism has become stuck, however, and fails to push the button. I also fail to press my button. The explosion happens. Sartorio’s argument then proceeds as follows: 1) I am responsible for the explosion in Two Buttons. 2) My failure to press my button didn’t cause the explosion in Two Buttons – One Stuck. 3) If my failure to press my button didn’t cause the explosion in Two Buttons  – One Stuck, then it didn’t cause it in Two Buttons. 4) Therefore, my failure to press my button didn’t cause the explosion in Two Buttons (from 2 and 3). 5) No other action or omission of mine caused the explosion in Two Buttons. 6) Therefore, Causal Condition is false (from 1, 4, and 5). It’s worth noting that many real-life cases have a similar structure to Two Buttons. Suppose I fail to repair the brakes on your car, but you fail to push the brake pedal before colliding with V;21 or suppose I fail to properly warn of the risks of a drug on its label, but you fail to even check the label before giving the drug to V;22 or suppose you perform an operation on V which shouldn’t be performed on patients with a certain condition, the details of which I failed to include in V’s medical records which you failed to request.23 If Sartorio is right, all such cases are cases of derivative responsibility without causation. The crucial premise for our purposes is 2). Sartorio begins by noting that the explosion would still have happened in Two Buttons – One Stuck whether or not I had pressed my button; my failure to press the button, therefore, made no difference to whether the explosion occurred. Now, it’s true that the explosion wouldn’t have occurred if I had pressed my button and the mechanism hadn’t been stuck. But if this is what makes my failure to press my button a cause of the explosion, “we would be committed to much more causation by omission than we are prepared to accept” (Sartorio 2004: 321). Consider a random omission, like Ronaldo’s failure to score last week. Now consider the following ‘obstacle’ to the explosion being prevented – Messi’s failure to form the conditional intention to travel to the nuclear plant and press the other button if Ronaldo scores. Though the explosion would still have occurred had Ronaldo scored, it wouldn’t have occurred if Ronaldo had scored and Messi had formed the relevant conditional intention. The reasoning above would therefore seem to imply that Ronaldo’s failure to score, and by extension every omission, is a cause of the explosion. However, there are other grounds for thinking that my omission is a cause of the explosion in Two Buttons  – One Stuck which don’t generalize in this way. One involves

171

Alex Kaiserman

appealing, once again, to Minimal Sufficiency. My failure to press the button was sufficient, in the circumstances, for the explosion. Ronaldo’s failure to score, by contrast, was not. Nor, moreover, does it seem to be a non-redundant part of any plurality of events/omissions jointly sufficient in the circumstances for the explosion. Thus Minimal Sufficiency has the resources to explain why my failure to press the button was a cause of the explosion in Two Buttons – One Stuck and Ronaldo’s failure to score was not. We can find further support for this strategy by looking at Sartorio’s argument for premise 1). Suppose I am not responsible for the explosion in Two Buttons. Presumably, though, I would have been responsible for the explosion had you pressed your button, because then the explosion wouldn’t have happened if I had pressed mine. But this, Sartorio says, seems to generate a problematic kind of moral luck. To say that whether I am responsible for the explosion depends on whether you press your button amounts to saying that two wrongs make a right – or more carefully, that “two wrongs that are generally conducive to a certain type of harm can neutralize each other” (Sartorio 2004: 319). It’s not clear to me that there is anything particularly problematic about this kind of moral luck, however. Consider the following case: Poison: One dose of substance X is deadly. But two doses neutralise each other, resulting in a harmless substance. You and I each independently inject V with one dose of X, intending to kill him. Shortly after V dies of something unrelated. Intuitively, neither you nor I are responsible for V’s death in Poison. Yet it seems to exhibit exactly the same kind of moral luck described previously – two wrongs that are generally conducive to a certain type of harm neutralize each other, so that neither you nor I end up responsible for V’s death even though each of us would have been had we acted alone. In later work, Sartorio (2012) acknowledges this point, but insists that there is a disanalogy between Poison and Two Buttons. In Two Buttons, each failure to press the button was individually sufficient for the explosion. In Poison, by contrast, neither action was individually sufficient for V’s death, even though if you hadn’t administered your dose my action and your failure to administer your dose would have been jointly sufficient in the circumstances for V’s death. Thus, Sartorio concludes, the principle that ‘two wrongs can’t make a right’ should be understood as restricted to cases like Two Buttons. It’s not clear what the rationale for this restriction is supposed to be, however. Moreover, the difference between Two Buttons and Poison can easily be explained by the combination of Causal Condition and Minimal Sufficiency. According to Minimal Sufficiency, my omission caused the explosion in Two Buttons, since it was individually sufficient in the circumstances for it. But my action wasn’t even a cause of V’s death in Poison, since neither my action nor yours was a non-redundant part of any plurality of events that were jointly sufficient for V’s death – that plurality would presumably have to include whatever actually killed V, which was already sufficient by itself for V’s death. My response to Sartorio’s argument, therefore, is to deny premise 2) – I am responsible for the explosion in Two Buttons simply because my omission caused the explosion. This is compatible with Causal Condition, and is what we get from applying an independently plausible account of causation, Minimal Sufficiency, to the case – an account that Sartorio appears to draw on, albeit implicitly, in her argument for why I  am responsible for the explosion. Admittedly, this view has some counterintuitive consequences. It implies, for example, that my failure to press my button caused the explosion – and hence, presumably, 172

Responsibility and causation

that I am responsible for the explosion – in Two Buttons – One Stuck as well.24 But it doesn’t follow from this that every omission is a cause of the explosion, contrary to what Sartorio seems to suggest. And it’s perhaps worth reminding ourselves just how plausible Causal Condition seemed at first blush. The idea that we are responsible only for what we cause is a deeply ingrained feature of our ordinary thinking about responsibility – the argumentative burden on one who wishes to deny it is high. I submit that this burden has not yet been discharged.

Conclusion In this chapter, I have argued that cases of collective action and omission – though they raise interesting causal questions, such as how to think about the difference between ‘threshold’ and ‘non-threshold’ cases, the difference between partially causing and causing a part, and so on  – in fact pose no threat to the fundamental idea that we are responsible only for what we cause. If correct, this conclusion has important consequences for the philosophy of collective responsibility. The idea that we need a ‘new ethics for a collective age’ is one which has gained currency in recent years, in part because of purported counterexamples to Causal Condition.25 I have not argued against that idea directly here. Nevertheless, the arguments of this chapter might be seen as lending support to the view that there is, in fact, nothing morally sui generis about the responsibility we bear for things we bring about with others.26

Acknowledgements I’m grateful to Maximilian Kiener and Carolina Sartorio for helpful written feedback on an earlier draft of this chapter, and to Imogen Rivers and audiences in Oxford, Amsterdam, Vienna, and San Francisco for valuable discussion.

Notes 1 Clarke (1993), O’Connor (1995), Griffith (2010). 2 “The metaphysical theories of agent-causation . . . try to force autonomy into the objective causal order – giving a name to a mystery” (Nagel 1986: 115). 3 “[I]f a causing of an event need not be an action, what principle can an advocate of the traditional conception of agent causation invoke, to distinguish between those causings of events which are actions and those which are not?” (Alvarez & Hyman 1998: 222) 4 O’Connor (2000: ch.5). 5 Kane (1996), Franklin (2018), Ekstrom (2019). 6 For a summary of the challenges involved, see Dowe and Noordhof (2004). 7 Sartorio (2016), Kaiserman (2022). 8 For a review of different approaches, see Kaiserman (2018). 9 For different perspectives on the epistemic condition, see Robichaud and Wieland (2017). 10 For one recent view on the topic, see Tadros (2021). 11 There are of course some problems with Minimal Sufficiency. First, it’s not clear that there is any non-reductive way of understanding what it is for some events to be ‘sufficient’ for another which rules out the possibility of effects being sufficient for their causes, or effects of a common cause being sufficient for each other. And second, Minimal Sufficiency seems to struggle with so-called ‘late’ and ‘trumping’ pre-emption cases (Schaffer 2000). I won’t comment further on these issues here. For what it’s worth, I’m sceptical that causation is analysable in non-causal terms in the first place; my goal in introducing Minimal Sufficiency is primarily to show how more sophisticated

173

Alex Kaiserman accounts of causation, whether or not they’re ultimately correct, allow us to reconcile Causal Condition with our intuitions in collective action cases. 12 Borrowed from Bernstein (2017). 13 Minimal Sufficiency is more commonly known as the ‘NESS account’ of causation, where ‘NESS’ stands for ‘Necessary Element of a Sufficient Set’. I avoid this terminology, partly because it’s a mistake to think that effects are caused by sets as opposed to their members (compare: Principia Mathematica wasn’t authored by the set containing Russell and Whitehead, it was jointly authored by Russell and Whitehead), but also because it’s unclear from the NESS formulation alone what the members of the set have to be necessary for. Indeed, several authors appear to think (mistakenly) that for X to count as a NESS cause of Y, it must be necessary for the occurrence of Y (for examples of this mistake, see Bazargan-Forward (2017: 332), Kingston and Sinnott-Armstrong (2018: 172) and Kutz (2007: 298)). Minimal Sufficiency is not liable to being misunderstood in this way. See also Gunnemyr (2019: 409) and Rivers (2021) on this point. 14 Examples of this sort feature prominently in Parfit (1984), but can arguably be traced back to Glover (1975). Note that Parfit was concerned with the question of what makes the actions in such cases morally wrong, whereas I am only concerned with the question of what makes the actors responsible for the harm. These questions might be related, but I make no such assumption here. See Nefsky (2019) for a helpful overview of the available views. 15 See also Tiefensee (2022). 16 This is true not only on supervaluationist treatments of vagueness, but even on extremely strong logics of determinacy (e.g., those like S5, which validate the Barcan formula and its converse, and also rule out any higher-order vagueness). My thanks to Alex Roberts for valuable discussion here. 17 I include this qualification because, plausibly, a single event can contribute to different extents to multiple different causings of the very same effect – see Kaiserman (2016). 18 Philosophers disagree about how much responsibility for omissions is direct, as opposed to derivative. Of particular interest is the case of unwitting omissions, like forgetting to pick up some milk on the way home. Some think that we can be responsible for an unwitting omission only in virtue of being directly responsible for some earlier action or decision, of which the omission was a foreseeable consequence (e.g., Fischer & Tognazzini 2009); others disagree (e.g., Clarke 2017). I don’t want to get into this debate here, so I’ll focus on deliberate omissions, and assume along with most others that we can be directly responsible for those. 19 E.g., Schaffer (2005). 20 E.g., Lewis (2004). 21 Saunders System Birmingham Co. v. Adams 117 So. 72 (Ala 1928). 22 Restatement of the Law (Third) of Torts: Liability for Physical Harm §27, Reporters Comments (i). 23 Elayoubi v. Zipser [2008] NSWCA 335. I’m grateful to Abrams (2022) for collating these examples. 24 Interestingly, English law appears ambivalent on the question of whether defendants should be held liable in cases like Two Buttons – One Stuck. In Bailey v. Ministry of Defence [2009] 1 WLR 1052, the defendant was found liable for the claimant’s loss after providing negligent medical care, even though it couldn’t be established that, had the right care been provided, the claimant’s injuries wouldn’t have occurred. By contrast, in McWilliams v. Sir William Arrol & Co. [1962] UKHL 3, the defendant was found not liable for the claimant’s harm after failing to provide him with a harness, on the grounds that the claimant wouldn’t have worn the harness had he been provided with one. See Steel (2015: ch.1) for an attempt to reconcile these judgements. 25 The label is from Kutz (2000). 26 On this point, see also Kaiserman (2021b).

Further reading Kutz (2000) is a canonical example of the sort of view to which I have been responding in this chapter; see Bazargan-Forward (2022) for a more recent development. Sartorio (2016) is an important contribution to the literature on causation and direct responsibility. Nefsky’s (2017) work on

174

Responsibility and causation collective action cases is deservedly influential. For a discussion of the issues in this chapter from a legal perspective, see Wright (1985), Moore (2009) and Kaiserman (2017, 2021b).

References Abrams, Yuval. (2022). “Omissive Overdetermination: Why the Act-Omission Distinction Makes a Difference for Causal Analysis,” University of Western Australia Law Review, 49(1), 57–86. Alvarez, Maria, & Hyman, John. (1998). “Agents and Their Actions,” Philosophy, 73(2), 219–245. Bazargan-Forward, Saba. (2017). “Complicity,” in Marija Jankovic & Kirk Ludwig (eds.), Routledge Handbook on Collective Intentionality. New York: Routledge. ———. (2022). Authority, Cooperation, and Accountability. Oxford: Oxford University Press. Bernstein, Sara. (2017). “Causal Proportions and Moral Responsibility,” in David Shoemaker (ed.), Oxford Studies in Agency and Responsibility: Volume 4. Oxford: Oxford University Press. Clarke, Randolph. (1993). “Toward a Credible Agent-Causal Account of Free Will,” Noûs, 27(2), 191–203. ———. (2017). “Blameworthiness and Unwitting Omissions,” in Dana Kay Nelkin  & Samuel C. Rickless (eds.), The Ethics and Law of Omissions. Oxford: Oxford University Press. Dowe, Phil, & Noordhof, Paul (eds.). (2004). Cause and Chance: Causation in an Indeterministic World. New York: Routledge. Ekstrom, Laura W. (2019). “Toward a Plausible Event-Causal Indeterminist Account of Free Will,” Synthese, 196, 127–144. Fischer, John Martin,  &  Tognazzini, Neil A.  (2009). “The Truth About Tracing,” Noûs,  43(3), 531–556. Frankfurt, Harry. (1969). “Alternate Possibilities and Moral Responsibility,”  Journal of Philosophy, 66(23), 829–839. Franklin, Christopher Evan. (2018). A Minimal Libertarianism: Free Will and the Promise of Reduction. Oxford: Oxford University Press. Glover, Jonathan. (1975). “It Makes No Difference Whether or Not I  Do It,” Proceedings of the Aristotelian Society, 49, 171–190. Griffith, Meghan. (2010). “Why Agent-Caused Actions Are Not Lucky,”  American Philosophical Quarterly, 47, 43–56. Gunnemyr, Mattias. (2019). “Causing Global Warming,” Ethical Theory and Moral Practice, 22(2), 399–424. Kaiserman, Alex. (2016). “Causal Contribution,”  Proceedings of the Aristotelian Society,  116(3), 387–394. ———. (2017). “Partial Liability,” Legal Theory, 23(1), 1–26. ———. (2018). “ ‘More of a Cause’: Recent Work on Degrees of Causation and Responsibility,” Philosophy Compass, 13(7), e12498. ———. (2021a). “Reasons-Sensitivity and Degrees of Free Will,” Philosophy and Phenomenological Research, 103(3), 687–709. ———. (2021b). “Against Accomplice Liability,” in John Gardner, Leslie Green,  & Brian Leiter (eds.), Oxford Studies in Philosophy of Law: Volume 4. Oxford: Oxford University Press. ———. (2022). “Alternative Possibilities in Context,” Inquiry, 65(10), 1308–1324. Kane, Robert. (1996). The Significance of Free Will. Oxford: Oxford University Press. Kingston, Ewan, & Sinnott-Armstrong, Walter. (2018). “What’s Wrong with Joyguzzling?” Ethical Theory and Moral Practice, 21(1), 169–186. Kutz, Christopher. (2000). Complicity: Ethics and Law for a Collective Age. Cambridge: Cambridge University Press. ———. (2007). “Causeless Complicity,” Criminal Law and Philosophy, 1(3), 289–305. Lawson, Brian. (2013). “Individual Complicity in Collective Wrongdoing,” Ethical Theory and Moral Practice, 16(2), 227–243. Lee, Samuel. (2022). “Collective Actions, Individual Reasons, and the Metaphysics of Consequence,” Ethics, 133(1), 72–105. Lewis, David. (2004). “Void and Object,” in John Collins, Ned Hall, & L. A. Paul (eds.), Causation and Counterfactuals. Cambridge, MA: MIT Press.

175

Alex Kaiserman Mackie, J. L. (1965). “Causes and Conditions,” American Philosophical Quarterly, 2(4), 245–264. Moore, Michael S. (2009). Causation and Responsibility: An Essay in Law, Morals and Metaphysics. Oxford: Oxford University Press. Nagel, Thomas. (1986). The View from Nowhere. Oxford: Oxford University Press. Nefsky, Julia. (2017). “How You Can Help, Without Making a Difference,”  Philosophical Studies, 174(11), 2743–2767. ———. (2019). “Collective Harm and the Inefficacy Problem,” Philosophy Compass, 14(4), e12587. O’Connor, Timothy. (1995). “Agent Causation,” in Timothy O’Connor (ed.), Agents, Causes, and Events: Essays on Indeterminism and Free Will. Oxford: Oxford University Press. ———. (2000). Persons and Causes: The Metaphysics of Free Will. Oxford: Oxford University Press. Parfit, Derek. (1984). Reasons and Persons. Oxford: Clarendon Press. Rivers, Imogen. (2021). “Against Making a Difference,” accessed January 12, 2023, from blog.practicalethics.ox.ac.uk/2021/03/oxford-uehiro-prize-in-practical-ethics-against-making-a-difference/. Robichaud, Philip, & Wieland, Jan Willem (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(1), 315–336. ———. (2012). “Two Wrongs Do Not Make a Right: Responsibility and Overdetermination,” Legal Theory, 18(4), 473–490. ———. (2016). Causation and Free Will. Oxford: Oxford University Press. Schaffer, Jonathan. (2000). “Trumping Preemption,” Journal of Philosophy, 97(4), 165–181. ———. (2005). “Contrastive Causation,” Philosophical Review, 114(3), 327–358. Sinnott-Armstrong, Walter. (2005). “It’s Not My Fault: Global Warming and Individual Moral Obligations,” in Walter Sinnott-Armstrong & Richard Howarth (eds.), Perspectives on Climate Change: Science, Economics, Politics, Ethics. Amsterdam: Elsevier. Steel, Sandy. (2015). Proof of Causation in Tort Law. Cambridge: Cambridge University Press. Tadros, Victor. (2021). “Two Grounds of Liability,” Philosophical Studies, 178(11), 3503–3522. Tiefensee, Christine. (2022). “Indeterminacy and Collective Harms,” Philosophical Studies, 179, 3307–3324. Wright, Richard W. (1985). “Causation in Tort Law,” California Law Review, 73(6), 1735–1828. Young, Iris Marion. (2003). “Political Responsibility and Structural Injustice,” The Lindley Lecture, University of Kansas, May 5.

176

13 RESPONSIBILITY AND THE DEEP SELF Monika Betzler

Introduction Philosophers have long sought to identify the relevant agential capacities that allow us to hold one another responsible, attribute actions to one another, and regard each other as appropriate targets of moral assessment. According to a widespread intuition shared by many contemporary theorists of responsibility, we can only be held responsible for what we do when we, as agents, are in some normatively relevant sense at the root of our actions. This seems to be the case if our actions express or disclose something about ourselves that accounts for who we are (Smith 2008: 368). The idea that we must be “behind” what we do and that our actions must “speak for” us if we are to be held responsible for them can be bolstered if we consider a person who acts on a desire that she does not want to have. The unwilling addict, for example, is someone who acts on her desire to take the drug, but who is overwhelmed by a desire she does not really want to act on. Even though the unwilling addict’s desire is in some sense part of her psychology, it is external to her in that it does not stand for her as an agent (Frankfurt 1971: 12; Wolf 1990: 30–33). In contrast to mere desires, urges, passions, compulsions or random firings of our brains that can cause us to act, philosophers make reference to the deep self to demarcate when the agent may be held responsible for an action, and when they may be excused or exempted from responsibility.1 So-called Deep Self Views (which sometimes refer to the true or real self, or are termed identificationist, attributionist, self-disclosure or self-expression views) are put forward to identify those mental states that represent or express the respective person’s genuine agential point of view. Susan Wolf pioneered giving voice to Deep Self Views (though she did not defend them) in her claim that if we are responsible agents, it is not just because our actions are within the control of our wills, but because, in addition, our wills are not just psychological states in us,

177

DOI: 10.4324/9781003282242-21

Monika Betzler

but expressions of characters that come from us, or that at any rate are acknowledged and affirmed by us. (Wolf 1987: 49) The idea of the deep self is thus supposed to pin down the particular way in which we are related to our actions so that they can be attributable to us, and so that in turn – in the absence of certain excusing or exempting conditions – we can also be asked to justify them and therefore be answerable for them and held to account (Gorman 2019; Smith 2012). To understand this, it is helpful to contrast the case of genuine responsibility with that of causal responsibility. For instance, a person might have inadvertently caused an accident because her car’s brakes malfunctioned. She thus brought about an outcome, such as, for example, injuring a pedestrian. She is causally responsible for the accident without having in any way intended it. If she had recently had the brakes checked, she might not have been able to prevent or foresee them failing. In such a case, it becomes quite obvious that this person is also not genuinely responsible for what she causally brought about. But it seems necessary for genuine and morally relevant responsibility that there is a self that must somehow be specified as the locus of the agent’s point of view in order to be a proper target of responsibility: a person P is responsible for A-ing if and only if her deep self is expressed in what she does (Shoemaker 2011: 501; Sripada n.d.: 5). The aim of this chapter is to first highlight the theoretical benefits of Deep Self Views, more generally. In the next section, the most prominent versions of these views to have been developed (mostly in our theorizing about self-governance and autonomy) will be assessed. How well can they account for the deep self? More recent criticisms of these Deep Self Views by responsibility theorists will then be presented. These criticisms refer to cases in which the deep self seems either unnecessary or insufficient for ascriptions of responsibility. I will show how these criticisms motivate refined versions of the Deep Self View and initiate new debates about different kinds of responsibility and the justification of blame. Finally, I will highlight the challenges that Deep Self Views continue to face.

In defense of deep self views There are at least four theoretical features in support of Deep Self Views. First, Deep Self Views seem phenomenologically adequate. We often allude to the idea of a deep self in everyday parlance: we distance ourselves from what we did by explaining that our actions were not expressive of who we are. This idea is suggested in locutions such as: “I was not myself when I did that.” Similarly, we also excuse or exempt others by saying that “they were not really themselves when they acted.” We thereby acknowledge that they were caused to act by forces other than their deep self, such as the drug addict or the person whose deep self is a result of manipulation, such as brainwashing. Deep Self Views can thus make sense of the fact that there is a normatively important difference between forces that just cause us to act and our own agential point of view. While the former can give rise to experiences of alienation, the latter accounts for our sense of proper agency. Second, and as a result, Deep Self Views have explanatory power. They can quite commonsensically account for an important distinction in our responsibility practices and identify those conditions under which we can be held responsible for what we do, and differentiate them from conditions under which we are excused or exempted from responsibility. Deep 178

Responsibility and the deep self

Self Views can narrow the scope of responsibility ascriptions to instances where agents are properly “behind” their action. For example, the compulsion moving a kleptomaniac to steal is such that it exempts them from being morally responsible as it is not the agent acting but rather their kleptomania. Deep Self Views can thus quite easily explain why we excuse or exempt agents who act out of (momentary or deep-seated) compulsions, psychological disorders or mental disabilities. They are thus more informative than the so-called Quality-of-Will Theory (even though these theories can be thought to be related), according to which it is the pure quality of will of the agent (good or bad), which is the target of our responsibility ascriptions (Strawson 1962). But Strawson does not explain when exactly an action displays good or ill will. Deep Self Views can show that the will is good or bad when it issues from the agent’s point of view. Similarly, they show why adult human agents are responsible in a way that animals, infants, and machines – which all lack deep selves – are not (Wolf 1987: 51). Third, Deep Self Views seem to provide good guidance for our responsibility ascriptions. This is because they sidestep metaphysical debates about the compatibility of determinism and freedom, and are thus metaphysically parsimonious. They build on an actualist notion of the deep self that does not presuppose alternative possibilities for ascriptions of moral responsibility (at least in cases of morally criticizable behavior). As a result, whether or not an agent is responsible is dependent not on what that agent would have done in a counterfactual world, but only on how it came about that they did what they did in the actual sequence of events. Hence, the freedom required for responsibility is simply the freedom to express the agential point of view such that the action in question can be said to have been the agent’s action (Strabbing 2016: 738). Similarly, it is also not dependent on any counterfactual attitudes that the agent would have, were they placed in some epistemically more favorable circumstances (Sripada n.d.). Deep Self Views do not establish compatibilism, but they imply that there are no other factors – like alternative possibilities – that render responsibility incompatible with determinism (Lippert-Rasmussen 2003: 350). Fourth, and despite metaphorical language surrounding Deep Self Views that refers to actions “speaking for” the agent or to the agent “owning” their actions, Deep Self Views are not metaphysically mysterious, but conducive to a naturalist explanation of what is at the root of action. As will be highlighted in more detail in the next section, none of the Deep Self Views on offer resorts to some obscure and unanalyzable idea of the self, but rather identifies some distinct and privileged subset of mental states or a connection of such states to account for the agent’s point of view (Gorman 2022a). Despite the differences among various Deep Self Views, they are unified as a class of theories that tries to make sense of the idea that, to ascribe responsibility to us as agents, our actions must reflect the agent’s point of view in some normatively relevant way. Since we shape our relationships in large part on the basis of what the actions of others mean to us, Deep Self Views seem to serve our practical interest in understanding other people through their actions (Reis-Dennis 2018: 484). It should be noted that any conception of the deep self is ontological, not psychological (Jaworska 2007; Sripada 2016: 1212). It thereby focuses on the mental states and their interrelated structures which constitute the deep self in an ontological sense. Deep Self Views therefore typically refer to the agent as “being identified” or revealing themselves as self-integrated. This is to be distinguished, however, from the merely psychological question of what stand the agent takes to their self, what the agent considers as self-disclosing or alienating, and how the agent actively identifies with some of their mental states (Smith 179

Monika Betzler

2008). If the agent is not self-deceived, the psychological sense of a deep self may coincide with its ontological sense, but it is often difficult to avoid errors with regard to what qualifies one’s self as deep. What matters for the ascription of responsibility, however, is the ontological sense of a deep self. Non-alienation is therefore merely an indicator to isolate the deep self.

The variety of deep self views Philosophers have gone a long way in proposing various accounts of Deep Self Views. To isolate what the deep self is, it is necessary to come up with a plausible criterion of “internal sourcehood” (Sripada 2016: 1204), and thus pinpoint the proper location of the deep self with which, as Wolf puts it, “the agent is to be properly identified” (Wolf 1990: 30), and is therefore revealed in acting. All Deep Self Views share the attempt to isolate a particular mental state or a specific connection or “mesh” between mental states that accounts for the agent’s deep self. There are different ways to characterize and compare these accounts. One such way is to distinguish them by virtue of the kinds of mental states they take to be relevant to identify the self. We can thus differentiate between non-cognitive and cognitive accounts of the deep self. Another way, which is orthogonal to the first one, is to account for the deep self by distinguishing between how mental states have to be related. This leads us to differentiate endorsement accounts from accounts referring to coherent structures of mental states defined by their functional role to establish agential identity over time. The different versions of the Deep Self View focus mainly on what the deep self is and how best to capture it, quite independently from questions about responsibility. I will focus on three main contenders and show how they developed in part against the backdrop of criticisms that befall the respective rival versions concerning the proper criterion of internal sourcehood.2 As it will turn out, they all illicitly refer to different values underlying our understanding of self-governance and autonomy. In typically focusing on a single value, all three contenders are beset by problems, due either to their neglect of the other value or to an erroneous view about how these two values are to be connected. Which mental state or set of mental states can be isolated such that the agent can be regarded as being identified with (and potentially non-alienated from) them? According to the hierarchical model of desires, broadly Humean in spirit, it is the specific connection between different kinds of non-cognitive mental states – that is, different kinds of desires – that isolates the agent’s deep self. Harry Frankfurt introduces the idea that what characterizes us as persons is that we can take a stance in relation to our first-order desires. While first-order desires are desires for particular actions, second-order desires – “volitions” in Frankfurt’s terminology – are desires that one of our first-order desires becomes effective and thus our “will” (Frankfurt 1971). To the extent that we act on a first-order desire that is in line with our volition and thus issues from a will we want to have, the action is the result of what is endorsed by us and thus attributable to us, even if we could not have done otherwise (Frankfurt 1969). The action thus issues from our deep self (even though Frankfurt does not use that term). One of the problems that the hierarchical desire model faces is that of an infinite regress. It is by no means clear why volitions – which continue to be non-cognitive states – carry the right kind of authority to account for the agent’s point of view, and need not be further authorized by mental states of an even higher order (Watson 1975: 217; Betzler 2013). 180

Responsibility and the deep self

Frankfurt acknowledges this problem, and proposes various solutions to it (Frankfurt 1992: 104–105). One way to stop the regress is to specify a further characteristic that qualifies the deep self. Frankfurt mentions “wholehearted identification” and “satisfaction” as two ways by which the agent can be thought to issue their point of view. But it remains elusive how the compulsive agent can be distinguished from a non-compulsive agent. After all, a willing addict can wholeheartedly identify or be satisfied with the first-order desire to take the drug and yet not qualify as fully responsible given that their identification or satisfaction is at least partly determined by their addiction. In light of these problems, Gary Watson proposes a Platonic model of the deep self, which resorts to cognitive mental states and abandons the hierarchical model. It is the distinction between judging a state of affairs as good and merely desiring it that helps Watson to show that it is the agent acting in light of their values that accounts for their point of view, and thus their deep self: It is because valuing is essentially related to thinking or judging good that it is appropriate to speak of the wants that are (or perhaps arise from) evaluations as belonging to, or originating in, the rational (that is, judging) part of the soul; values provide reasons for action. (Watson 1975: 208) Only those desires that are grounded in what the agent in question judges as good are those that constitute the deep self. Hence, we are responsible only when we act in accordance with our values. The idea is that an agent cannot be alienated from his entire valuational system as it is that set of considerations that yields judgments about the good (Mitchell-Yellin 2015, 380). Even though the Platonic model is not confronted either with the problem of an infinite regress or with the problem of failing to sufficiently distinguish the deep self from compulsion, there are a host of concerns that can be marshaled against it due to its underlying idealized picture of a unified and rational agent. The first problem is that judgments about what is good neither necessarily account for what is really important to the agent nor capture whether the agent really stands behind what they judge to be good. There is an important difference between judging good and valuing that the Platonic model does not sufficiently appreciate. Consider, for example, a student who decides to go to law school because he thinks a career in law is a good thing to pursue. It might well be the case, though, that the student never manages to value law as a career for him. He might just be unable to be affected by the values connected to law. The recognition of value in judging something to be good is compatible with not being drawn to or attracted by it if one pursues it (Frankfurt 1999: 158). What the student judges to be good may thus not be at all important to him. A second worrisome implication of the Platonic model is that actions cannot be properly attributed to an agent if they issue from some motive other than their value judgments. Cases of so-called inverse akrasia highlight that agents can act on account of an emotion against what they judged best and thereby reveal their deep self. Huck Finn, for example, reflectively judges that it would be best to turn in his friend Jim, the slave. But he can’t bring himself to do this and spontaneously lies about Jim’s whereabouts to those hunting him (Arpaly & Schroeder 1999; Betzler 2022). If the Platonic model were correct, we could not make sense of the idea that we consider Huck Finn as someone who very much acts out of his deep self (Sripada 2016: 1214). 181

Monika Betzler

These criticisms surrounding the Platonic model lend support to a further version of the Deep Self View, the non-cognitive care model. Cares or carings are sui generis mental states that typically consist of a set of emotional, motivational, commitmental, and judgmental dispositional tendencies. In contrast to the Platonic model, the conception of caring that Frankfurt introduces in his later writings shows that we might be unaware of or even mistaken in what we care about, and our cares can even contradict what we judge to be good. But to distinguish caring from mere compulsions, he claims that carings must endure through an exercise of our own “volitional activity” and that the agent in question proves “unwilling to give up” their carings by identifying with a higher-order desire (Frankfurt 1999: 160–161). While Frankfurt’s conception of caring is less dependent on conscious endorsement – thus highlighting that carings reveal themselves rather than being consciously selected by the agent – he cannot do without a variant of reflectiveness. Agnieszka Jaworska, in contrast, argues that to represent the agent’s point of view, caring does not have to be reflexive, but can simply function “to support the psychological continuities and connections that constitute the agent’s identity and cohesion over time” (Jaworska 2007: 559; see also Shoemaker 2003: 90, 2011: 506). She views caring as a structured set of various, less complex emotions, emotional dispositions, and desires that unfold reliably over time in response to relevant circumstances. For example, for a person P to care, she rejoices when the object of her care is flourishing, and is sad or frustrated if it does not fare well. She takes pride in the success of the object of her care and is fearful when it is endangered, etc. This emotional attunement to the object of care and its various circumstances is what imbues this object with importance and reveals that it is a source that commands one’s emotional vulnerability (Jaworska 2007: 561). Cares or carings are thus defined by their unique functional role as this complex set of mental states accounts for the agent’s cohesion over time, and this is what ultimately captures the agent’s point of view (Sripada 2016: 1209–1210). The problem with the care model, however, is that we can care deeply about things we do not want to care about. Even though Sripada thinks that there is a basic conceptual tie between what an agent cares about and what matters to her, it remains questionable whether either one accounts for the criterion of internal sourcehood and thus for the agent’s deep self (Sripada 2016: 1211). Consider someone who was raised in such a way that she cannot help but care about being liked by others. Even if she is prone to the sets of mental states mentioned previously, it remains unclear why this expresses her deep self given that she herself has come to believe that this care is due to influences she herself deems problematic, and wants to get rid of. The care model seems unable to distinguish between the “authentic self” and the “authoritative self,” or between the condition of “non-alienation” and the condition of “sovereignty” (Enoch 2022). While the authentic self seems to be about what the agent cares about, the authoritative self is about the mental states that manifest the agent as the “author of his own life” (Lippert-Rasmussen 2003: 368). These two senses of self may overlap and coincide, but they can also come apart. For example, an agent can decide to take a stand on acting against what they deeply care about. Similarly, an agent can decide to conform to societal expectations. It is thus conceivable that an agent can be inauthentic, yet still authorize what they want to do. The discussion of these three different contenders of the Deep Self View highlights that they trade on different values underlying our understanding of self-governance and 182

Responsibility and the deep self

autonomy. While the care model locates the deep self in the authentic self, the Platonic model regards the authoritative self as the locus of the deep self. Frankfurt’s hierarchical model can be interpreted such that it accounts for the deep self in virtue of a proper connection between the authentic self and the authoritative self. The agent authorizing their authentic self is what seems to isolate the deep self. But while the care model cannot account for authority, the Platonic model cannot account for authenticity (or non-alienation). Frankfurt’s hierarchical model seems unable to properly locate the right connection between authority and authenticity. Consequently, none of the Deep Self Views on offer seems able to fully meet the criterion of internal sourcehood, if this is to account for all the evaluative dimensions that we take to underlie self-governance or autonomy. A way out of this predicament is to focus more on the practical interest we have in referring to Deep Self Views. While the different contenders on offer are mainly concerned with how to account for the criterion of internal sourcehood quite independently from questions about responsibility, responsibility theorists regard our intuitive verdicts about responsibility as important resources to shed new light on how to properly isolate the deep self. Responsibility theorists who defend a deep self theory of responsibility try to be as ecumenical as possible with regard to different versions of the Deep Self Views on offer and articulate the conditions of responsibility that focus on a pluralist understanding of the deep self (Shoemaker 2015a, 2015b; Gorman 2022a). Or, they favor and refine one particular version of the Deep Self View, such as the care model (Sripada 2016), or introduce a new judgment-dependent version (Smith 2005, 2008; Strabbing 2016). Whatever stand one takes on a particular version of the Deep Self View, the question that needs to be addressed – and which the aforementioned versions of the Deep Self View typically do not tackle head on – is how exactly the deep self is connected to our verdicts about the ascription of responsibility. Is the deep self necessary and/or sufficient to account for ascriptions of responsibility? Answers to these questions motivate more sophisticated variants of the Deep Self View.

Ascriptions of responsibility and the deep self: the necessary condition A first set of criticisms points to various kinds of non-deliberate conduct for which we typically take agents to be responsible, yet which cannot be regarded as issuing from their deep selves. For example, Deep Self Views have to regard the weak-willed agent as excused or exempted from responsibility ascriptions. An agent who judges that dieting is best, yet acts on the first-order desire to indulge in eating a cake, does not act on their deep self if this is defined by the desire they want to be their will or by their best judgment or what they care about over time. Since Deep Self Views cannot account for this, the deep self does not seem necessary for responsibility. Similarly, we hold agents responsible for actions resulting from a whim or some spontaneous emotion which sometimes leads us to regard such actions as being “out of character” yet of such a kind that the agent in question qualifies as being responsible for them. For example, an agent who cheats because just once they feel like it, but neither wants this to be their will nor judges this as the best course of action nor really cares about cheating, seems responsible for their cheating even though it does not result from their deep self. In a similar vein, smashing a vase against the wall out of anger does not excuse or exempt us from responsibility, even if we do it “despite ourselves.” 183

Monika Betzler

Sometimes we even fail to do things we should do because we simply forget about them or do not think sufficiently about doing them, like watering the plants or calling a friend on their birthday. Again, we can be held responsible for our failings even though they did not issue from our deep self (Smith 2008; Shoemaker 2011: 505). As a result, responsibility ascriptions extend to cases in which we do not act from what we really want, judge best, or care about. There are at least three broad approaches by which responsibility theorists try to rescue the Deep Self View in light of the challenges that these cases of non-deliberate conduct, including failings, pose. The first response holds that we are responsible not only for our actions, but also for our attitudes, thereby extending that for which we can be held responsible. The so-called rational-relations view maintains that some of our intentional mental states, such as our desires, emotions, intentions, and beliefs, directly or indirectly reflect our evaluative judgments. There is either a direct and thus conceptual connection between the anger we feel, for example, and the judgment that what we are angry about is anger-provoking, or there is an indirect rational connection between our attitudes, including patterns of awareness and our evaluative judgments. If we care about our friend, we should not fail to notice when they need our help. As a result of these connections, we are justified to ask those who hold them to defend or modify them (Smith 2005: 251–258, 2008: 369–371). The rational–relations view locates the deep self (again, without mentioning that term) in the evaluative judgment and its normative connections to other intentional mental states.3 It thus seems able to explain why we can also hold an agent responsible for their nondeliberate conduct and failings. This is the case insofar as the mental states leading to nondeliberate action or failure to act are normatively connected to the evaluative judgments the agent can be taken to hold. Smith’s view of evaluative judgments is broader than Watson’s in that she does not align them with consciously held beliefs but rather considers them as “tendencies to regard certain things as having evaluative significance” that comprise things we care about (Smith 2005: 251). This also explains why weak-willed actions can be attributed to an agent to the extent that they reveal their tendencies to regard what led to the weak-willed action as significant and make it appropriate to ask the agent to defend or modify it. It also explains why compulsions are not subject to responsibility ascriptions as these are not connected to such tendencies. A second response to the criticisms regarding non-deliberate conduct tries to relax the criterion of internal sourcehood that a plausible Deep Self View must meet. It argues that it is necessary to further specify in what way and how much the self must be expressed in an action such that the agent in question qualifies as being responsible for it. It acknowledges that the self is a “mosaic,” potentially conflicted and imperfectly cohering (Arpaly & Schroeder 1999; Shoemaker 2003; Sripada 2016: 24; Sripada n.d.: 4–5), to then qualify which parts of an agent’s self get to count as self-expressive, even if there is no unified self. Such a view is taken to be compatible with different Deep Self Views, including a refined version of the care model (Sripada 2016). Once we understand that attributability is a gradable notion that corresponds to the gradable approval of an agent with regard to their desires or cares to act, the hope is that Deep Self Views can account for the responsibility of agents acting due to weakness of will or other non-deliberate motives. To the extent, then, that an agent would at the time of action approve to some degree of acting on the desire they in fact act upon for some aim of theirs, their action qualifies as sufficiently self-expressive. This makes room for cases

184

Responsibility and the deep self

of weak will and other motives resulting in non-deliberate conduct since all the aims an agent desires are given some weight and are treated as having some weight via the agent’s hypothetical and partial approval or endorsement (Gorman 2019, 2022b; Sripada 2016: 1228). In the case of compulsive actions, in contrast, it is the sheer force of a desire in the absence of further aims that moves the agent to act simply to rid themselves of the compulsive motive, and explains why the agent cannot be held to account for lack of any approval. This so-called minimal approval account of attributability only requires hypothetical and partial endorsement and therefore “less-than-complete identification with one’s action” (Gorman 2019: 145). A third response holds that a judgment responsiveness view can do justice to the challenges mentioned while giving up on the idea of the deep self altogether. What an agent is responsible for is simply broader than their deep self, but can be captured by what the agent did if this results either from their responding to at least one of their judgments about the reasons they had in favor of doing the action in question by doing it, or their failing to exercise their capacity to respond to their judgment about the reasons that they have against doing the action by not doing it (Strabbing 2016: 759). It seems, however, that it is a matter of dispute whether the judgment responsiveness view is different from a Deep Self View or rather an important qualification of it so as to account for motives resulting in non-deliberate conduct and failed actions.

Ascriptions of responsibility and the deep self: the sufficient condition Another major objection holds that Deep Self Views fail to provide a sufficient condition of responsibility. The actions of evil psychopaths are a case in point. Susan Wolf introduces JoJo, the son of a dictator who, despite the fact that he acts from his deep self, is not responsible for his heinous deeds. What distinguishes JoJo from responsible human beings is that he is not sane. To meet the condition of sanity, an agent must know what they are doing and know that what they are doing is right or wrong. Sanity is thus the “minimally sufficient ability cognitively and normatively to recognize and appreciate the world for what it is” (Wolf 1987: 56). The sane deep-self view that Wolf offers can explain why JoJo – similarly to people with deprived childhoods and those who cannot but follow the problematic norms of their societies – is not fully responsible. Since JoJo cannot help but have a mistaken conception of value and thus is deprived of an ability to know right from wrong, he is “unable to evaluate his character and revise it from a rational and reasonable standpoint” (Wolf 1987: 59). It is this ability that explains why only sane selves can properly be accorded responsibility. As a result, a further normative competence condition must be met by an appropriate view of the deep self. A further concern that Wolf raises against Deep Self Views is that they commit us to a superficial kind of responsibility. We might be open to various forms of evaluations on the basis of our deep selves, but we are not open to blame or praise that marks out responsibility. This is the case because it is not the agent’s fault that they are the way they are. Deep Self Views thus only allow us to identify the self’s role in the causal series that brings about the action in question, but they do not let us hold the agent responsible in “a more serious way” (Wolf 1990: 41). In response to this concern, philosophers have distinguished between different kinds or “faces” of responsibility. Deep Self Views explain why an action is attributable to

185

Monika Betzler

an agent. But, in response to Wolf, Watson denies that attributions of responsibility are superficial. After all, they often reveal something morally relevant about the agent’s “fundamental evaluative orientation” (Watson 1996: 271). However, in addition to responsibility-as-attributability, we also hold agents accountable for what they did, which “is not just a matter of the relation of an individual to her behavior” (Watson 1996: 262). Instead, we demand certain conduct from each other by, for example, our blaming practices. But it is this feature of responsibility-as-accountability that raises issues of fairness. After all, it seems unfair to hold accountable agents for whom it is simply too hard to meet the standards by which we would typically hold them to account (Watson 1996: 281). Accordingly, Watson thinks that our practices of accountability should be moderated in such cases. The rational–relations view aligns with Watson’s account in that it takes a responsible agent’s behavior to reveal something morally relevant about that agent, no matter what their circumstances (Smith 2005, 2008; see also Sher 2006). As Smith, for example, points out, it is an agent’s judgments that, in their action, reveal themselves to be quite independent of their history. She also concedes, however, that an agent, although fully responsible, can still be blameworthy in different ways (Smith 2008: 268). In a similar vein, some philosophers hold that moral protest can be more appropriate in such cases than blame (Hieronymi 2004). While the minimal approval account of attributability similarly recognizes that further conditions that must be met for an agent whose action can be attributed to them can be thought to merit blame or other reactive attitudes (Gorman 2019: 141), it has also been argued that blame can be justified even if it is not grounded in the deep, but rather in the “shallow self” (Reis-Dennis 2018: 487). Reis-Dennis refers to the example of a college student who drunkenly stumbles into his friends and starts a racist and sexist rant. Even though his rant may be genuinely out of character, it would still be fitting for his friends to feel indignant and expect an apology (Reis-Dennis 2018: 489). If this is right, we can hold others justifiedly accountable without being able to ground blame in the deep self. If the deep self is not necessary for blame, we need a different justification for blame. The suggestion is that blame is shaped by other factors that affect the action’s meaning, such as the resultant consequences, its bearing on the relationship in question, and the legitimate expectations of affected others that might be violated. Blame and other reactive practices allow us to stand up for others and ourselves, restore the social order, and, eventually, restore our relationships. Deep Self Views are thus taken to overestimate the extent to which our responsibility ascriptions target agents and their internal states rather than actions and “the whole state of affairs those actions give rise  to” (ReisDennis 2018: 498).

Whither the deep self? Deep Self Views are motivated by the hope that we can discern what accounts for the agent’s point of view – the criterion of internal sourcehood – which is relevant for ascriptions of genuine responsibility. None of the Deep Self Views, as they have been initially put forward, meets this criterion satisfactorily. As I have tried to show, this is due to their implicit commitment to different values that we take to underlie our understanding of selfgovernance and autonomy. None of them seems able to account for all the values related to the authoritative self, the authentic self, and the proper connection between them. 186

Responsibility and the deep self

Responsibility theorists have then taken different directions to hold on to the view that the deep self is in some sense necessary and sufficient for responsibility. After all, we need an account of the agent’s point of view to make sense of our ascriptions of responsibility. Responsibility theorists have therefore extended or relaxed the criterion of internal sourcehood to make sense of our intuitions about responsibility ascriptions. Even though these new proposals make it possible to ascribe responsibility for nondeliberate conduct and failings, they either affirm the value of the authoritative self in a revised judgment-dependent version (Smith 2008; Strabbing 2016), or they offer a watereddown model of both the authoritative and the authentic self such that nothing particularly deep seems to remain about the alleged deep self (Gorman 2019; Sripada 2016). As for ascriptions of responsibility, the rational–relations view considers evil psychopaths as responsible despite their lack of normative competence. The minimal approval view contents itself with responsibility-as-attributability, but it leaves us at a loss for how to make sense of accountability. As a result, the further developments that responsibility theorists offer to rescue the Deep Self View also struggle to account for the right kind of connection between the agent’s point of view and our ascriptions of genuine responsibility, not just in the minimal sense of attributability, but also in the full sense of accountability.

Notes 1 While an exemption denies responsibility altogether, an excuse accepts responsibility but averts (or mitigates) blameworthiness (Duff 2007). 2 For reasons of space, I will not discuss Michael Bratman’s theory about self-governing policies that is also one of the Deep Self Views to which responsibility theorists refer (Bratman 1996). 3 While Smith does not explicitly consider herself to be a deep self theorist, it makes sense to interpret her rational–relations view as an attempt to locate internal sourcehood (see also Shoemaker 2011).

Further reading Frankfurt, H. (1971). “Freedom of the Will and the Concept of a Person,” in H. Frankfurt (ed.), The Importance of What We Care About: Philosophical Essays (pp. 11–25). Cambridge: Cambridge University Press, 1988. Frankfurt’s account of persons can be regarded as the first contemporary Deep Self View. Gorman, A. (2019). “The Minimal Approval View of Attributability,” in D. Shoemaker (ed.), Oxford Studies in Agency and Responsibility, vol. 6 (pp. 140–164). Oxford: Oxford University Press. This is one of the most sophisticated Deep Self Theories on offer. Sripada, C. (2016). “Self-Expression: A  Deep Self Theory of Moral Responsibility,”  Philosophical Studies, 173, 1203–1232. This is another sophisticated Deep Self Theory that has recently been developed. Watson, G. (1996). “Two Faces of Responsibility,” in G. Watson (ed.), Agency and Answerability: Selected Essays (pp. 260–288). New York: Oxford University Press, 2004. G. Watson develops his view in response to S. Wolf’s claim that responsibility in the case of evil psychopaths is not deep. He shows that we can distinguish between responsibility-as-attributability and responsibility-as-accountability. Wolf, S. (1987). “Sanity and the Metaphysics of Moral Responsibility,” in F. Schoeman (ed.), Responsibility, Character, and the Emotions: New Essays in Moral Psychology (pp. 46–62). Cambridge: Cambridge University Press. This Is the Article in Which S. Wolf Gives Voice to Deep Self Views and Develops Her Sane Self View in Response. ———. (1990). Freedom Within Reason. New York: Oxford University Press. A  groundbreaking book on responsibility where S. Wolf argues that responsibility ascriptions in the case of evil psychopaths are not deep.

187

Monika Betzler

References Arpaly, N., & Schroeder, T. (1999). “Praise, Blame and the Whole Self,” Philosophical Studies, 93, 161–188. Betzler, M. (2013). “Einleitung: Begriff, Kontext und Konzeptionen der Autonomie,” in M. Betzler (ed.), Autonomie der Person (pp. 7–36). Münster: Mentis. ———. (2022). “Inverse Akrasia: A Case for Reasoning About One’s Emotions,” in C. Bagnoli (ed.), Time in Action: The Temporal Structure of Rational Agency and Practical Thought (pp. 243–263). London: Routledge. Bratman, M. (1996). “Identification, Decision, and Treating as a Reason,” Philosophical Topics, 24, 1–18. Duff, A. A. (2007). Answering for Crime: Responsibility and Liability in Criminal Law. Oxford: Hart Publishing. Enoch, D. (2022). “Autonomy as Non-Alienation, Autonomy as Sovereignty, and Politics,” The Journal of Political Philosophy, 30, 143–165. Frankfurt, H. (1969). “Alternate Possibilities in Moral Responsibility,” in H. Frankfurt (ed.), The Importance of What We Care About: Philosophical Essays (pp.  1–10). Cambridge: Cambridge University Press, 1988. ———. (1971). “Freedom of the Will and the Concept of a Person,” in H. Frankfurt (ed.), The Importance of What We Care About: Philosophical Essays (pp. 11–25). Cambridge: Cambridge University Press, 1988. ———. (1988). The Importance of What We Care About: Philosophical Essays. Cambridge: Cambridge University Press. ______. (1992). “The Faintest Passion,” in H. Frankfurt (ed.), The Importance of What We Care About: Philosophical Essays (pp. 95–107). Cambridge: Cambridge University Press, 1999. ———. (1999). Necessity, Volition, and Love. Cambridge: Cambridge University Press. Gorman, A. (2019). “The Minimal Approval View of Attributability,” in D. Shoemaker (ed.), Oxford Studies in Agency and Responsibility, vol. 6 (pp. 140–164). Oxford: Oxford University Press. ———. (2022a). “Demystifying the Deep Self View,” Journal of Moral Philosophy, 19, 390–414. ———. (2022b). “What Is the Difference between Weakness of Will and Compulsion?” Journal of the American Philosophical Association online first, 1–16. Hieronymi, P. (2004). “The Force and Fairness of Blame,” Philosophical Perspectives, 18, 115–148. Jaworska, A. (2007). “Caring and Internality,” Philosophy and Phenomenological Research, 74, 529–568. Lippert-Rasmussen, K. (2003). “Identification and Responsibility,” Ethical Theory and Moral Practice, 6, 349–376. Mitchell-Yellin, B. (2015). “The Platonic Model: Statement, Clarification and Defense,” Philosophical Explorations, 18, 378–392. Reis-Dennis, S. (2018). “Responsibility and the Shallow Self,” Philosophical Studies, 175, 483–501. Sher, G. (2006). “Out of Control,” Ethics, 116, 285–301. Shoemaker, D. W. (2003). “Caring, Identification, and Agency,” Ethics, 114, 88–118. ———. (2011). “Moral Responsibility and the Self,” in S. Gallagher (ed.), The Oxford Handbook of the Self (pp. 487–518). Oxford: Oxford University Press. ———. (2015a). Responsibility from the Margins. New York: Oxford University Press. ———. (2015b). “Ecumenical Attributability,” in R. Clarke, M. McKenna, & A. Smith (eds.), The Nature of Moral Responsibility: New Essays (pp. 115–140). New York: Oxford University. Smith, A. M. (2005). “Responsibility for Attitudes: Activity and Passivity in Mental Life,” Ethics, 115, 236–271. ———. (2008). “Control, Responsibility, and Moral Assessment,”  Philosophical Studies, 138, 367–392. ———. (2012). “Attributability, Answerability, and Accountability: In Defense of a Unified Account,” Ethics, 122, 575–589. Sripada, C. (2016). “Self-Expression: A  Deep Self Theory of Moral Responsibility,”  Philosophical Studies, 173, 1203–1232. ———. (n.d.). “At the Center of Agency, the Deep Self,” unpublished manuscript. Strabbing, J. T. (2016). “Responsibility and Judgment,” Philosophy and Phenomenological Research, XCII, 736–760.

188

Responsibility and the deep self Strawson, P. (1962). “Freedom and Resentment,” Proceedings of the British Academy, 48, 187–211. Watson, G. (1975). “Free Agency,” The Journal of Philosophy, 72, 205–220. ———. (1996). “Two Faces of Responsibility,” in G. Watson (ed.), Agency and Answerability: Selected Essays (pp. 260–288). New York: Oxford University Press, 2004. Wolf, S. (1987). “Sanity and the Metaphysics of Moral Responsibility,” in F. Schoeman (ed.), Responsibility, Character, and the Emotions: New Essays in Moral Psychology. Cambridge: Cambridge University Press, 46–62. ———. (1990). Freedom Within Reason. New York: Oxford University Press.

189

14 RESPONSIBILITY AND THE EMOTIONS Andreas Brekke Carlsson

To be morally responsible for an action is closely connected with being held responsible for that action. According to the Strawsonian tradition, which will be the focus of this chapter, a person is responsible for an action just in case it is appropriate to hold them responsible for that action. We can respond to a moral wrongdoing with a wide range of emotional reactions. We feel disappointment, anger, resentment, indignation, disgust, contempt, disdain, or experience hurt feelings. When we act wrongly ourselves, we often experience guilt, remorse, regret, disappointment, sadness or shame. On an inclusive view, all of these reactions are ways of holding ourselves or others morally responsible; on a restrictive view only some of them are. The aim of this chapter is to provide an overview of some of the many connections between emotions and responsibility. In section  1, I  will highlight a couple of important methodological principles that have framed the debate about moral responsibility since the publication of PF Strawson’s seminal paper, Freedom and Resentment. In section 2, I will present the standard account of the blaming emotions, according to which blame is understood in terms of resentment, indignation, and guilt. In section 3, I will canvas different ways in which these blaming emotions might be appropriate. Sections 4 and 5 critically discuss two common strategies of deriving the conditions of blameworthiness by considering whether and when blaming emotions are appropriate. Section 6 discusses other emotions that do not fit neatly into this picture but that nevertheless seem to be responsibility responses: disappointment, contempt, shame, and hurt feelings. A common thread in the chapter will be a challenge for the Strawsonian tradition. Given that this tradition understands moral responsibility in terms of the appropriateness of blaming emotions, it must provide a realistic picture of these emotions and the practices that rely on them. On the other hand, it also aims to develop a normatively plausible account of the conditions of responsibility. I will try to show that it is often difficult to strike a satisfactory balance between these two aspirations.

1  The connection What is it to be morally responsible for an action, an attitude or an omission? We can approach this question from two different directions. On the one hand there are the DOI: 10.4324/9781003282242-22 190

Responsibility and the emotions

conditions of responsibility. A seagull or a toddler might harm others in various ways, but we don’t normally think of them as morally responsible for such harms. This is because seagulls and toddlers lack certain capacities. Because of this lack, they are exempted from responsibility. But even people who possess the general capacities required for moral responsibility might not be responsible for the actions they perform. On any occasion they might lack control, knowledge, be under duress, etc. In such cases, they might have an excuse. By systematizing what gives rise to exemptions and excuses, we might piece together the conditions of moral responsibility. But note that the way we discover what the conditions of responsibility are is by thinking about under which conditions it would be appropriate to hold someone responsible. This suggests another way of approaching the question of what it is to be morally responsible, namely by asking what it is to be held responsible for an action, attitude or omission. Responses to wrongdoing are multifaceted, and include punishment, sanctions, and differential distribution of resources. However, according to one very influential line of thought, the core of our responsibility practices is non-institutional and interpersonal. According to Strawson (1962), holding someone responsible is essentially a matter of responding to their actions, attitudes and omissions with reactive attitudes, such as resentment, indignation or gratitude. Such attitudes – “the non-detached attitudes and reactions of people directly involved in transactions with each other”  – presuppose participating in human relationships; they are emotional responses to the good or ill will of others. As such they differ from the objective attitude we take towards people when we see them as someone “to be managed or handled or cured or trained; perhaps simply to be avoided” (1962: 52). Focusing on the situations when we excuse or exempt people from the reactive attitudes will provide the conditions of moral responsibility. Strawson focuses on resentment. He asks us to think about the occasions in which an offended person will “naturally or normally” experience resentment. He then asks us to “consider what sorts of special considerations might be expected to modify or mollify this feeling or remove it altogether” (1962: 79). Typical excusing considerations will be: ignorance (“she didn’t know”, “she didn’t mean to”), duress (“he was pushed”, “he had to do it”), lack of control (“she couldn’t help it”). Typical exempting consideration will involve more general features which make resentment inappropriate: “He’s only a child”, “His mind has been systematically perverted”, “That’s purely compulsive behaviour on his part” (1962: 79).1 Given this interconnection between the conditions of responsibility and our practices of holding someone responsible, most participants in the responsibility literature accept something like the following biconditional: Strawson’s principle: An agent S is responsible for an act, attitude or omission X if and only if it is appropriate to hold S responsible for X.2 The most familiar way of holding someone responsible is by blaming them. Of course, we are also morally responsible for good actions and attitudes. But in this chapter, I will focus on blame and blame-like responses rather than praise (see Telech, this volume). Adapting the previous schema to blameworthiness in particular rather than responsibility in general, we get: An agent S is blameworthy for an act, attitude, or omission X if and only if it is appropriate to blame S for X. This widely accepted generic schema raises two questions: What is blame? (see also Menges, this volume), and what do we mean by “appropriate”? 191

Andreas Brekke Carlsson

2  Emotions and blame While there is disagreement concerning the exact nature of emotions, most philosophers agree upon three features that emotions share and which, taken together, set them apart from other mental states such as moods, beliefs, or desires (Menges 2017). First, emotions typically have a characteristic phenomenology. It feels like something to be in the grip of an emotion. Anger, for example, is often accompanied by sensations of one’s raised heart rate and muscular tension and a heated, aggressive feeling. Emotions also often have a hedonic tone. Guilt, for example, is a painful feeling, although the pain of course can be experienced in different strengths and shades. Second, emotions also motivate different kind of actions; they have action tendencies. The action tendency of guilt is to express the emotion to the victim of one’s wrongdoing and attempt to repair the relationship that has been damaged. Shame on the other hand motivates withdrawal or escape from the shame-inducing situation. The action tendency of anger is to confront the offender and lash out at him (D’Arms & Jacobson 2022: 193). Finally, emotions have representational content. Fear represents its object as being dangerous or threatening. Envy portrays one’s rival as, roughly, having something that one lacks, and casts this circumstance in a negative light. Regret represents one’s action as a mistake. What exactly the blaming emotions resentment and guilt represent is a contested issue, which we will return to in the next section. It is important to note that the notion of representational content need not be understood as a thought or belief (Rosen 2015; Tappolet 2016). The content can also be understood as quasi-perceptual seeming, or what Gendler (2008) calls an alief. According to D’Arms and Jacobson (2022), the representational content of an emotion is not something that can be distinguished from the emotion’s phenomenology and its action tendencies. On a Strawsonian account, to blame someone is to target them with one of the reactive emotions. But which ones? The focus, at least since Wallace’s influential Responsibility and the Moral Sentiments (1994), has been on resentment, indignation and guilt.3 We feel resentment when the wrongdoing is directed towards ourselves and indignation when the wrongdoing is directed against others. When we feel guilt, we perceive ourselves as wrongdoers. Resentment and indignation are angry emotions, and the action tendency of anger  – to confront and lash out  – sets them apart from other hostile emotions such as contempt and disdain. Guilt on the other hand is characterized by a painful affect as well as the motivation to express one’s guilt and repair the relationship that has been damaged or threatened by one’s wrongdoing. Despite these differences, resentment, indignation and guilt are often taken to be unified by a common representational content. This was Wallace’s motivation for focusing on this narrow class of reactive attitudes. It would be difficult to find a shared content for love, hurt feelings, resentment and forgiveness. But resentment, indignation and guilt, according to Wallace, all represent the wrongdoer as having violated a legitimate normative expectation. Several other competing suggestions have been made in the literature. These emotions may represent an agent as having acted with insufficiently good will (Graham 2014) and/ or as having had the capacity to act better than she did (Strabbing 2019). We will return to this issue shortly. For now, it is sufficient to note that the assumption of a shared representational content provides the resentment, indignation, and guilt with a unity which strengthens the case for thinking that just these emotions should be singled out as the blaming emotions and thus figure in our account of blameworthiness. 192

Responsibility and the emotions

If we understand blame in terms of these reactive attitudes, we get the following account of blameworthiness: Blameworthiness: Agent S is blameworthy for X if and only if it is appropriate to respond to S’s X-ing with the reactive attitudes of resentment, indignation or guilt (in the reflexive case). This leaves us with our second question. How should we understand the notion of appropriateness?

3  Blaming emotions and their justifications Suppose Smith feels resentment towards Jones. This does not yet settle the question of whether Jones is blameworthy. For him to be blameworthy, Smith’s resentment must be appropriate. There are three main ways of interpreting the notion of appropriateness in the current literature. Forward-looking moral views: One possibility is to justify the expression of blaming emotions on broadly consequentialist grounds (Smart 1961). Recently, some sophisticated versions of this approach have been developed. Vargas (2013) argues that whereas our system of holding each other responsible should be justified by its beneficial consequences, our everyday blaming practices should be justified by what the agent has done. McGeer (2019) and Jefferson (2019) argue that blame is justified in virtue of fostering reason– responsiveness. Fricker (2016) holds that blame may function as a proleptic mechanism: If the wrongdoer did not recognize the reasons he had to act differently, one might nevertheless treat him as if he did recognize these reasons, thereby bringing him to recognize reasons he did not recognize at the time of action. Pereboom, drawing on Talbert (2012) and Smith (2013), has developed a protest view of blame on which blame is justified by largely forward-looking considerations. While forward-looking justifications often draw on a reactive attitude account of blame, they do not have to. In the remainder of the chapter, I will therefore focus on justifications that are more closely connected to emotions. Backward-looking moral views: Another alternative is to understand appropriateness in terms of a backward-looking moral notion such as fairness (Wallace 1994; Watson 2004; Nelkin 2011) or desert (McKenna 2012; Pereboom 2014; Clarke 2016; Carlsson 2017). One way to motivate a moral notion of appropriateness is by considering the action tendencies of the other-directed blaming emotions. Resentment and indignation are forms of anger and the action tendency of anger is to confront and lash out. Watson (2004) and Wallace (1994) emphasize that reactive attitudes come with a disposition to sanctioning behavior. Watson claims that “blaming attitudes involve a readiness to adverse treatment” (2004: 275). Wallace argues that blame “involves a disposition to engage in a variety of sanctioning activities” (1994: 94). It makes sense to ask whether such sanctions are fair or deserved.4 Another motivation takes its starting point in the phenomenology of self-blame. Guilt has a negative hedonic tone; it is partly constituted by its painfulness. Given this, we might ask whether the painfulness of guilt is deserved (Clarke 2013; Carlsson 2017; Portmore 2019). Desert is often understood as a moral notion, which is a consideration of justice and entails non-instrumental goodness.5 Fittingness views: The final view prefers a non-moral kind of propriety. Fittingness is a non-moral normative relation that obtains between, e.g., shame and the shameful, 193

Andreas Brekke Carlsson

admiration and the admirable, blame and the blameworthy, etc. For a response to be fitting is for its object to call for, or merit, or be worthy of that response. One way of understanding the fittingness relation, particularly popular in the literature on blame and emotions, is in terms of truth or correctness.6 As noted, emotions have representational content. An emotion, on this view, is fitting only if what the emotion represents is true. In this sense, the fittingness of an emotion is similar to the truth of a belief (D’Arms & Jacobson 2000). Rosen (2015) calls it the “Alethic view”.7 Fittingness in the alethic sense is not a moral kind of propriety. It can be fitting to feel envy towards one’s rival or to be amused by an immoral joke. Shame is fitting to what is shameful; amusement is fitting to what is amusing, etc. Fear of tigers is fitting; fear of kittens is not. It is not fitting to envy one’s rival for being wealthier than you when she is in fact approaching bankruptcy; and it is not fitting to regret something that was not a mistake. Blaming emotions will thus be fitting only when their representational content is correct. If resentment represents the wrongdoer as having violated a legitimate expectation, resentment is only fitting if the wrongdoer in fact has violated a legitimate expectation; if resentment represents the wrongdoer as having acted with ill will, resentment is only fitting when the wrongdoer acted with ill will.

4  Fair and deserved emotions The notion of appropriateness and the account of the nature of the blaming emotions one opts for will have important implications for how one thinks about the conditions of blameworthiness. Proponents of the backward-looking moral view have typically defended rather strict conditions of knowledge, control, reason–responsiveness and voluntariness. The thought is simple: if blame is harmful to its recipient, that harm will be unfair or undeserved, unless she satisfies certain stringent conditions. We can illustrate this with the following argument.8

The interpersonal argument for control (1) An agent S is blameworthy for X only if S deserves to be targeted with the reactive attitudes resentment or indignation because of X. (2) To be targeted with the reactive attitudes resentment or indignation is to suffer. (3) S deserves to suffer for X only if X was under S’s control. Therefore (4) S is blameworthy for X only if X was under S’s control. Is premise (2) true? That depends on how we understand the notion of blame. To “target” someone with resentment or indignation could be understood either as experiencing or expressing these emotions. Blame can be private or outwardly expressed. If blame is identified with the reactive attitudes, it seems natural to understand blameworthiness in terms of the appropriateness of experiencing these emotions. After all, there seem to be numerous instances of blame that are not expressed. We blame people from afar and we blame the dead. Even though experiencing the blaming emotions involves a disposition to treat the wrongdoer in a certain way, this disposition need not be manifested (Nelkin 2013: 124; Graham 2014: 391). Moreover, even when the blaming emotions are expressed, they will 194

Responsibility and the emotions

not always be experienced as sanctions. Sometimes we do not care about the disapproval of others. If this understanding of blame is correct, it is not obvious that blame entails suffering. An alternative approach would be to focus on guilt, rather than resentment and indignation (Clarke 2013, 2016; Carlsson 2017, 2019; Portmore 2019, 2022). Guilt doesn’t need to be expressed in order for the person at whom it’s targeted to suffer, because suffering is necessarily part of experiencing guilt. This suffering comes in different degrees. It can vary from a mild discomfort to a prolonged state of agony (Clarke 2013: 155). But if the emotional state does not involve suffering at all, it is not guilt. This provides the material for another, intrapersonal, argument for why control is required for blameworthiness (Carlsson 2017; Portmore 2019):

The intrapersonal argument for control (1) An agent S is blameworthy for X only if S deserves to feel guilt for X. (2) To feel guilt is to suffer. (3) S deserves to suffer for X only if X was under S’s control. Therefore, (4) S is blameworthy for X only if X was under S’s control. This argument faces a different set of challenges. First, one might object to the claim that emotions can be deserved. According to Shoemaker (2015), questions of desert only arise when some kind of treatment is at issue. But experiencing a reactive attitude is not a treatment. Second, and relatedly, considerations of parsimony might favor the fittingness view. Consider the relation between emotion and value. Something is admirable if and only if it is fitting to admire, regrettable if and only if it is fitting to regret. These biconditionals are widely accepted (Howard 2018; Berker 2022).9 D’Arms and Jacobson (2022) ask why blameworthiness should be different in that it, alone among these values, requires desert and not just fittingness to be appropriate.

5  Fitting emotions Another way of linking emotions and blameworthiness is the following: The Alethic Account of Blameworthiness: An agent S is blameworthy for X if and only if S is a fitting target of a blaming emotion (Hieronymi 2004; Graham 2014; Rosen 2015; Shoemaker 2015; Strabbing 2019; Portmore 2022). The alethic view is not merely an account of what it is to be blameworthy. It also aims to provide an explanation of the conditions of blameworthiness. For any condition of blameworthiness, such as freedom, control, bad quality of will, knowledge, etc., there must be a corresponding thought or representation present in the blaming emotion. If any such thought is not present in the blaming emotion, it is not a condition on blameworthiness. This follows from the claim that an agent is blameworthy if an only if the thoughts or representation involved or constitutive of the blaming emotions are true of her. 195

Andreas Brekke Carlsson

There will be some constraints on any plausible alethic account. First, the account should be informative. Suppose that the representational content of blaming emotions is simply that the person being blamed is blameworthy. This would be phenomenologically plausible, but would not yield an informative account of moral blameworthiness. After all, the alethic account seeks to illuminate the conditions of blameworthiness by considering what blaming emotions represent. Second, the representational content of the blaming emotions should not entail any clear conflicts with our considered judgments about the conditions of blameworthiness. Suppose that the content of the blaming emotions were simply that the agent acted wrongly and with a bad or insufficient quality of will. This will result in a lot of false positives. It seems that small children for example can act wrongly, and with a bad quality of will. Yet many would be reluctant that say that they are blameworthy in virtue of these features of their act. Their bad quality of will does not make it fitting for us to resent them. The natural explanation for why small children are not blameworthy is that they lack some capacity or form of competence necessary for being blameworthy. More generally, commonsense morality recognizes a wide variety of excuses and exemptions: ignorance, lack of control, compulsion, etc. One challenge for the alethic account is to identify a representational content of blaming emotions that can capture these conditions. Third, the account should provide a psychologically plausible account of what blaming emotions represent. For this reason, it seems clear that the representational content cannot be that the agent acted wrongly with an insufficient quality of will and did not act under compulsion, with sufficient control, with sufficient knowledge, etc. This would be an implausibly complex representation. One possibility would be to say that the blaming emotions represent the wrongdoer as having committed an unexcused wrongdoing. While solving the problem of how to make sense of excuses, this suggestion gives rise to other problems. As Owens (2012: 33) has pointed out, excused wrongdoing is just another name for inappropriate blame. The previous suggestion would thus be uninformative. A better proposal comes from Strabbing (2019) who argues that the representational content of blaming emotions is that 1) the agent acted from an insufficiently good will and 2) that she could have done better. This would account for why we often think that children, psychopaths, and manipulated agents are not blameworthy even though they can manifest a bad quality of will: they lack the capacity to act better than they did. However, it is not obvious that this account is psychologically plausible. We often feel guilty when we cause other people harm, even though we could not have acted differently (more on this shortly). Or consider being mistreated by someone who lacks the capacity to act better. Resentment may seem, even on reflection, a natural reaction. However, if guilt or resentment represent the agent as having the capacity to act better, it seems that guilt and resentment should evaporate once we become aware that these agents do not have the capacity to act better. But it is not clear that they do. One way to amend this problem would be to postulate that desert is part of the representational content of blame. According to Rosen (2015) resentment is partly constituted by the thought that the wrongdoer deserves to suffer for what she has done. According to Portmore (2019), guilt is partly constituted by the representation that the wrongdoer deserves the unpleasantness of her guilt. For both Rosen and Portmore, blame is therefore only fitting if either suffering in general or the unpleasantness of guilt is deserved. This leaves room for ethical theorizing about when suffering is deserved. Rosen’s and Portmore’s accounts can thus incorporate the ethical arguments we encountered in the previous section. One can 196

Responsibility and the emotions

argue that people do not deserve harm or unpleasantness unless certain conditions – control and knowledge, say – are satisfied. At the same time, they can maintain that blaming emotions are fundamentally experienced rather than expressed, and that the relevant norm of propriety for all emotions is fittingness rather than desert. The reason for this is that desert enters the picture as the emotion’s representational content rather than its norm of propriety. It is a neat package. However, its plausibility relies on the claim that desert in fact is part of the representational content of the blaming emotions. This is not obvious. Desert, on most accounts, entails non-instrumental goodness. But it might be possible to resent a family member or a co-worker without having the thought that it is non-instrumentally good that they suffer. Portmore’s desert claim is more specific, since it is restricted to the unpleasantness of guilt, but faces another problem. In general, it is unusual that emotions are constituted by a thought or representation concerning their own justification. Emotions do not tend to be self-referential in this way: The representation content of regret, for example, seems to be that I made a mistake, but not that I made a mistake and that it is appropriate that I pained by this mistake. In addition, it’s crucial to Portmore’s project of deriving the conditions of blameworthiness from the representational content of guilt that desert is part of what guilt represents. But even if guilt is self-referential in this way, it is hard to establish that the relevant notion is desert, rather than some other normative notion.10 Alethic accounts face a challenge. On the one hand, they need to capture at least some of the commonly accepted conditions on blameworthiness and the corresponding excuses and exemptions. On the other, they need to give a psychologically realistic picture of the reactive attitudes. That this might be difficult to achieve is nicely illustrated by some recent work on guilt and resentment. Standard accounts of guilt take guilt to represent the agent as having violated a legitimate expectation, acted with ill will or acted wrongly and betrayed a personal relationship. What these views have in common is that they all presume that the agent has acted objectionably. Zhao (2020) points to common experiences of guilt that do not fit this pattern. Survivors of atrocities often feel guilty because they survived and others did not. Moreover, we often feel guilty in cases where we are merely causally responsible for bad outcomes. A natural response would be to insist that while these cases may be instances of guilt, this guilt is not fitting. When we encounter people in the grip of guilt for actions for which they are not responsible, we will typically point out that they are not responsible for them, it wasn’t their fault, they couldn’t have acted differently, etc. The presumption seems to be that these considerations undermine the appropriateness of the emotion. However, if these emotional experiences are in fact instances of unfitting guilt, people in the grip of them must represent themselves as having violated a legitimate expectation or acted with ill will or having betrayed a personal relationship. Zhao argues that such interpretations are implausible. Instead he proposes the following representational content of guilt: “Guilt represents some part of the self as being implicated in a bad state of affairs” (Zhao 2020: 23). Next consider resentment. Reis-Dennis (2021) notes that it is typically considered unfitting for adults to resent children when they act wrongly. Strawson, as we noted previously, took childhood to be a prime example of something that would exempt someone from resentment. Children, however, do resent each other and this resentment often seems fitting: “We know from observing children, from consuming literature and film, and from our own memories of growing up, that a child’s full-blooded resentment in response to being bullied is often intelligible and apt” (2021: 5). Reis-Dennis’ explanation 197

Andreas Brekke Carlsson

is that children are typically too socially weak to be fittingly resented by adults, but not by each other. He argues that a psychologically realistic account will show that resentment is sensitive to relative social power. As a consequence, some adults might fulfill all the traditional conditions of moral responsibility (knowledge, control, bad quality of will) and nevertheless not be fittingly resented if their social status is too low. Children, on the other hand, might fittingly resent each other even though they would lack some of the conditions normally taken as necessary for fitting resentment (knowledge, control, or reason–responsiveness). These revisionary accounts of guilt and resentment raise two issues. First, if Zhao and Reis-Dennis are correct, this will create difficulties for the alethic account by leading to the following dilemma: We could maintain that agents are blameworthy if and only if resentment and guilt are fitting. But this would seem to give us the unwelcome result that agents can be blameworthy for merely being implicated in a bad state of affairs, or that people can fail to be blameworthy although they fulfill the traditional criteria for blameworthiness, if their perceived social status is sufficiently low. Alternatively, we might try to avoid these results by distinguishing between different kinds of guilt and resentment.11 One might allow that there are kinds of guilt and resentment that do not seem to track culpable wrongdoing, but insist that there are other forms of guilt and resentment that do. One problem for this approach is that it may seem ad hoc. After all, the guilt and resentment described by Zhao and Reis-Dennis share both the phenomenology and action tendencies with guilt and resentment as these emotions are normally conceived. The second worry is methodological. Even if Zhao’s and Reis-Dennis’ analyses of guilt and resentment turn out to be incorrect, it seems possible that they, or something like them, could have been correct. Indeed, given that psychological states are the product of our cultural and evolutionary history, it seems to me that we should not be too surprised if resentment and guilt fail to match up perfectly with our considered judgement about the conditions of blameworthiness. The representational contents of our blaming emotions seem contingent in a way the conditions of blameworthiness arguably are not. The question, given the Strawsonian methodology, is what to do with such a divergence.

6  Other ways of holding responsible: disappointment, contempt, shame, and hurt feelings Although the main focus of Strawson’s essay is on resentment, he understands the reactive attitudes as a big and diverse family, including attitudes like “gratitude, resentment, love, and hurt feelings”. Recently there has been a resurgence of work on the broader class of reactive attitudes. First, consider disappointment. Some forms of disappointment are impersonal. I can be disappointed that it rains or that I was rejected. But I can also be disappointed in a person for something they did. Let us call the latter reactive disappointment (Telech & Katz 2022).12 Disappointment is often viewed as a non-retributive alternative to blame, as an emotional response that would be justified even if we lacked the kind of control necessary to deserve to be objects of angry emotions like resentment and indignation (Pereboom 2014; Milam 2016: for discussion, see Menges 2020). Telech and Katz (2022), however, argue that we should see reactive disappointment as form of blame, rather than as an alternative. They note that disappointment, although not an angry emotion, certainly has condemnatory force. When

198

Responsibility and the emotions

you say: “I’m disappointed in you!” you address me and expect an uptake in the form of acknowledgement and repair. Indeed, a common response to reactive disappointment is the feeling of guilt. Consider next contempt and shame. These emotions differ from resentment and guilt in several important respects. First, they are not merely responses to moral failures. We can feel contempt towards the moral failings of others and shame for our own moral shortcomings. But we can also feel contempt towards someone for being bad at sports or choosing a safe but boring occupation. We can feel ashamed of our cowardice, but also of our bodies, or our parents. Second, contempt and shame are globalist attitudes: they target what we are, rather than what we do (Mason 2003; Bell 2013; Deonna et al. 2012). Third, they have different action tendencies. Whereas guilt and resentment, as we have seen, come with a disposition to engage with the victim or wrongdoer, the characteristic action tendency of both contempt and shame is withdrawal. Nevertheless, it seems plausible that at least some instances of contempt and shame are forms of blame. As hostile, affective evaluations of agential features they are ways of holding each other responsible. Just like resentment and indignation, they fall between mere grading and sanctioning. For our purposes there are two things to note about disappointment, contempt and shame. First, they are all harmful. Contempt and disappointment have a sting, or condemnatory force. It is painful to be on the receiving end of them. It is by no means obvious that it is better to be the object of these emotions rather than resentment and indignation (Menges 2020). Shame on the other hand is intrinsically painful, just as guilt. Nevertheless, it also seems clear that the conditions for when it is appropriate to feel disappointment, contempt and shame are less strict than conditions for resentment, indignation and guilt. Telech and Katz note that disappointment is a response to failing to live up to a normative hope. But what we can hope for is different from what we can demand or reasonably expect. It seems that I can be disappointed in a friend for not meeting me, even though I know that his depression makes it unreasonable to expect this from him. We feel contempt for aspects of people’s personality over which they do not have control. It is controversial whether it would be appropriate to feel guilt for unwitting omissions, actions done from moral ignorance, and our involuntary emotional reactions. But it does seem appropriate to feel ashamed for these things if they reflected our character. One way to deal with this issue is to make a distinction between different kinds of responsibility (Watson 2004; Shoemaker 2015). According to Watson (2004) we should distinguish between responsibility as attributability and responsibility as accountability. To be responsible in the attributability sense means that a certain kind of evaluation, what Watson calls aretaic appraisal, is appropriate. This evaluation concerns how well or poorly an agent’s actions, omissions or attitudes reflect on her character, on what she cares about or stands for: she might be kind, cowardly, brave, or selfish. To be blameworthy in the accountability sense means that some more robust kind of blame than mere aretaic appraisal would be licensed. One could then argue that contempt, disappointment and shame, as global evaluations of one’s character and cares, are blame in the attributability sense, whereas resentment and indignation are blame in the accountability sense (Shoemaker 2015; Carlsson 2019).13 One could also argue that the relevant norm of propriety for accountability is desert, whereas the relevant norm of propriety for attributability is fittingness (Carlsson 2019; Portmore 2019). This solution is not entirely satisfactory, though.

199

Andreas Brekke Carlsson

We still need a plausible explanation of why guilt, but not shame, disappointment and contempt should be governed by desert.14 Contempt, shame and disappointment differ from resentment and guilt both in their objects, and in the conditions that make them appropriate. Yet all of them can plausibly be viewed as kinds of blame, and all of them require some of the traditional conditions of responsibility in order to be appropriate. Now consider the following cases from Shoemaker (2019: 130): Trashed gift: Mitzy drops by her close friend Livia’s house unannounced only to see a birthday gift she had given Livia last week in the trash. Better caul Saul: After a lot of bad blood between them, an older brother, Chuck, tells his younger brother Jimmy (who has idolized Chuck) that it’s time to move on, saying, “The truth is, you never really mattered that much to me”. When faced with these events, Shoemaker points out that it would be natural for Mitzy and Jimmy to experience hurt feelings: distressing affective responses to emotional injury. These painful emotional states are typically caused by perceived relationship denigration, but can also result from for example betrayal, humiliation, verbal aggression, shock, ill-conceived humor, or discouragement. There are many situations in which hurt feelings would be appropriate, but resentment would not. Hurt feelings, Shoemaker notes, pose a methodological conundrum for the Strawsonian tradition. On the one hand, Shoemaker makes a strong case for the claim that interactions involving hurt feelings are an integral part of our responsibility practices. We experience hurt feelings as responses to an agential feature that is attributable to the one that hurts us, namely, what they think and feel about us. This agential feature can cause a fitting reactive attitude (hurt feelings). When hurt feelings are expressed, they often have a blaming edge to them. Moreover, the response one seeks from someone who hurt our feelings are those characteristic of our responsibility practices: apologies, acknowledgement, and guilt (Shoemaker 2019, 2022). On the other hand, Shoemaker argues, hurt feelings seem fitting in absence of any of the commonly accepted conditions of responsibility. Given that hurt feelings are reactions to what other think and feel about us, they do not require knowledge, control, reason-responsiveness, or a bad quality will. We can hurt others’ feelings while being ignorant, coerced, and with the very best of intensions. A  small child can hurt its grandmother’s feelings by saying he doesn’t like her, and a cat might hurt its owner’s feelings by preferring someone else’s lap. Because of these considerations we might be tempted to discard hurt feelings when theorizing about responsibility. But this seems to entail a break with the Strawsonian methodology of taking our practices of holding each other responsible seriously.

7  Concluding remarks In this chapter, I have highlighted the challenge of making sense of the conditions of responsibility while at the same time remaining faithful to the ways we hold each other responsible by experiencing and expressing emotions. This is not an easy task once we take an inclusive view of the emotions involved in holding each other responsible, as exemplified by disappointment, shame, or hurt feelings. But even if we take a restrictive view and focus on resentment and guilt, the challenge remains difficult. 200

Responsibility and the emotions

Notes 1 Strawson’s further strategy was to argue that none of these familiar excuses and exemptions concerns determinism. 2 There is a lively debate on whether the left side of this biconditional explains the right side or vice versa, i.e., whether someone is responsible in virtue of being appropriately held responsible or the other way around. See Brink and Nelkin (2011), McKenna (2012) Shoemaker (2017), Todd (2016), Menges (2021), and Clarke and Rawling (2023). 3 See for example: McKenna 2012; Graham 2014; Rosen 2015; Strabbing 2019. 4 This view would be even more plausible if we take expressed blame to the fundamental notion in our analysis of blameworthiness (McKenna 2012). 5 Desert is also sometimes seen as species of fittingness. See McKenna (2022). 6 It is important to note that many philosophers working on fittingness explicitly reject the equation between fittingness and correctness. See for example Howard (2019) and Naar (2021). 7 Rosen (2015: 70–71) uses the term “alethic” rather than the term “fitting”, because it takes fittingness to be a sui generis non-moral relation. 8 For versions of this argument, see Wallace (1994); Watson (2004); Nelkin (2011); and Pereboom (2014). 9 However, see Achs and Na’aman (forthcoming) for important qualifications. 10 See Achs (2022) for a view on which emotions are self-referential, and the relevant normative notion is not desert. 11 For the view that there are two distinct kinds of guilt, see Nichols and Prinz (2010: 134). 12 As opposed to “propositional disappointment”. See Menges (2020) for a similar distinction. 13 However, Wang (2021) argues that shame or a subcategory of shame is a reactive attitude tied to accountability, and Telech and Katz (2022) argue that reactive disappointment is a form of accountability blame. 14 Portmore’s explanation is that guilt but not shame has desert as part of its representational content. This solution, though, is open to the worries mentioned in the previous section.

Further reading A natural starting point for the connection between emotions and moral responsibility is Strawson’s “Freedom and Resentment” (1962). For a systematic treatment of resentment, indignation, and guilt, and the moral view of their justification, see Wallace, Responsibility and the Moral Sentiments (1994). Shoemaker’s Responsibility from the Margins (2015) gives a more inclusive account of the emotions involved in holding each other responsible and focus on their fittingness. Gideon Rosen’s “The Alethic Account of Moral Responsibility” (2015) provides a clear presentation of the alethic view. For a forward-looking justification of the reactive attitudes, see McGeer’s “Scaffolding Agency: A Proleptic Account of the Reactive Attitudes” (2019). For more on self-directed blaming emotions, see the chapters in my edited volume Self – Blame and Moral Responsibility (2022).

References Achs, R. (2022). “Blame’s Commitment to Its Own Fittingness,” in C. Howard  & R. J. Rowland (eds.), Fittingness: Essays in the Philosophy of Normativity. Oxford: Oxford University Press. Achs, R., & Na’aman, O. (Forthcoming). “The Subtleties of Fit,” Philosophical Studies. Bell, M. (2013). Hard Feelings: The Moral Psychology of Contempt. Oxford: Oxford University Press. Berker, S. (2022). “The Deontic, the Evaluative, and the Fitting,” in C. Howard & R. J. Rowland (eds.), Fittingness: Essays in the Philosophy of Normativity. Oxford: Oxford University Press. Brink, D., & Nelkin, D. (2011). “Fairness and the Architecture of Responsibility,” in D. Shoemaker (ed.), Oxford Studies in Agency and Responsibility. Oxford: Oxford University Press. Carlsson, A. B. (2017). “Blameworthiness as Deserved Guilt,” The Journal of Ethics, 21, 89–115. ———. (2019). “Shame and Attributability,” in D. Shoemaker (ed.), Oxford Studies in Agency and Responsibility (6, pp. 112–139). Oxford: Oxford University Press.

201

Andreas Brekke Carlsson Clarke, R. (2013). “Some Theses on Desert,” Journal of Ethics, 16, 153–164. ———. (2016). “Moral Responsibility, Guilt, and Retributivism,” Journal of Ethics, 20, 121–137. Clarke, R., & Rawling, P. (2023). “Blameworthiness and Dependence,” The Philosophical Quarterly, https://doi.org/10.1093/pq/pqad011. D’Arms, J.,  & Jacobson, D. (2000). “The Moralistic Fallacy,” Philosophy and Phenomenological Research, 61, 65–90. ———. (2022). “The Motivational Theory of Guilt (and its implications for Responsibility),” in Carlsson (ed.), Self-Blame and Moral Responsibility. Cambridge: Cambridge University Press. Deonna, J., Rodogno, R., & Teroni, F. (2012). In Defense of Shame. Oxford: Oxford University Press. Fricker, M. (2016). “What’s the Point of Blame?” Nous, 50, 165–183. Gendler, T. S. (2008). “Alief and Belief,” The Journal of Philosophy, 105(10), 634–663. Graham, P. A. (2014). “A Sketch of a Theory of Moral Blameworthiness,” Philosophy and Phenomenological Research, 88(2), 388–409. Hieronymi, P. (2004). “The Force and Fairness of Blame,” Philosophical Perspectives, 18, 115–148. Howard, C. (2018). “Fittingness,” Philosophical Compass, 13(11), e12542. Jefferson, A. (2019). “Instrumentalism About Moral Responsibility Revisited,” The Philosophical Quarterly, 69(276), 555–573. Mason, M. (2003). “Contempt as a Moral Attitude,” Ethics, 113(2), 234–272. McGeer, V. (2019). “Scaffolding Agency: A Proleptic Account of the Reactive Attitudes,” European Journal of Philosophy, 27(2), 301–323. McKenna, M. (2012). Conversation and Responsibility. NewYork: Oxford University Press. ———. (2022). “Fittingness as a Pitiful Intellectualist Trinket?” in C. Howard  & R. J. Rowland (eds.), Fittingness: Essays in the Philosophy of Normativity. Oxford: Oxford University Press. Menges, L. (2017). “The Emotion Account of Blame,” Philosophical Studies, 174(1), 257–273. ———. (2020). “Blame It on Disappointment: A Problem for Skepticism about Angry Blame,” Public Affairs Quarterly, 34(2), 169–184. ———. (2021). “Responsibility and Appropriate Blame: The No Difference View,” European Journal of Philosophy, 29(2), 393–409. Milam, Per-Erik. (2016). “Reactive Attitudes and Personal Relationships,” Canadian Journal of Philosophy, 46(1), 102–122. Naar, H. (2021). “The Fittingness of Emotions,” Synthese, 199(5–6), 13601–13619. Nelkin, D. (2011). Making Sense of Freedom and Responsibility. Oxford: Oxford University Press. ———. (2013). “Desert, Fairness, and Resentment,” Philosophical Explorations, 16(2), 117–132. Nichols, S., & Prinz, J. (2010). “Moral Emotions,” in J. M. Doris (ed.), The Moral Psychology Handbook (pp. 111–146). New York: Oxford University Press. Owens, D. (2012). Shaping the Normative Landscape. Oxford: Oxford University Press. Pereboom, D. (2014). Free Will, Agency, and Meaning in Life. Oxford: Oxford University Press. Portmore, D. (2019). “Desert, Control and Moral Responsibility,” Acta Analytica, 34(4), 407–426. ———. (2022). “A Comprehensive Account of Blame,” in A. Carlsson (ed.), Self-Blame and Moral Responsibility. Cambridge: Cambridge University Press. Reis-Dennis, S. (2021). “Rank Offence: The Ecological Theory of Resentment,” Mind, 130(520), 1233–1251. Rosen, G. (2015). “The Alethic Conception of Moral Responsibility,” in R. Clarke, M. McKenna, & A. Smith (eds.), The Nature of Moral Responsibility: New Essays. Oxford: Oxford University Press. Shoemaker, D. (2015). Responsibility From the Margins. Oxford: Oxford University Press. ———. (2017). “Response- Dependent Responsibility: Or a Funny Thing Happened on the Way to Blame,” Philosophical Review, 126, 481–527. ———. (2019). “Hurt Feelings,” The Journal of Philosophy, 116(3), 125–148. ———. (2022). “Responsibility: The State of the Question Fault Lines in the Foundations,” The Southern Journal of Philosophy, 58(2), 205–237. Smart, J. C. C. (1961). “Free-will, Praise, and Blame,” Mind, 70, 291–306. Smith, A. (2013). “Moral Blame and Moral Protest,” in Randolph Clarke, Michael McKenna,  & Angela Smith (eds.), The Nature of Moral Responsibility: New Essays. Oxford: Oxford University Press.

202

Responsibility and the emotions Strabbing, J. T. (2019). “Accountability and the Thoughts in Reactive Attitudes,” Philosophical Studies, 176(12), 3121–3140. Strawson, P. F. (1962). “Freedom and Resentment,” Proceedings of the British Academy 48: 1–25. Reprinted in and all page references from Gary Watson (ed.), Free Will, (2nd ed., pp.  72–93). Oxford: Oxford University Press, 2003. Talbert, M. (2012). “Moral Competence, Moral Blame, and Protest,” The Journal of Ethics, 16(1), 89–109. Tappolet, C. (2016). Emotions, Values, and Agency. Oxford: Oxford University Press. Telech, D., & Dahan Katz, L. (2022). “Condemnatory Disappointment,” Ethics, 132(4), 851–880. Todd, P. (2016). “Strawson, Moral Responsibility, and the ‘order of explanation’: An Intervention,” Ethics, 127(1), 208–240. Vargas, M. (2013). Building Better Beings: A Theory of Moral Responsibility. Oxford: Oxford University Press. Wallace, R. J. (1994). Responsibility and the Moral Sentiments. Cambridge, MA: Harvard University Press. Wang, S. T. (2021). “Shame and the Scope of Moral Accountability,” The Philosophical Quarterly, 71(3), 544–564. Watson, Gary. (2004). Agency and Answerability. Oxford: Oxford University Press. Zhao, M. (2020). “Guilt Without Perceived Wrongdoing,” Philosophy and Public Affairs, 48, 285–314.

203

15 VARIETIES OF ANSWERABILITY Maximilian Kiener1

Many scholars characterize responsibility in terms of answerability. In so doing, they elaborate on when and why it is apt, fitting, or even obligatory for a person to explain, justify, or excuse their conduct to others, not because doing so would have beneficial consequences but rather because there is some intrinsic connection between said conduct and a person’s moral agency. Answerability, so understood, does not yet imply liability to praise or blame. One can be answerable for causing some harm, yet still avert blame by providing an excuse or justification. However, answerability implies one’s eligibility for such a response, making one a proper target for some form of moral assessment.2 In this chapter, I shall discuss three of these approaches in greater detail. They all start from the description of answerability given so far but differ from each other by explaining answerability more specifically in terms of the reflection of evaluative judgments (section 1), as a social or legal practice (section 2), or as a disposition for an inner dialogue (section 3). For each approach, I shall lay out its key ideas and present a novel response to a core challenge. My overall aim is to show that an adequate view of moral responsibility must incorporate insights from each of these approaches and, most notably, make use of their potential to cross-fertilize each other.

I  Answerability and the reflection of judgment Various moral philosophers have considered responsibility in terms of answerability. For instance, Angela Smith holds that ‘a morally responsible agent is one who can intelligibly be asked to “answer for” her attitudes and conduct’ (Smith 2015: 103). T.M. Scanlon states that someone is responsible for his conduct when ‘it is appropriate in a general sense to ask him to defend it or to disown it’ (Scanlon 2000: 289). And David Shoemaker claims that (at least one important type of) responsibility ‘is something agents have in virtue of their ability (in principle) to “answer for” their actions’ (Shoemaker 2015: 64). To answer for one’s conduct requires a person ‘to give her (justificatory)  reasons for thinking, feeling, or acting in the way she has’ (Smith 2015: 103). Thus, the kind of explanation which the answerable person is expected to give is not just a purely non-rational causal explanation. Rather, it is a request for justification, for an account of someone’s DOI: 10.4324/9781003282242-23 204

Varieties of answerability

motivating reasons, aims, and values, or simply for an explanation of what good they envisaged as arising from their conduct. For this reason, an answerable person is significantly different from an eyewitness who provides testimony, or from a behavioral scientist who provides scientific explanations. But philosophers also add that a person is answerable for something, such as some conduct, action, or even belief, if and only if it reflects this person’s judgment. For instance, Smith explains that ‘[i]n order for an agent to be answerable for something, . . . the thing in question must in some way reflect her own judgment or assessment of reasons’ (Smith 2015: 103). Shoemaker (2015: 82) and Scanlon (2000: 276) agree with this requirement too. The relevant (evaluative) judgments or assessments, Smith further specifies, ‘are not necessarily consciously held propositional beliefs, but rather tendencies to regard certain things as having evaluative significance’ (Smith 2005: 251). Moreover, Smith adds, ‘such judgments make  up the basic evaluative framework through which we view the world. They comprise the things we care about or regard as important or significant’ (Smith 2005: 251–252). But what exactly does it mean that some conduct or action reflects an evaluative judgment? In David Shoemaker’s view, conduct or action reflects a person’s evaluative judgment when it attracts the first-personal emotional reactions of ‘regret and pride, although the wider emotional family here includes disapproval, disappointment, irritation, frustration, and shame, in negative cases, and approval, appreciation, and approbation, in positive cases’ (Shoemaker 2015: 65). Moreover, Shoemaker is more specific: in cases of harmful or wrongful action, answerability attracts the kind of regret that comes with ‘a motivational impulse to change one’s deliberative policies in the future’ (Shoemaker 2015: 67). Shoemaker calls this ‘agential regret’ and explicitly contrasts it with Bernard Williams’s ‘agent-regret’ (Williams 1981), which need not include any motivation to change one’s deliberative policies at all (Shoemaker 2015: 68). In Smith’s view, on the other hand, evaluative judgments do not have to correlate with a particular set of emotional responses. Smith also focuses more closely on mental states, rather than action or conduct, and describes three different ways in which such mental states can reflect evaluative judgments: directly, indirectly, and through what I term ‘idealised reflection’. Mental states, including emotions and attitudes, directly reflect evaluative judgments when there is ‘a conceptual connection between them [in the sense that] . . . [t]hese attitudes are partly constituted by these evaluative judgments’ (Smith 2005: 270). For instance, if I face an alligator, my fear is partly constituted by my judgment that I am in danger, since my fear is directly based on my explicit judgment about the threat posed by this animal and would neither arise nor could be understood without reference to this judgment. Mental states indirectly reflect evaluative judgments when ‘[t]here is, if not a conceptual connection, at least a rational connection between . . . unreflective patterns of awareness and what we care about or regard as important or significant’ (Smith 2005: 270). For instance, if I  regularly fail to notice when my friends need my help, this may reflect an implicit evaluative judgment about the value of our friendship, even if it is not as explicit as my belief about the alligator’s threat from the previous example. However, despite the differences between direct and indirect reflection, Smith is keen to emphasize that, in both cases, there is a rational connection, not a purely causal connection, between mental states and evaluative judgements, and in both cases the protagonists 205

Maximilian Kiener

are active rather than passive with regard to these mental states since they involved ‘the activity of evaluative judgment’ (Smith 2005: 237). Therefore, these cases are different from situations in which, for instance, I feel nauseous before speaking in public because I judge it to present a risk of embarrassment or failure, or feeling nervous before boarding a rollercoaster that I think is scary. In these situations, my emotions and mental states may also depend on my judgments. But the relation between them and my judgements, Smith adds, is ‘purely causal’ rather than ‘rational’ (Smith 2005: 258), and we are ‘essentially “passive” with regards to these states’ (Smith 2005: 257). Finally, Smith offers a third possibility for reflecting judgments, which  – for want of a better term – I call idealised reflection. Smith attends to this third possibility after acknowledging that the rational connection between our mental states and our evaluative judgments can sometimes fail without us ceasing to be answerable. For instance, ‘I may continue to fear spiders even though I  judge them not to be dangerous or threatening in any way’ (Smith 2005: 253). Here, Smith states, my fear does not reflect my actual judgment, yet I  remain answerable for what appears to be an irrational emotion. For this reason, Smith specifies that, for someone to be answerable, it is sufficient that mental states are ‘linked to particular judgments in such a way that if one sincerely holds a particular evaluative judgment, then the mental state in question should (or should not) occur’ (Smith 2005: 253). Here, the ‘should’ is ‘the should of rationality and, therefore, marks a normative ideal which our actual attitudes may not always meet’ (Smith 2005: 253, emphasis added). Accordingly, I  remain answerable for my irrational fear in the spider example because Fear is the kind of state that we take to be, ‘in principle,’ sensitive to our evaluative judgments; the fact that such states are not always ‘in fact’ sensitive to these judgments shows only that we are not always fully rational. (Smith 2005: 253) This is a striking feature of Smith’s view since it shows that, on her account, the scope of answerability is ultimately determined by what an idealised rational person would think, feel, or do. As a result, Smith, unlike Shoemaker, allows for answerability even when a person’s mental states do not reflect any of their actual judgments. Yet, they both agree that ‘a connection to choice is not an essential condition of responsibility’ (Smith 2005: 237). One can be answerable for attitudes and mental states even if one did not, or could not, choose or voluntarily control them. One of the key controversies around answerability concerns the question of whether answerability is just one type of responsibility among others, or if it is the core of responsibility. Shoemaker supports the former view. He presents the example of a parent of a serial killer, who considers their child a ‘worthless human being’ (Shoemaker 2015: 55). But despite that judgment, the parent also bursts into tears when learning about their child’s upcoming execution (Shoemaker 2015: 55). Shoemaker comments on this case and says that ‘[a]nswerability demands . . . would be senseless’ (Shoemaker 2011: 614): A demand for justification presupposes that the agent as he is is capable of offering a reason-based account of his actions or attitudes. (Shoemaker 2011: 614, n. 30);

206

Varieties of answerability

Your demand to me to justify an attitude reflecting a groundless emotional commitment will be without a point as a demand, for I am simply devoid of the resources necessary to engage with your communicative attempt. (Shoemaker 2011: 611) Shoemaker concludes that accounting for this case requires another type of responsibility, which he calls responsibility as attributability, according to which someone is responsible in this sense when it reflects their character, rather than (as in the case of answerability) their judgment. Shoemaker argues that explaining how the parent is ‘responsible’ for their reaction requires going beyond the confines of answerability (Shoemaker 2015). By contrast, Smith claims ‘that there is a single (and unified) concept of moral responsibility underlying our actual moral practices . . . “responsibility as answerability” ’ (Smith 2015: 101). And Smith argues that answerability can deal with the case of the parent. She says: ‘When people are asked to defend their deepest emotional commitments, they are generally not mute: they can and do point to what they take to be justifying reasons’ (Smith 2012: 582). It is most noteworthy that Smith and Shoemaker seem to agree that a person can be answerable for their conduct only if they are able to give a reason-based account of their emotional commitment. However, while Shoemaker denies that the parent can give such an account, Smith argues that they can. Perhaps the parent would just say ‘He’s my son’. The result of this dispute is then an intuitive stand-off over the question of what people can offer in terms of reason-based explanations of their emotional commitments. Smith explicitly concedes that her reply to Shoemaker is an ‘unsatisfying response’ since it does not provide a way of ‘rationally adjudicating our disagreement over these sorts of cases’ (Smith 2012: 582). For this reason, I propose a different reply. I would like to question the view that Smith and Shoemaker agree on here, namely that a person can be answerable for their conduct only if they are able to give a reason-based account of that conduct. To cast doubt on this assumption, consider the following passage from Westlund’s discussion of what she calls dialogical answerability (which I shall discuss in greater detail in section 3): The self-responsible agent may not always know exactly how to answer a particular challenge, but she does appreciate when an answer is owed and is disposed to respond in ways that express this appreciation – by attempting a partial reply, for example, even while admitting its imperfections, or by acknowledging that she must consider the matter further. (Westlund 2003: 506) Westlund alerts us to an important point: sometimes we lack an answer to a question, yet still need to acknowledge the applicability or legitimacy of that question. I suggest that this insight could apply to the serial killer’s parent: the parent may lack a reason-based account for their emotional commitment, yet still accept the applicability or legitimacy of the question put to them, and hence their answerability on this matter. Thus, there are two importantly different types of what we may term a ‘negative’ answer to the demand for a justification. Both answers are negative in the sense that they state the agent under consideration cannot give a reason for their conduct or attitudes. Yet, one negative reply is a dismissal because it also rejects the applicability or legitimacy of the demand

207

Maximilian Kiener

for a justification, whereas the other negative reply is a concession because it accepts the applicability or legitimacy of the demand for a justification. People can reply with a dismissal when they are asked why their nose is so long, but they may only reply with a concession, so I argue, when they are in a situation akin to that of the parent in Shoemaker’s example. (Thus, concessions are cases in which people are unable to give any answer or reason at all for their conduct or attitude, rather than just an answer or reason that falls short of a proper justification.) This new perspective could invite an important conceptual clarification, namely that there can be answerability without an answer (cf. Hubbs 2013). It is surprising that Smith did not allude to this possibility, given that her previous work could accommodate it. Recall Smith’s example of a person who fears spiders, despite judging them not to be dangerous and thus lacking an answer to the question of why they fear spiders. Smith explained that the person remains answerable for their fear because ‘[f]ear is the kind of state that we take to be, “in principle,” sensitive to our evaluative judgments’ (Smith 2005: 253). Similarly, Smith could have replied to Shoemaker that emotional commitments are ‘in principle’ sensitive to our evaluative judgments and therefore the parent remains answerable, even if this particular parent cannot provide an answer. In this way, Smith could offer an account of a concession’ that focuses on the traits of an idealized rational agent, and contrast this view with Westlund’s approach, which would explain a ‘concession’ in terms of situations in which an answer is ‘owed’ in some morally significant way and where people strive for autonomy. Bringing both accounts into discussion on this matter provides a first opportunity for different accounts to cross-fertilize each other: it could help us better understand when and why people remain answerable, despite their inability to give an actual answer.

II  Strict answerability in the law I shall now proceed to an approach that considers answerability not primarily as being concerned with an intra-personal reflection of judgment, but rather with an inter-personal social practice. Here, the law provides a good starting point and can show how the treatment of legal answerability could enrich our understanding of moral answerability too. But first, let me take a step back. In the law, so-called ‘strict liability’ is well known. It means that a person can be convicted or punished under criminal law, or obliged to pay compensation under civil law, simply for performing a certain act (i.e., the actus reus), regardless of any or some so called mens rea like negligence, recklessness, or intent. For instance, if someone causes ‘poisonous, noxious or polluting matter . . . to enter any controlled water’, they commit an offense and are strictly liable for it even if they prove they took reasonable care to avoid the pollution (Water Resources Act 1991, 85 (1); R v Milford Haven Port Authority (2000) 2 Cr App R (S) 423).3 On the other hand, it less well-known that the law also recognizes that answerability can be strict too. Answerability is strict when, again, solely by virtue of having performed a certain act, and regardless of mens rea, a person is obliged to answer in court and, thus, the law is permitted to request a justification. Cases of strict answerability arise in various contexts, including in the criminal law. For instance, if someone sells contaminated food, they can be charged with the offense of ‘[s]elling food not of the nature or substance or quality demanded’, under Section 14 of the Food Safety Act 1990. Yet, at this point, they do not yet incur liability because Section 21(1) also states that it is ‘a defence for the person 208

Varieties of answerability

charged to prove that he took all reasonable precautions and exercised all due diligence to avoid the commission of the offence’. Thus, in this case, liability is not strict because it still requires fault, or mens rea. Yet, answerability is strict because it requires only the selling of contaminated food, irrespective of fault. Such cases of strict answerability can be understood as cases of reverse burdens of proof. Normally, the prosecution must prove both the actus reus and the mens rea before anyone can be summoned to court and held answerable (cf. Woolmington v DPP (1935) AC 462, 481). Yet, the European Court of Human Rights and many domestic legal systems state that the prosecution’s burden of proof can be reduced (Salabiaku v France (1988) ECHR App no 10519/83). In some cases, namely those of strict answerability, the prosecution only needs to prove the actus reus and it is then up to the defendant to disprove the mens rea. In other words, there is now ‘a burden on the defendant to prove some matter, the effect of which is that he is not guilty of the offence charged’ (Dennis 2005: 901). Some arguments for reverse burdens of proof, and thus also for strict criminal answerability, are purely pragmatic, that is they relate to the efficiency of prosecuting and preventing crime. Others, however, are genuine moral arguments. For instance, Simester explains that ‘allocating the burden of proof is part of balancing the costs and risks of an activity between actors and potential victims’ (Simester 2021: 302). If someone voluntarily engages in an optional activity from which he stands to gain but, in so doing, also imposes special risks on others (e.g., by running some profit-oriented but risky business), it seems morally appropriate that he not only receive whatever benefits flow from this activity but that he also bear some of the costs and risks. This moral dimension indicates that strict answerability may be more than a legal peculiarity. It may have fundamental moral significance, and even point to a form of genuine moral responsibility. In fact, Antony Duff argues for exactly this point: For in moral contexts answerability is typically strict: I must answer morally for the harms that I cause even if I cause them through non-culpable accident or inadvertence. If I knock over your vase, or take your umbrella, you will rightly expect me to answer for what I have done. My answer might exculpate me, averting blame by showing that I lacked the moral equivalent of mens rea: I was pushed into the vase by a rushing child; I reasonably believed that it was my umbrella. But it is up to me to provide such an explanation – which is to say that I am morally answerable even for what I do through non-culpable accident or inadvertence. (Duff 2009b: 893–984) (Cf. Duff 2009a: 220, 230) Thus, Duff argues that what is the exception in legal contexts, i.e., that answerability is strict, is normal in moral contexts. In morality, people are obliged to answer for their harmful conduct irrespective of fault. Their explanation, if it is an excuse or justification, may avert liability to blame, but it still admits responsibility as answerability. This approach could follow the characterization of moral answerability from the previous section, namely that ‘a morally responsible agent is one who can intelligibly be asked to “answer for” her attitudes and conduct’ (Smith 2015: 103), where ‘to answer for’ means ‘to give her (justificatory) reasons for thinking, feeling, or acting in the way she has’ (Smith 2015: 103). Yet, this approach is then committed to a less demanding sufficient condition, namely that a person is morally answerable if the person caused some harm and had a prospective obligation not to do so. 209

Maximilian Kiener

As a result, this legally inspired approach to moral answerability differs from the approach to moral answerability in the previous section in two significant ways. First, with strictness, the scope of moral answerability becomes much wider. On the strict answerability view, agents like Bernard Williams’s lorry driver who, through no fault of his own, runs over a child, are morally answerable (Williams 1981). From the perspective of Smith and Shoemaker, however, the lorry driver is a paradigmatic example of someone who is not answerable. Second, the explanation of what makes people answerable is different. Whereas theorists like Shoemaker explain answerability in terms of reflection of judgment, theorists like Duff explain answerability in terms of harms that an agent ought not to have caused. The key objection against strict answerability comes from Shoemaker, who claimed that strict answerability is a contradictory expression: either one is answerable, but then there is no strictness, or there is strictness but then there is no answerability (Shoemaker 2013: 164). This is because, as Shoemaker elaborates, being answerable for something requires ‘access’, not only to the reasons in favor of a course of action, but also to the reasons against it, which he calls ‘instead-of’ reasons (Shoemaker 2015). And, in cases of the completely faultless causation of harm, a person does not have the relevant access. For example, the lorry driver did not have ‘access’ to the relevant reasons because he could not possibly have known that there were reasons against driving down that road at that time. Shoemaker does not define what he means by the term ‘access’, but I understand his meaning to be the ability and opportunity to respond to the relevant reasons. Thus, the lorry driver may have had the ability to respond to the reasons, but lacked the opportunity, and so cannot be answerable. When Duff responded to Shoemaker, he accepted the underlying assumption in Shoemaker’s view, namely that a person is answerable for some harm if and only if (s)he had ‘access’ to the reason not to do it. Yet, Duff offers a different take on ‘access’. He says: ‘what matters is that . . . that reason [i.e., the reason not to do it] is accessible to me, in the sense that I can understand it with hindsight, and could have understood it at the time’ (Duff 2019: 171, emphasis added). In this passage, Duff accepts the requirement of an ability to respond to the relevant reasons at the time of the action, but he rejects an opportunity to have a fair chance to appreciate the reasons in the situation. I think Duff was too quick here because there is an alternative reading of opportunity that is compatible with strict answerability and can answer Shoemaker’s concern more convincingly. The lorry driver has a prospective obligation to guard against the risk of hitting people with his vehicle. Moreover, since this risk is known in advance and is inseparable from driving, he can also take meaningful precautions, such as ensuring the maintenance of his vehicle, driving within the speed limit, and so on. But taking such precautions is essentially a way of responding to the reasons not to hit people because these precautions are ultimately motivated, justified, and demanded by these reasons. Thus, when people have certain specific prospective obligations, they also have an opportunity to take meaningful precautions, and therefore also an opportunity to respond to reasons against certain conduct, or what Shoemaker calls ‘instead-of’ reasons. Thus, there are two readings of ‘opportunity’. In Shoemaker’s view, opportunity requires a fair chance to respond to the reasons against some form of conduct in the situation at hand. By contrast, on an alternative reading, opportunity requires a fair chance to take meaningful precautions. The latter may be sufficient for answerability and thereby show how one could defend strict moral answerability without relinquishing the requirement of an opportunity altogether. Thus, it 210

Varieties of answerability

is here that the proponents of answerability from section 1, who focus on reflection of judgment, and the scholars of strict answerability, discussed in this section 2, could benefit from bringing together their hitherto separate views on what is required to have an ‘opportunity’ to respond to different types of reasons.

III  Dialogical answerability, autonomy, and virtue Finally, Andrea Westlund offers another account of answerability. Westlund asks how autonomy is possible in certain non-ideal situations, e.g., when one lives under oppressive circumstances, and claims that autonomy can persist even in some of these circumstances, as long as a person shows ‘dialogical answerability’ (Westlund 2003, 2009, 2011). Westlund’s view contributed to my earlier discussion of the debate between Smith and Shoemaker in section 1. In what follows, I shall show how it can also enrich our understanding of the relation between responsibility and virtue. By ‘dialogical answerability’, Westlund means the ‘disposition to hold oneself answerable to external critical perspectives on one’s action-guiding commitments’ (Westlund 2009: 26). It is a disposition in the sense that it is about being inclined and ‘prepared to take up and respond to the critical perspectives of others’ (Westlund 2009: 29), it is about one’s ‘readiness to answer . . . in the face of critical challenges’ (Westlund 2009: 34), and it is about ‘be[ing] willing to be engaged in a form of potentially open-ended justificatory dialogue about one’s action-guiding commitments’ (Westlund 2003: 495). Westlund makes clear that this ‘disposition to hold oneself answerable to others is . . . a feature of the agent’s psychology’ (Westlund 2009: 34). Yet, despite being a feature of the agent’s psychology, dialogical answerability also has a clear external component. Westlund says answerability ‘is nonetheless a disposition to be engaged by what is external to the agent, that is, by points of view other than her own’ (Westlund 2009: 34). The disposition to face these external critical perspectives is what gives answerability its ‘irreducibly dialogical form of reflectiveness and responsiveness to others’ (Westlund 2009: 28). But such answerability need not involve an actual dialogue or conversation: [O]ne who holds herself answerable needn’t wait for an actual critic to come along in order to submit her endorsements to critical scrutiny, but will be disposed to engage in an intrapersonal version of critical dialogue in the process of forming and reforming her attitudes towards her desires, preferences and so forth. (Westlund 2003: 497) In addition, dialogical answerability does not require one to give an actual answer or reason, as mentioned in section 1: the ‘self-responsible agent may not always know exactly how to answer a particular challenge, but she does appreciate when an answer is owed’ (Westlund 2003: 506). As a paradigmatic example of someone who falls short of dialogical answerability, Westlund presents Virginia Woolf’s Angel in the House, an exaggerated depiction of the Victorian woman: She was utterly unselfish. She excelled in the difficult arts of family life. She sacrificed herself daily. If there was chicken, she took the leg, if there was a draught, she sat in 211

Maximilian Kiener

it – in short, she was so constituted that she never had a mind or wish of her own, but preferred to sympathise always with the minds and wishes of others. (Woolf 2011: 179, in: Professions for Women [1931]) A woman without a ‘mind or wish of her own’ suffers from self-denial and toxic deference to others that precludes dialogical answerability and thus autonomy. On the other hand, an example of someone who retains their autonomy despite oppressive circumstances, Westlund adds, could be a woman who willingly embraces the ‘strictures imposed by the Taliban in pre-2001 Afghanistan’ (Westlund 2009: 28). To be clear, Westlund still emphasizes that ‘by no means . . . [do] all or even many “Taliban women” accept their condition freely or authentically’ (Westlund 2009: 29). Yet, some of them might, namely insofar as they exhibit ‘a dialogical disposition to hold [themselves] answerable for elements of that hierarchy in the face of critical challenges posed by other agents’ (Westlund 2009: 33). But, in addition to linking her view on answerability to autonomy, Westlund also connects answerability to responsibility. She says that the ‘disposition to answer for oneself, or to be “self-representing” in justificatory dialogue, constitutes a self-orientation to which I refer alternatively as “responsibility for self” or as “self-responsibility” ’ (Westlund 2003: 485). At this point, Westlund’s focus is on responsibility for and to oneself. However, she then goes on to emphasize that such responsibility for and to oneself, which she identifies with dialogical answerability, also holds important connections to responsibility to others, which we normally associate with moral responsibility. As mentioned before, Westlund argues that properly holding oneself answerable requires being ‘prepared to take up and respond to the critical perspectives of others’ (Westlund 2009: 29, emphasis added), so that answerability to oneself and answerability to others are ‘structurally identical’ (Westlund 2011: 170). And conversely, answering to others adequately requires that one also hold oneself answerable because only then would one be properly concerned to give a satisfactory answer to the challenge posed [in the way dialogical answerability suggests] . . . and not just to give whatever answer will in fact suffice, for the time being, to get [one] off the hook with a particular, fallible interlocutor. (Westlund 2003: 497) Thus, in Westlund’s view, meaningful responsibility to others presupposes responsibility to oneself. Stoljar acknowledged Westlund’s attempt to link dialogical answerability to responsibility to others. What is more, Stoljar even highlighted further overlap between Westlund’s view with Angela Smith’s work on moral responsibility. As Stoljar points out, Westlund, like Smith, proposes a notion of answerability that is both relational – because it requires that the agent be thought of as open to interpersonal demands for justification – and nonvolitional. . . . [They both argue] that agents can be answerable and morally responsible even for nonvolitional attitudes and activities such as failing to notice, forgetting, or spontaneous action. (Stoljar 2018: 242)

212

Varieties of answerability

Yet, there are also crucial differences, Stoljar adds, since Westlund portrays answerability as a ‘subjective readiness to hold oneself answerable, whereas . . . [Smith] does not require this’ (Stoljar 2018: 242). As a result, Stoljar objects, Westlund’s ‘dialogical answerability is neither necessary nor sufficient for moral responsibility’ (Stoljar 2018: 243). It is not necessary because ‘people can be answerable and potentially blameworthy even when their characters or their evaluative judgments themselves block the development of the disposition to hold themselves answerable’ (Stoljar 2018: 242). And it is not sufficient because a person’s disposition to hold himself answerable might go too far: Even if the agent with OCD [Obsessive-Compulsive Disorder] has a subjective disposition to answer for his failure to visit his friend, he is not answerable (or morally responsible) because the failure by hypothesis does not bear the appropriate relation to his evaluative judgments. (Stoljar 2018: 243) Stoljar’s points are powerful because they show that a person’s contingent disposition to hold themselves answerable may well diverge from the actual conditions of responsibility. And we may indeed doubt that one’s subjective dispositions should have any impact on one’s moral responsibility at all. One potential reply would be to say that Stoljar misses the point since Westlund’s primary aim was never to outline the conditions of moral responsibility, as opposed to the responsibility to self, or some ideal of being responsible to others. Plausible as this response may be, I think it threatens to hide an important insight into moral responsibility that Westlund’s view, or a further elaboration of it, can provide. To retrieve this insight, we need to look at the debate surrounding moral luck. When discussing cases of moral luck, especially that of Williams’s lorry driver, Susan Wolf places a focus on ‘a willingness to be held accountable for what one does’ (Wolf 2001: 13), which resembles Westlund’s focus on the ‘willing[ness] to be engaged in a form of potentially open-ended justificatory dialogue about one’s action-guiding commitments’ (Westlund 2003: 495). Wolf then admits that people can do too much or too little in this regard, just as Stoljar pointed out. Yet, Wolf also insists that people can get it right, which is ‘a matter of offering the right amount [of willingness to be held accountable] . . . at the right time to the right person in the right way’ (Wolf 2001: 13). And if people get it right, the ‘willingness to be held accountable’ constitutes ‘a virtue that I  suspect we all dimly recognize . . . the virtue of taking responsibility for one’s actions and their consequences’ (Wolf 2001: 13). Wolf says that it is ‘a virtue with no name’ (Wolf 2001: 13), but in some important way ‘akin to the well-established virtue of generosity’ (Wolf 2001: 14) because the person would be less concerned about saving face or retreating to an area where others cannot hold him accountable, and instead would offer himself up as the responsible agent. As Wolf puts it: ‘Generosity generally involves a willingness to give more  – more time, more money, more love, more lenience, more, in one way or another, of oneself than justice requires’ (Wolf 2001: 14), and this is what a person disposed to offer themselves as a responsible person could do. Moreover, Wolf thinks the willingness to be held accountable, especially in cases of moral luck, also expresses ‘our recognition that we are beings who are thoroughly in-the-world, in interaction with others whose movements and thoughts we cannot fully control’ (Wolf 2001: 14).

213

Maximilian Kiener

In these statements, we can find a potential response for Westlund to Stoljar’s criticism. Following Wolf, Westlund could admit that there is room for misuse of the willingness to see oneself as responsible. But she could also insist, as Wolf did, that this still leaves room for proper use. And when people get it right, they have an opportunity for virtue. In fact, Wolf’s view is compatible with Westlund’s account because Wolf too focuses on responsibility to oneself, and not just on responsibility to others. Wolf’s key idea in her paper on moral luck is that people like Williams’s lorry driver, who cause harm without fault or only minimal negligence, can rationally hold themselves responsible to a greater extent than others could hold them responsible. For this reason, Wolf’s view on the virtue of holding oneself responsible is directly relevant to Westlund’s view, and also invites further inquiry into what was key for Westlund as well, namely the relation between responsibility to self and responsibility to others. However, Wolf and Westlund also differ in their views. Wolf is concerned with accountability and blame, whereas Westlund focuses on answerability and dialogue. But this difference need not prevent Westlund from making use of Wolf’s insight about the nameless virtue, precisely because holding oneself responsible in either the accountability or answerability sense can be seen as an act akin to generosity. In addition, Westlund’s emphasis on answerability, the assessment of reasons, the probing of one’s conduct and attitudes, rather than the fittingness of blame, even allows us to go beyond Wolf’s view and identify another equally nameless virtue. What I have in mind is a virtue opposed to complacency and self-indulgence, a virtue which motivates a person to recognize moral and rational demands on their conduct, and not to shy away from them. It is a virtue that makes a person internalize the critical perspectives of others and empathically identify with them through an inner moral dialogue. It is a virtue that resembles, at least in some sense, an ‘ambition’ to strive to align with ideals of rational and moral principles. Pettigrove and Meyer can help us unpack this idea. They suggested that ‘the truly good person is characteristically ambitious about self-improvement’ (Pettigrove & Meye, 2009: 284). And they further claim that ‘Moral ambition is a . . . virtue . . . by dint of its fundamental orientation towards the improvement of the moral quality of a person’s life’ (Pettigrove & Meyer 2009: 294) (see also Pettigrove & Meyer 2009: 297). A person with the disposition to dialogical answerability may embody this virtue of moral ambition because that person will be concerned to give a satisfactory answer to the challenge posed – an answer that meets intersubjectively shareable standards of acceptability, for example – and not just to give whatever answer will in fact suffice, for the time being, to get her off the hook with a particular, fallible interlocutor. (Westlund 2003: 497) Moreover, and as highlighted earlier, the dialogically answerable person will show such moral ambition through her independent moral scrutiny, given that, as Westlund pointed out, she ‘needn’t wait for an actual critic to come along in order to submit her endorsements to critical scrutiny’ (Westlund 2003: 497). Hence, in addition to Wolf’s statements on generosity, Westlund’s view can also, through a similar approach, identify another nameless virtue akin to moral ambition. Therefore, even if Stoljar’s claims about over- and under-inclusiveness are correct, Westlund’s account of dialogical answerability may still be relevant to our understanding of responsibility. The insights from Westlund’s view include the connection between answerability, generosity, 214

Varieties of answerability

and moral ambition, as well as the relation between answerability to oneself and answerability to others. For these reasons, Westlund’s view should not be considered as contributing to the debate on autonomy alone but also be recognized as an important part of our understanding of moral responsibility more generally.

Conclusion This chapter considered three approaches to answerability and addressed a key controversy in each case. Section 1 discussed an intra-personal approach to moral answerability, where answerability is understood in terms of reflecting evaluative judgments, including the challenge of whether answerability always requires the ability to give an actual answer. Section 2 discussed an inter-personal approach, where answerability is understood in terms of a social or legal practice, including the challenge of whether so-called ‘strict’ answerability is a contradiction in terms. Finally, section 3 offered an autonomy-based approach, where answerability is understood as integral not only to responsibility but also to autonomy under non-ideal circumstances, including the challenge of how a certain psychological disposition could ever be relevant to responsibility. Each of these approaches and the discussion of their challenges hold unique insights into moral responsibility, including into the connections between responsibility and judgment (section 1), responsibility and the obligation to explain (section 2), and responsibility and virtue (section 3). What is more, there is potential for the different approaches to complement and inform one another. For instance, Smith (section 1) and Westlund (section 3) could together advance our understanding of when and why people remain answerable, even when they cannot give an answer. Shoemaker’s view (section 1) and the discussion of strictness (section 2) could, in concert, help us to refine the type of ‘opportunity’ to respond to reasons that is required for genuine moral answerability. And the discussion of judgments and obligations (sections 1 and 2) could, together with my interpretation of Westlund’s view (section 3), explore how deontic and aretaic dimensions intersect in moral responsibility. For these reasons, an adequate understanding of responsibility as answerability requires that we bring together, rather than separate, the insights from these three different approaches.

Notes 1 For very helpful comments on earlier versions of this chapter, I thank Hannah Altehenger, Holger Baumann, Juan Pablo Bermúdez, Susanne Burri, Gabriel De Marco, Emma Dore-Horgan, Thomas Douglas, Viktor Ivanković, Sebastian Köhler, Peter Königs, Leonhard Menges, Daniel Miller, Sven Nyholm, Peter Schaber, Marcel van Ackeren, and Andrea Westlund. I especially thank the Leverhulme Trust for outstanding support (project ID: ECF-2021–176). 2 However, some ‘hard incompatibilists’ argue that genuine blame and praise is never appropriate, while still accepting some form of answerability as a framework for valuable moral conversations (cf. Pereboom 2001: 127). 3 I am indebted to Duff’s discussion of these and other cases in Duff (2009a).

Further reading Smith (2005, 2015) and Shoemaker (2015) provide the key texts explaining answerability in terms of reflection of judgment. Antony Duff (2009) offers the most elaborate account of legal answerability, as well as strict answerability. Westlund (2003) presents key ideas on the connection between answerability, autonomy, and responsibility, while Wolf (2001) is a key text on the connection between accountability and virtue.

215

Maximilian Kiener

References Dennis, I. (2005). “Reverse Onuses and the Presumption of Innocence: In Search of Principle,” Criminal Law Review, 901–936. Duff, A. (2009a). Answering for Crime: Responsibility and Liability in the Criminal Law. Oxford: Hart. ———. (2009b). “Legal and Moral Responsibility,” Philosophy Compass, 4(6), 978–986. DOI: 10.1111/j.1747-9991.2009.00257.x. ———. (2019). “Moral and Criminal Responsibility. Answering and Refusing to Answer,” in J. Coates & N. Tognazzini (eds.), Oxford Studies in Agency and Responsibility Volume 5. Oxford: Oxford University Press. Hubbs, G. (2013). “Answerability Without Answers,” Journal of Ethics and Social Philosophy, 7(3), 1–15. Pereboom, D. (2001). Living Without Free Will. Cambridge and New York: Cambridge University Press. Pettigrove, G. A., & Meyer, M. J. (2009). “Moral Ambition,” Australasian Journal of Philosophy, 87(2), 285–299. Scanlon, T. (2000). What We Owe to Each Other. Cambridge, MA and London: Belknap Press of Harvard University Press. Shoemaker, D. (2011). “Attributability, Answerability, and Accountability: Toward a Wider Theory of Moral Responsibility,” Ethics, 121(3), 602–632. DOI: 10.1086/659003. ———. (2013). “On Criminal and Moral Responsibility,” in M. Timmons (ed.), Oxford Studies in Normative Ethics: Volume 3 (pp. 154–178). Oxford: Oxford University Press. ———. (2015). Responsibility from the Margins. Oxford: Oxford University Press. Simester, A. P. (2021). Fundamentals of Criminal Law: Responsibility, Culpability, and Wrongdoing. Oxford: Oxford University Press. Smith, A. (2005). “Responsibility for Attitudes: Activity and Passivity in Mental Life,” Ethics, 115(2), 236–271. DOI: 10.1086/426957. ———. (2012). “Attributability, Answerability, and Accountability: In Defense of a Unified Account,” Ethics, 122(3), 575–589. ———. (2015). “Responsibility as Answerability,” Inquiry, 58(2), 99–126. Stoljar, N. (2018). “Answerability,” in Social Dimensions of Moral Responsibility (pp.  231–252). Oxford: Oxford University Press. Westlund, A. C. (2003). “Selflessness and Responsibility for Self: Is Deference Compatible with Autonomy?” The Philosophical Review, 112(4), 483–523. ———. (2009). “Rethinking Relational Autonomy,” Hypatia, 24(4), 26–49. ———. (2011). “Autonomy, Authority, and Answerability,” Jurisprudence, 2(1), 161–179. Williams, B. (1981). Moral Luck: Philosophical Papers, 1973–1980. Cambridge, UK: Cambridge University Press. Wolf, S. (2001). “The Moral of Moral Luck,” Philosophic Exchange, 31(1), 2–16. Woolf, V. (2011). The Collected Essays of Virginia Woolf. Oxford: Benediction Classics.

Legal References Food Safety Act. (1990). UK. R v Milford Haven Port Authority. (2000). 2 Cr App R (S) 423. Salabiaku v France. (1988). ECHR App no 10519/83. Water Resources Act. (1991). UK. Woolmington v DPP. (1935). AC 462.

216

Section 6: The conditions and challenges of responsibility

16 THE CONSEQUENCES OF INCOMPATIBILISM Patrick Todd

There are several well-known pitfalls when it comes to the task of characterizing incompatibilism concerning responsibility and determinism, pitfalls that are primarily associated with further characterizing the two theses at stake in the alleged incompatibility. In particular: What is determinism? And what is moral responsibility? Both questions have been the subject of vigorous debate. On the first question, P.F. Strawson went so far as to begin his familiar essay on the compatibility problem – “Freedom and Resentment” – by claiming that he doesn’t know what the thesis of determinism is meant to be in the first place.1 On this question, however, the standard – albeit perhaps problematic – characterization in the responsibility literature has been some variant of the following: determinism is the thesis that the past and the laws of nature entail one unique future.2 On the second question, there is vastly more to say than I  could say even in a short book. However, at least since Strawson’s same article, it has been common (although certainly not universal) to characterize moral responsibility in terms of the fairness or desert or propriety of the reactive attitudes, especially the key attitudes of resentment and indignation. This chapter is about certain ways of misconstruing incompatibilism. However, this chapter is not about problems that arise in connection with understanding the two terms of the relevant incompatibility. In this chapter, I address what I regard to be ways of misconstruing incompatibilism that arise in connection, so to speak, with the “incompatibilism” part of incompatibilism – that is, in connection with the relationship envisaged between the two given theses, whatever precisely they are. Most philosophers, of course, know the strict definition of incompatibilism: that there is no possible world in which anyone is both morally responsible (in the sense at sake) and determined (in the sense at stake). More generally: p and q are incompatible if and only if there is no possible world in which both p and q. If we limit our attention to the characterization of responsibility in terms of the fairness/ desert/propriety of resentment and indignation, as I shall in this essay, the strict definition has it that there is no possible world in which anyone is both a fair/deserving/appropriate target of resentment and indignation, and yet also determined. Incompatibilism, however, is often presented directly in terms of its alleged consequences. My focus in this essay is on

219

DOI: 10.4324/9781003282242-25

Patrick Todd

what I take to be an implicit argument operating in the background of much of the literature on responsibility, viz.: The strict modal thesis. There is no world in which anyone is both determined and a deserving target of the reactive attitudes. Therefore, The consequence. If we found out that determinism is true, we would have to give up the reactive attitudes. More generally, my sense is that many philosophers regard this consequence as so obvious so as not to be worth mentioning – so obvious that we can simply and directly present incompatibilism as the second thesis, the thesis that if we found out that determinism is true, we would have to give up the reactive attitudes. For instance, Gary Watson writes that “incompatibilists insist that the truth of determinism would require us to take the objective attitude [i.e., give up resentment and indignation] universally.”3 My claim is that this is a serious mistake, and that there is an attractive, neglected form of incompatibilism that denies the consequence. There are, I  contend, two distinct ways in which the incompatibilist might deny the consequence, corresponding to which of the two premises in the following argument he or she denies: (1) If we found out determinism is true, we would find out that no one deserves the reactive attitudes. (2) If we found out that no one deserves the reactive attitudes, we would have to give up the reactive attitudes. So, (3) If we found out determinism is true, we would have to give up the reactive attitudes. The flip-flopping incompatibilist denies (1), and what I propose to call the innocent incompatibilist denies (2). I investigate each option in turn, although I focus most of my attention on a denial of (2).

Incompatibilities and indicatives I begin with a puzzle about the indicative conditional, which I shall represent throughout as ‘if p, q’. Consider an example. Consider an atheist who accepts, on a priori grounds, the standard argument from evil, according to which the existence of God is incompatible with the existence of evil. That is, this atheist is a God/evil incompatibilist: there is no world in which both God exists and evil exists. Since this atheist takes it as obvious a posteriori that evil does exist, this atheist concludes, of course, that God does not. Now the puzzle. Is this atheist committed to the truth of the following indicative conditional? (4) If God exists, then there is no evil. Not obviously. Indeed, if anything, it is obvious that our atheist will not be willing to accept (4). That is, the atheist who accepts the standard argument from evil is not likely to think that if God actually does exist, then there is no evil. Instead, this atheist is likely to think that if God actually does exist, then her a priori argument from evil is somewhere mistaken, even if she can’t say where. In other words, the atheist is likely to accept that if God actually does exist – has existed this whole time – then of course there is still evil; it is just that God 220

The consequences of incompatibilism

and evil are after all somehow compatible. So this is the puzzle. One can rationally accept that there is no possible world in which both p and q, and yet not accept (the indicative conditional) that if p, ~q. This result is puzzling, but it is compelling – and it is even more compelling on a theory that links belief in an indicative conditional ‘if p, q’ to a disposition to infer q on accepting p.4 In particular, the relevant atheist probably isn’t disposed to conclude that there is no evil on coming to accept that God actually exists, but instead to conclude that the argument from evil was somewhere mistaken. The general point here is the following. One can accept on a priori grounds that there is no world in which both p and q. However, it may nevertheless be the case that one is not disposed to reject q on finding out that the actual world is a p-world – instead, finding out that p may lead one to reject the incompatibility of p and q. The upshot here is perhaps obvious. The incompatibilist about moral responsibility and determinism can certainly believe that there is no world in which anyone both deserves the reactive attitudes and is determined. And yet she may deny premise (1), that if we found out that determinism is true, then we would find out that no one deserves the reactive attitudes. Consider van Inwagen’s well-known stance that if he found out determinism is true, then he would reject his argument that free will (understood as the freedom to do otherwise) is incompatible with determinism.5 Fischer has protested that van Inwagen’s position is somehow unstable, but his objections can, I believe, be countered.6 More generally, however, the God/evil example shows, I believe, that there is in principle nothing problematic about believing that there is no world in which both p and q, and yet not accepting the indicative ‘if p, ~q’. And this is what matters.7 But perhaps we should momentarily back up. The suggestion here is that some incompatibilists will accept that there is no world in which anyone is both responsible and determined, and yet will not accept that if determinism is actually true, no one is responsible. Now, which incompatibilists will take this sort of line? Plainly, it is only the incompatibilist who is also a libertarian – that is, who believes that we are in fact responsible. In point of fact, it is the libertarian who ranks the claim that we are morally responsible above the claim that responsibility is incompatible with determinism in terms of overall plausibility. In other words, we can distinguish between two different types of libertarian: resolute responsibility theorists, and resolute incompatibility theorists. The latter are such that if they found out that determinism is in fact true, they would conclude that we are not responsible. The former, however, are such that if they found out that determinism is in fact true, they would conclude that incompatibilism is false – and so we still are responsible. My point here is simple. To present incompatibilism directly as the view that if we found out that determinism is true – or, more simply, that if determinism is true – then we would have to give up the reactive attitudes is to neglect, without consideration, the views of the libertarians who are resolute responsibility theorists. But this is to say that it is no part of what it is for incompatibilism to be true for the relevant practical thesis to be true. A final point. Consider another salient property commonly attributed to indicative conditionals: the assertion of an indicative ‘if p, q’ presupposes that p is a live epistemic possibility. For instance, consider the following: # (5) See, Anders isn’t in his office. If Anders is in his office, the lights are on. (5) seems infelicitous. Why? Short story far too short: ‘if p, q’ presupposes that p is epistemically possible.8 Now the point. Consider: 221

Patrick Todd

(6) If determinism is true, then no one is responsible. What will the libertarian make of (6)? Prima facie, she shouldn’t accept (6), for (6) carries the presupposition that determinism could be true – which the libertarian will reject. Of course, the libertarian could accommodate the antecedent in (6); she could say, “Of course, determinism isn’t true. But look, I’m fallible, so if I’m wrong about that, and determinism is true, then . . .” But then there is no reason why she couldn’t add, “. . . then the most natural thing to think is that responsibility is somehow compatible with determinism.” Thus, the libertarian needn’t accept (6). Because she needn’t accept (6), she needn’t accept (1), and thus needn’t accept (3) – i.e., the consequence. But this is to say that the strict modal thesis does not entail the consequence.

Denying (2): innocent incompatibilism Recall the key argument from before: (1) If we found out determinism is true, we would find out that no one deserves the reactive attitudes. (2) If we found out that no one deserves the reactive attitudes, we would have to give up the reactive attitudes. So, (3) If we found out determinism is true, we would have to give up the reactive attitudes. We have briefly considered a way in which some incompatibilists will deny (1). But what about (2)? The incompatibilist who rejects (2) may (or may not) admit that to find out that determinism is true would be to find out that no one is responsible. However, she denies that this would be to find out that we thereby have a requirement: to stop treating people as responsible – that is, to refrain from engaging in the reactive attitudes. But first we need to back up and slow down. Obviously, (2) is importantly ambiguous. To whom does “we” refer here, or to whom does “us” refer when Watson writes that “incompatibilists insist that the truth of determinism would require us to take the objective attitude [i.e., give up resentment and indignation] universally”? I thus proceed in two stages. I first consider a reading of this conditional on which the individual alone finds out that determinism is true. I then consider an interpretation on which some relevant community finds out that determinism is true. First, then, consider the following (from the perspective of an incompatibilist): (3*) If an oracle told me that determinism is true, then I would have to give up the reactive attitudes. (Let’s simply grant that the oracle is reliable, and not worry about how we know this, or how the oracle knows what he knows. The point of the oracle is simply to allow us to bypass questions about the mechanism of discovery.9) For my own part, even speaking as an incompatibilist, I am inclined to reject (3*). Below I consider several different reasons for this conclusion. But first, what is the prima facie case for (3*)? The thought appears simple: if an oracle told me that determinism is true, I would thereby come to know that no one is blameworthy (deserves the reactive attitudes), and, necessarily, it is morally wrong

222

The consequences of incompatibilism

to (knowingly) blame those who aren’t blameworthy. In other words, the thought appears to be something like: Wrong to blame the blameless. Necessarily, it is morally wrong to knowingly blame those who aren’t blameworthy. But matters are not so simple. Again, why should even an incompatibilist doubt (3*)? Several (sometimes overlapping) considerations appear relevant. (a) Inescapability. First of all, am I able to give up the reactive attitudes? Is it really possible for me, as I actually am, to cease becoming angry and resentful, at least some of the time, while still maintaining my usual day-to-day relationships – that is, whilst not hermetically isolating myself, or (worse) taking my own life? Perhaps – but this isn’t obvious. Needless to say, it is one thing to accept the thesis that no one is blameworthy; it is something entirely else to live moment by moment as if this is true, and to regulate one’s emotional life accordingly. However, it is at least very plausible that I would have to give up the reactive attitudes only if I am able to give up the reactive attitudes. Note: the ability here needn’t be anything so strong as the incompatibilist’s free-will ability; all that is required is the idea that I have to do something only if that thing is appropriately sensitive to my trying to do that thing. And it seems questionable that I could give up the reactive attitudes, even if I tried. (b) Difficulty. However, let me simply grant that I could  – with sufficient effort  – train myself in the relevant way. Even still, it should be uncontroversial that it would at least be extremely difficult for me to train myself in such a way that I  never blame those around me, to adopt what we earlier called “the objective attitude” at all times. But now the claim that I must do this seems to be seriously doubtful. In particular, the claim that I must do this now must be weighed against other claims on my time and attention – in other words, it must be weighed in terms of the gains and losses both to my own life and to the lives of those around me. Let’s imagine, for purposes of illustration, that I could train myself in the relevant way only by devoting myself to a certain program for three hours a day for three months. (Perhaps this is unrealistic, but it isn’t clear what is realistic.) But let’s also suppose that I am heart surgeon with a busy schedule, and I certainly can’t take three hours off per day for three months, at least not without alienating my confused family and friends (who would regard my quest as at best quixotic, if not wholly absurd and misguided), risking my job, jeopardizing the health of my patients, and seriously inconveniencing many others along the way. Well, must I do this, for the sake avoiding becoming resentful and indignant with people who I have come to discover do not merit that resentment and indignation? Not obviously. Indeed, this is not obvious even if I am not a heart surgeon, but instead, perhaps, a runof-the-mill academic with the merely ordinary demands on his time and attention.   The point here is a simple one: whatever effort of will I  expend on the project of reforming my basic disposition to have the reactive attitudes is effort I cannot spend elsewhere, on other things that matter – and perhaps matter much more. This point is naturally related to a further one. Suppose you had a choice about which of two projects I pursue: (a) spending 30 hours reforming my disposition to blame, so that I less frequently (or perhaps never) blame the blameless, or (b) spending 30 hours at the soup

223

Patrick Todd

kitchen, or visiting lonely elderly people at a nursing home, or cleaning the polluted river, or . . . pick your favorite cause. First, it isn’t obvious that the goods here are even commensurable, and second, supposing that they are commensurable, it isn’t clear that the goods at issue in (a) outweigh the goods at issue in (b). For my own part, for most ordinary people in the kinds of circumstances with which I am most familiar, there is a whole range of activities that I would rather see those people devote their time to, rather than the project of eliminating their dispositions to blame, even on the assumption that those they blame are never in fact blameworthy. (c) Harms and benefits. One salient consideration involved in this judgment is the question of who is harmed by my blame, if anyone is harmed at all. If my blaming someone is seriously harmful, and yet she does not deserve that harm, then perhaps there is very strong reason against my blaming that person – enough reason to make it a requirement to regulate myself in such a way that I do not do so. First, however, it seems fairly obvious that my hostile blaming attitudes towards Vladimir Putin do not harm, in any straightforward way, Vladimir Putin. Thus, if we are going to find plausible subjects who are non-trivially harmed by my blame, this will have to be subjects with whom I  have some kind of personal relationship. Consider, then, a mundane case: my wife sleeps late yet again, hurries out the door, and leaves me with a mess to clean up. In other words: mundane grist for resentment’s mill. Suppose I resent her to the appropriate degree and in the appropriate way (on the assumption that she is blameworthy). Is she harmed by this resentment, in itself? The question is a difficult one, and one I am not sure how to resolve. But it is at least not obvious that she is harmed, or harmed very substantially. Consider another case. My friend Neal promises to pick me up from the airport, and then forgets. I am naturally annoyed and angry. But then suppose I suddenly pass away from a heart attack, standing there waiting for him, in my annoyed and angered condition. Was Neal harmed by my harboring these attitudes? Not obviously. Of course, if I survive, Neal might be harmed by a decision of mine no longer to trust him, or no longer to be his friend, on account of his manifest unreliability. But then what is harming him here is not my blame – my hostile attitudes in themselves – but instead something else, something that was in any case liable to happen without the hostile attitude of resentment. But this is simply a different matter. In other words, in many cases in which it may appear that resentment is harmful, the resentment is merely epiphenomenal with respect to what actually causes the harm. By way of summing up the previous considerations, consider an analogy. Suppose that due to natural facts about me, I am deathly afraid of spiders – not venemous spiders, mind you, but the harmless variety in my vicinity. An oracle then gives me a convincing demonstration that the local spiders are in fact perfectly harmless, and thus reveals to me that my fearing those spiders is inappropriate; my fear represents them as fearsome – worthy of fear – and what I have discovered is that these spiders do not merit this attitude: they are not in fact worthy of fear at all. Does this discovery then impose on me a duty to rid myself of this fear? Well, not obviously. First, I  may not really be able to overcome the fear anyway, even knowing that it mistakenly represents its object as fearsome; second, even if I could do this, it might be very difficult, and frankly I might have better things to do. Finally, the spiders are not substantially harmed by my fearing them when they don’t deserve it. It is thus unclear why the discovery of the fact that the spiders are not fearsome imposes on me an all things considered duty to rid myself of this fear; perhaps I should just recognize this 224

The consequences of incompatibilism

unfortunate fact – that what I fear doesn’t merit it – and carry on. Mutatis mutandis, what we’ve said about fear likewise goes for blame. To support the key claim from before, we need to be told why the cases are different.10 The considerations I have been mentioning are all fairly ordinary considerations commonly thought to bear on whether and when I have a duty. Before moving on, however, I want to mention at least one decidedly more speculative source of concern about the key claim, that if I found out that no one deserves the reactive attitudes, I would have to give up the reactive attitudes. (d) Demandingness and Authority. If incompatibilism is true and yet determinism is true, then “from the start”, as it were, God, the gods, or nature has given us a set of dispositions and practices that presupposed for their fairness something that was false. But what this then suggests is that fairness (in this sense) is somehow explanatorily disconnected from what we are like; in other words, there are facts about fairness, but those facts play no explanatory role in what reality is actually, concretely like. But this then suggests that these facts lack the right kind of relationship to us in order to be fully authoritative for us. Prima facie, if someone makes a demand of us, the normative force of this demand is a function of at least two things: the authority of the one who demands, and the costs involved in complying. Roughly: the greater the authority, the more ability to impose more demanding demands; the less demanding the demand, the less authority is needed for the demand to be authoritative. Now, the demand to refrain from the reactive attitudes may be extremely demanding, and impose on me/us serious costs. But when the moral standards that make this demand are themselves shallow in the structure of the universe (cf. Mavrodes 1986), they lack substantial authority, and specifically the authority needed to authoritatively demand that we reform our basic dispositions at substantial cost to ourselves and others. This would not necessarily be to say that incompatibilism is false; rather, it is to say that the moral facts that would make it true would also make a certain kind of moral rationalism false, a moral rationalism on which the moral point of view is always, all things considered, authoritative for us. Perhaps the concern for fairness is a consideration I am not morally permitted to ignore, but am nevertheless, all things considered, permitted to ignore. Plainly, the argument here involves an enormous number of complications and further considerations, none of which I can so much as mention. But the considerations here are deep, and are almost always overlooked in discussions concerning the consequences of incompatibilism. I thus take it as highly non-obvious that, even qua resolute incompatibilist, if I find out that determinism is true, then I would have to give up the reactive attitudes – or indeed, do anything differently at all. Of course, I concede that this discovery would occasion certain attitudinal changes in me – it would perhaps prompt the thought that the human predicament is in many ways absurd, or unfair, and so on. And it may prompt an ongoing or intermittent reflective ambivalence about both my own dispositions to reactive anger, and that of those around me. But this is not to say that I would have to do anything about all of this. Those who say and write otherwise need to make their case.11 But let us now consider the alternative reading of the relevant conditional, where “we” refers to some relevant community. (3) If we found out determinism is true, we would have to give up the reactive attitudes. 225

Patrick Todd

The problems in assessing (3) are legion. Who are we to imagine “finding out” that determinism is true – and how? “We” community of philosophers? “We” educated folk who stay on top of the latest scientific discoveries? “We” politicians who determine the trajectory of our nation’s laws and customs? “We” autocrats? Or everyone – literally everyone – all at once? No matter which interpretation we pick, it seems, problems parallel to those considered previously will seem to arise: can we do this? And supposing we could, would it be all things considered good if we did, or better than other things we could be doing instead – like addressing the problems of poverty, or climate change? (And anyway, why we do have to do this for the sake of conforming to the standards of fairness, when those standards play no role explanatory role making the universe what it is like?) Fischer has asked us to imagine a scenario in which a team of physicists makes a series of discoveries, and we subsequently wake up to a New York Times headline, “Determinism is true!”12 Making certain (generous) allowances, this is indeed something we can imagine. But whereas I can imagine a team of physicists “announcing” that determinism is true – and the press being swept along for the ride – I can also imagine laughing at those physicists, in particular, at their startling combination of hubris and naivete. Fischer seems to write as if physicists could “discover” the truth of determinism in much the same way as physicists recently did discover the Higgs boson. But these cases are profoundly different; this is certainly not the place to discuss these matters, but note that it is unclear in principle what empirical discoveries could be made by physicists that would decisively establish the truth of determinism. (Indeed, we are told by some, the relevant interpretations of the physical theories are empirically equivalent.) The questions here are accordingly much more global and theoretical in nature. Further, though we can allow that we should defer to physicists regarding the nature of the laws of physics, the question whether all laws of nature reduce to laws of physics (for instance, the question whether there are any emergent powers13) is itself not a question settled by the physicists. Thus, even if physicists converge on the thesis that the laws of physics are deterministic, this leaves the question of determinism still open: for perhaps there are processes governed by laws that neither reduce to nor supervene upon those laws. Who are the physicists to tell us otherwise? The point here is that if we are to imagine suddenly gaining decisive evidence for the truth of determinism, then we are entering into the realms of science fiction.14 I have no objection to entering such realms, but we should tread cautiously when assessing what we must do in certain science fiction scenarios; the contours of these scenarios are often unclear. For instance: perhaps aliens (or AI overlords?) emerge, and perhaps, through some convincing demonstration, they show us a series of parallel universes that are physical duplicates of our universe up to the moment of their intervention – there is my counterpart, doing exactly what I did, down to the finest details! – and perhaps this forces on us the realization that any universe with our starting conditions and laws is a duplicate of our universe, which is to say: everything we have done has been an inevitable consequence of the state of the universe “at the beginning”. Well, what must we do, after this revelation? Frankly, I struggle to know how to begin to answer this question. But is it outrageous to suggest that we might collectively decide that though this new revelation did overturn our impression of ourselves as responsible for what we do and become, well, our priorities remain in place, life must go on – and so carry on as you were? I don’t know, and this is partly because it is unclear what scenario I’m even imagining, and plainly lack the space to fill in more of the details one way or another. But then my point: philosophers who write

226

The consequences of incompatibilism

that, on incompatibilism, if we came to discover that determinism is true, then we would have to give up the reactive attitudes, have substantially more work to do to show us that this is true. How are we making this discovery, who is making it, and in what circumstances? Of course, on incompatibilism, there may be some way of answering these questions such that, given those answers, the people in this scenario now (all things considered) must relinquish what reactive attitudes they may have. But it could be that the imagined people in the imagined circumstances bear very little similarity to us as we currently are, and so what these people must do given this discovery tells us nothing about what we must do given this discovery. Let us sum up. It is one thing to say that there is no world in which anyone both deserves blame and is determined, and quite another to say that all agents at determined worlds have an all-things-considered moral requirement to abandon the practices of blame. It could be true that no agent at any determined world deserves blame; it could also be that some agents at some determined worlds aren’t required to abandon their practices of blame. (And it could be that one of those deterministic worlds is our own.) At the very least, this question is an open question, and one that can’t be settled by analytic facts alone.

Conclusion My overarching point in this chapter is that we need to discuss “incompatibilism” more carefully. First, we need to be very careful before writing, “According to the incompatibilist, if determinism is true, no one is responsible.” This goes for many incompatibilists, but not all – not those libertarians we called resolute responsibility theorists (or “flip-floppers”). Second, we need to be careful not to directly present incompatibilism in terms of an alleged conditional requirement – the requirement to give up some relevant attitude or practice, conditional on determinism. On this front, let us revisit Watson’s claim from before – that incompatibilists insist that the truth of determinism would require us to take the objective attitude universally. As I have argued, this characterization won’t do, at least not without further argument. At most, we can say the following: according to the incompatibilist who is resolute, the truth of determinism would give us pro tanto reason to take the objective attitude universally. That blame is or would be undeserved or unfair is certainly a pro tanto reason against blame. But whether and in what circumstances this pro tanto reason amounts to a decisive reason is a difficult further question – and one that deserves much more attention than it has thus far been given. Let me finish this essay with one remark about how these final points bear on the longstanding, torturous dispute between the compatibilists and the incompatibilists about responsibility and determinism. My sense – which I certainly cannot here justify – is that many compatibilists are motivated by the key thought that nothing could show that we have to give all of this up. That is, nothing could show that we have to give up what certainly seem to be central features of our “form of life” – our general readiness to respond to others with reactive attitudes. But if I am right, this is a thought with which the incompatibilist can in principle agree. If what is most plausible about compatibilism is that nothing could require us to give the attitudes up, and if what is most plausible about incompatibilism is that no one who is determined is also fairly blamed – that is, if incompatibilism simply and directly seems true15 – then perhaps these are two thoughts that can ultimately be reconciled. But this is a discussion we must leave for another occasion.16

227

Patrick Todd

Notes 1 Strawson (1962). 2 This sort of “entailment-based” characterization has been standard at least since van Inwagen (1983). However, see Helen Steward (2021) for objections to this characterization, and an alternative. 3 Watson (1987), in Watson (2004: 225). Cf. Fischer & Ravizza (1998: 253), who contrast their view with the incompatibilist view as follows: “We would not have to withhold our attributions of responsibility . . . if we were convinced . . . that the universe is – surprisingly – causally deterministic.” 4 Cf. Ramsey (1931: 249) and the “Ramsey Test,” or Mellor (1993: 236): “ ‘If P, Q’ . . . expresses a disposition to infer Q from P. In other words, fully to accept a simple ‘If P, Q’ is to be disposed fully to believe Q if I fully believe P.” For a recent development and extension of a view of this kind, see Khoo (2022). Note: I am assuming that the indicative conditional is not just a material conditional “p ⊃ q” (i.e., “~p ∨ q”). The atheist of course accepts “God doesn’t exist, or there is no evil” – the atheist accepts that God doesn’t exist. 5 Van Inwagen (1983: 221). 6 Fischer (2016); Bailey and Seymour (2021); Todd and Rabern (2023). 7 Another example: dualism. Imagine characterizing dualism – i.e., consciousness and “everything is physical” incompatibilism – as follows: “The dualist contends that if we found out [e.g., from an oracle] that everything is physical, we’d have to conclude that there is no pain.” Or: “According to the dualist, if everything is physical, there is no pain.” Both characterizations are plainly ridiculous. According to the dualist, if everything is physical, she is totally mistaken about the prerequisites of pain. 8 Stalnaker (1975). For recent discussion, see Holguín (2020). 9 I don’t mean to be dismissive of such questions; indeed, there are highly non-trivial questions about how an “oracle” could know any such thing, or why we would have reason to believe this oracle, even if the oracle did know such a thing. Cf. Hawthorne’s (2002) appeal to an oracle in the context of debates about dualism. 10 Many readers will have plainly noticed that I have adopted these considerations – slightly modified – from P.F. Strawson’s “Freedom and Resentment”. Strawson therein suggests that (a) the reactive attitudes are inescapable for us, but that (b) even if we did have a choice in this matter, which we don’t, this choice could only be made in light of the “gains and losses” to human life. These arguments have universally been taken to be in the service of compatibilism, and commentators have tried – in my opinion in vain – to see in them some way towards a compatibilist result. My argument here is that these points have no tendency whatsoever to support compatibilism; what they show, if they show anything, is that incompatibilism is practically irrelevant. But the thesis that incompatibilism is practically irrelevant must be distinguished from the thesis that it is false. 11 Pereboom has written: It is plausible that to a certain degree moral resentment and indignation are beyond our power to affect, and thus even supposing that a hard incompatibilist is thoroughly committed to morality and rationality, and that she is admirably in control of her emotional life, she might nevertheless be unable to eradicate these attitudes. Thus as hard incompatibilists we might expect people to be morally resentful in certain circumstances, and we would judge it to be in an important sense beyond the agent’s control when they are. However, we also have the ability to prevent, temper, and sometimes to dispel moral resentment, and, given a belief in hard incompatibilism, we might attempt such measures for the sake of morality and rationality. (2007: 119, Cf. also 2014: 181) Yes, we might – but would we have to? 2 Fischer (2007, 2012). 1 13 Cf. Wilson (2021: Ch. 8). 14 Cf. the science fiction deterministic thought experiment described in Todd (2019). 15 For instance, on the basis of the manipulation arguments; cf. Todd (2013, 2017); Pereboom (2014: Ch 4).

228

The consequences of incompatibilism 16 For helpful comments on previous drafts of this chapter, I  wish to thank Michael McKenna, Mark Balaguer, Daniel Telech, Simon-Pierre Chevarie-Cossette, Brian Rabern, Andrew Bailey, Max Kiener, and Justin Capes.

Further reading When we ask whether “responsibility” is or is not consistent with determinism, there are different notions of responsibility we may have in mind; the focus on “reactive attitudes” is most intimately associated with Strawson (1962). Various philosophers have written or suggested that on incompatibilism, if determinism is true, we must abandon the reactive attitudes; see Watson (1987) and Fischer (2016). Bailey and Seymour (2021) is a recent reply to Fischer’s criticisms of “flipflopping”; see also Todd and Rabern forthcoming.

References Bailey, Andrew M.,  & Seymour, Amy. (2021). “In Defense of Flip-Flopping,”  Synthese, 199, 13907–13924. Fischer, John Martin. (2007). “Compatibilism,” in M. Vargas (ed.), Four View on Free Will (pp. 44–84). Malden, MA: Blackwell Press. ———. (2012). “Semicompatibilism and Its Rivals,” Journal of Ethics, 16, 117–143. ———. (2016). “Libertarianism and the Problem of Flip-flopping,” in K. Timpe & D. Speak (eds.), Free Will and Theism. Oxford: Oxford University Press. Fischer, John Martin,  & Ravizza, Mark. (1998). Responsibility and Control: A Theory of Moral Responsibility. Cambridge: Cambridge University Press. Hawthorne, John. (2002). “Advice for Physicalists,” Philosophical Studies, 109, 17–52. Holguín, Ben. (2020). “Knowledge in the Face of Conspiracy Conditionals,” Linguistics and Philosophy, 44, 737–771. Khoo, Justin. (2022). The Meaning of “If”. Oxford: Oxford University Press. Mavrodes, George. (1986). “Religion and the Queerness of Morality,” in William Wainwright  & Robert Audi (eds.), Rationality, Religious Belief, and Moral Commitment: New Essays in the Philosophy of Religion (pp. 213–226). Ithaca: Cornell University Press. Mellor, D. H. (1993). “How to Believe a Conditional,” The Journal of Philosophy, 90, 233–248. Pereboom, Derk. (2007). “Hard Incompatibilism,” in M. Vargas (ed.), Four View on Free Will. Malden, MA: Blackwell Press. ———. (2014). Free Will, Agency, and Meaning in Life. Oxford: Oxford University Press. Ramsey, Frank. (1931). The Foundations of Mathematics and Other Logical Essays. London: Kegan Paul, Trench, Trubner & Co. Stalnaker, Robert. (1975). “Indicative Conditionals,” Philosophia, 5, 269–286. Steward, Helen. (2021). “What Is Determinism? Why We Should Ditch the Entailment Definition,” in Marco Hausmann & Jörg Noller (eds.), Free Will: Historical and Analytic Perspectives (pp. 17–43). Cham: Springer Verlag. Strawson, P. F. (1962). “Freedom and Resentment,” Proceedings of the British Academy, 48, 1–25. Todd, Patrick. (2013). “Defending (a modified version of) the Zygote Argument,” Philosophical Studies, 164, 189–203. ———. (2017). “Manipulation Arguments and the Freedom to Do Otherwise,” Philosophy and Phenomenological Research, 95(2), 395–407. ———. (2019). “The Replication Argument for Incompatibilism,” Erkenntnis, 84(6), 1341–1359. Todd, Patrick,  & Rabern, Brian. (2023). “Resisting the Epistemic Argument for Compatibilism,” Philosophical Studies, 180(5), 1743–1767. Van Inwagen, Peter. (1983). An Essay on Free Will. Oxford: Oxford University Press. Watson, Gary. (1987). “Responsibility and the Limits of Evil: Variations on a Strawsonian Theme,” in F. Schoeman (ed.), Responsibility, Character and the Emotions: New Essays in Moral Psychology (pp. 256–286). Cambridge University Press. Reprinted in Watson, Agency and Answerability: Selected Essays. Oxford: Oxford University Press, 2004. Wilson, Jessica. (2021). Metaphysical Emergence. Oxford: Oxford University Press.

229

17 FREE WILL AND THE CASE FOR COMPATIBILISM Carolina Sartorio

This chapter makes a case for compatibilism. As I will understand it here, compatibilism is the idea that determinism (the assumption that everything we do is determined by the remote past and the natural laws) is compatible with free will (the kind of freedom or control relevant to our moral responsibility, and, in particular, to our being blameworthy and praiseworthy for what we do). The chapter consists of two parts. The first part briefly introduces the problem of determinism and free will and discusses the general motivations for being a compatibilist about that problem. The second part sketches two different ways of being a compatibilist (“leeway” compatibilism and “source” compatibilism) and ends with a discussion of the advantages that the second type of compatibilism has over the first type.

Part 1: Motivations for compatibilism Briefly, the problem of determinism and free will (the “classical” problem of free will) arises as follows. When we act, we don’t act in a void: what we do has prior causes. These causes are typically a combination of internal motivations (inclinations or reasons of various kinds) and external factors. Even when we act spontaneously, as when we absently touch our hair or tap our fingers on the table, we don’t regard these behaviors as uncaused, but as caused by prior factors that operate under our conscious radar. The problem arises once we realize that the causes of our acts include more than just their proximate causes: they also include the causes of those causes. For, surely, the causes of our acts also have prior causes that explain why they happened. And those causes, in turn, have other causes prior to those, which explain why they happened. And so on. Every time we act, then, there is a long chain of causes extending way back into the past – if not forever, at least back to a time before we were born. And the problem seems particularly pressing if those causes turn out to be necessitating causes – that is to say, if the existence of those causes guaranteed, given the natural laws that govern our world, that we would behave in a certain way in the present. For, if they are necessitating causes, this means that our acts had causes that are outside of our control (since they are in the remote past) that ensured that we would act in those ways: given those past causes and the natural laws,

DOI: 10.4324/9781003282242-26 230

Free will and the case for compatibilism

there is nothing else we could have done. And this includes everything that we do: our acts, choices, and deliberations (van Inwagen 1983; Ginet 1990). The thesis that our acts have remote necessitating causes is known as determinism (“causal” or “nomological” or “physical” determinism is sometimes used to distinguish it from other forms of determinism, such as logical or theological determinism; see Griffith 2013: chapter 2, and McKenna & Pereboom 2016: section 1.4). Note that the full set of causes in the remote past could potentially be extremely large – it could even be the whole state of the world at the time. Regardless of how large any such set of remote causes would have to be, the following holds: if determinism is true of our world, then, given the full state of the world at a time in the remote past (say, at the time of the big bang), which is not in our control, and given the natural laws, which are also not in our control, it follows that the world had to evolve in exactly the way it did – it couldn’t have evolved in any other way, including everything we do. And this seems to threaten our free will: how could we ever act freely, if everything we do is determined in this way? Here we are not going to get into the question of whether our world is in fact deterministic. This hinges on which laws are in fact true of our world, which is arguably an empirical question. So, in what follows I’ll assume that it’s at least an open question whether our world is deterministic and, in particular, that it’s an open question whether our actions are determined by the remote past and the laws. This is enough for the problem to arise, for, if it is indeed an open question, then this means that, for all we know, our world could be deterministic. Thus, it’s important to know whether we could still act freely if that were the case. Compatibilism about determinism and free will is the claim that determinism is compatible with free will: we could have free will, even if our acts were determined. Incompatibilism, on the other hand, is the claim that determinism is incompatible with free will: we couldn’t have free will if our acts were determined. We have already hinted at one main motivation for incompatibilism in our explanation of how the problem of determinism and free will arises (the idea that, if determinism is true, then there is only one thing we could have done at each point in time, given the past and the laws). In the remainder of this section, I’ll explain what I take some of the main motivations for compatibilism to be. I’ll give my own perspective of things, but I suspect that many other compatibilists see the issues in a similar way. To clarify: the considerations I’ll offer may not help sway those who are already convinced incompatibilists (or, more radically, skeptics about free will – those who believe we cannot have free will no matter what). Still, they can shed light on why some people are drawn to this kind of view. Also, they can help bring out the appeal of the view, in a way that might help sway other kinds of readers, such as those who are undecided about the problem of free will and those who approach the debate for the first time. The way I see it, the most natural way to arrive at a compatibilist view of free will is in three successive stages, each building on the prior one (Kane & Sartorio 2021, chapter 2):

Stage 1 At the first stage, we don’t even worry about the remote causes of our acts. We simply start by thinking about the contrast between seemingly free behaviors and seemingly unfree behaviors, or between behaviors for which people are blameworthy/praiseworthy and behaviors for which they are not at all responsible. For example, we may think about the

231

Carolina Sartorio

contrast between a compulsive hand-washer (someone who uncontrollably and irrationally washes her hands for the nth time when her hands are not dirty, and when there is no good reason for her to wash them) with a behavior that is the result of an ordinary process of rational deliberation or practical reasoning (and may be done for good reasons). By reflecting on the differences between these kinds of behaviors, we can get a rough and preliminary idea of what acting freely (and thus responsibly) might consist in, more generally. For example, we may arrive at the idea that free and unfree behaviors differ with respect to their causes: we may conclude that behaviors that are caused by mechanisms of one kind (such as compulsive mechanisms) are not free, and agents are not responsible for them, but those caused by mechanisms of another kind (such as ordinary deliberation on the basis of reasons) are free, and thus agents can be responsible for them. (On the distinction between causes that rob us of our freedom and causes that don’t, see, e.g., Ayer 1954.)

Stage 2 Next, at the second stage, we may notice that the differences we have identified between seemingly free behaviors and seemingly unfree behaviors have to do with the proximate causes of our behaviors. And, starting from those causal processes, we could then move, so to speak, “from the inside out” and inquire into the causes of those processes themselves: Where did those causes come from? Did something else cause them? If so, where did the causes of the causes come from? And so on. At this stage, the possibility of remote causes – and, in particular, deterministic ones – has been introduced. It is only then that the problem of determinism and free will arises. However, once this possibility is raised, it seems quite sensible to react by thinking: “Indeed, this seems to suggest that our acts have causes extending far back into the past – causes that could be, for all we know, deterministic.” However, on reflection, is this really that surprising? After all, doesn’t everything that happens have causes, including remote causes? Imagine, for example, that you see a leaf falling from a tree. You’ll probably think that there is an explanation of that event in terms of prior events and the laws of nature. This is a natural assumption to make, even if you might not know what that explanation is, or everything that is involved in it. Indeed, it would be odd to think of a quite ordinary natural event such as a leaf falling from a tree as lacking an explanation in terms of prior causes. In fact, in this type of case, it might even be odd to think of it as an event that lacks a full explanation in terms of prior causes: one that entails that the leaf would fall from the tree at that precise time, and in precisely the way it did (i.e., a deterministic explanation). Next, consider other natural events with more complex causes, such as a hurricane or an earthquake. As it happens, given the complexity of the mechanisms involved, scientists usually have trouble at predicting the exact path that a hurricane will take, or the occurrence of an earthquake. Still, we don’t tend to think of those events as lacking an explanation (even a full explanation) in terms of prior causes. Instead, we think of them as having an explanation whose details we aren’t aware of. These examples illustrate how natural it is to assume the existence of deterministic explanations, for at least events of this kind. Again, this is so even if we don’t know what those explanations are. For, in those cases where we don’t, we just assume that there is an explanation that escapes us. So now imagine that you believe that human acts, including mental acts such as the making of a decision, are just like those other events: they are events that take place in the 232

Free will and the case for compatibilism

natural world, and, as a result, they are subject to the same fundamental physical laws. Imagine that you believe, for example, that the event of somebody’s making a choice basically amounts to a certain kind of activity taking place in that person’s brain – patterns of neurons firing in a certain way, or something of that sort. If events of this kind are subject to the same physical laws as other events, then, if we think that other events have full explanations in terms of prior causes, we should probably think the same about human choices (even if, again, those causes may be too complex for us to understand, and to formulate reliable predictions on the basis of them). And we should probably think the same about the causes of human choices. And about the causes of the causes of human choices . . . And so on. As a result, we end up with a deterministic chain of causes that traces back to times in the remote past. This way of thinking can warm us up to compatibilism. For, as a result, instead of regarding determinism as a threat to our free will, we may be inclined to regard the existence of (possibly deterministic) remote causes as simply a consequence of a certain way of conceiving our world and our place in it: a naturalistic picture of the world and of human agency (see, e.g., McKenna & Pereboom 2016, section 2.4). According to such a naturalistic picture, an explanation in terms of prior causes is, at least in principle, just that: an explanation. It is not an excuse for our behavior, or something that exculpates us (or prevents us from getting credit) for what we do. For, intuitively, explanations are different from excuses. If determinism is true, then there is a sense in which the full state of the world in the remote past explains what you’re doing. But it explains what you’re doing in the same way that it explains everything that happens. In contrast, not everything – in fact, very little – can be explained by appeal to, say, an irresistible compulsion. That’s why the compulsion seems to be a much better excuse than the full state of the world at the time of the big bang: because, on those rare occasions where the compulsion obtains, we see it as something that abnormally constrains the behavior of individuals in ways that undermine their freedom and moral responsibility. That’s, at least, how compatibilists see things: determinism in and of itself is not something that robs us of our nature as morally responsible things.

Stage 3 Finally, there is a third and final stage that can also be part of the motivation for compatibilism. It is this. Consider, for example, the act by a non-compulsive agent – say, your friend Fred, who chooses to meet you for lunch on a certain day (freely, on the face of it). Now, let’s ask: Does Fred’s choice have causes in the remote past? There are three main possibilities: (1) It doesn’t have any such causes. (2) It has such causes, but they don’t determine Fred’s choice to meet you for lunch. (3) It has such causes, and they determine Fred’s choice. We may safely set option (1) aside because it is incredibly implausible: things don’t happen for no reason whatsoever – at least not in our world. Surely, Fred’s choice must have at least some remote causes. So that leaves (2) and (3) as the only two reasonable possibilities. As mentioned previously, we may not know (and may never know) for sure if our world is indeed deterministic; as a result, we may not know (and may never know) for sure if 233

Carolina Sartorio

(2) or (3) is true. Still, reflecting about the difference between (2) and (3) may lead to one final thought in favor of compatibilism. Here is how. The difference between (2) and (3) is just the difference between deterministic and indeterministic causes. If (2) is true, then the remote causes of Fred’s choice aren’t deterministic, because they didn’t guarantee that he would make that choice. This means that, given those causes, the world could have evolved in more than one way: it could have evolved in a way that resulted in Fred’s choosing to meet you for lunch (as in the actual case), or it could have evolved in a way that resulted in Fred’s not making that choice. Either could have happened, given the same combination of past and laws. Which of the two happened was genuinely undetermined, or random. Thus, this wasn’t up to Fred. (Or to anybody else, for that matter, but what’s important for our purposes here is that it wasn’t up to Fred.) In other words, indeterminism introduces an element of randomness or chance that escapes agential control (van Inwagen 1983; Mele 2006). If an event is truly undetermined, there is no “full” explanation of why it happened. At most, there is an explanation in terms of probabilities: given the prior state of the world, there was some probability that the world would evolve in that way, but also some probability that it wouldn’t, and there is no deeper explanation than this. We can think of this in terms of “possible worlds” or different ways in which the world could have evolved: given the same past and laws, we could have ended up in a world like the actual one (where Fred makes the choice to meet you for lunch), or in a world different from the actual one (where he doesn’t). Both kinds of worlds were consistent with the prior state of our world, and Fred doesn’t get to decide which world we live in. As a result, this may naturally lead you to think that, in circumstances like this, Fred doesn’t have more control than if his choice had been determined. For the existence of indeterminism doesn’t give him the capacity to “settle” which way things were going to go, or which choice he would end up making. Things just happened the way they did, and that’s the end of the story; there is no deeper explanation. And, if this is right, then it suggests that the falsehood of determinism (or the truth of indeterminism) wouldn’t help our freedom in any way. For it wouldn’t make us any more free, or any more in control of our acts, than if our acts had been causally determined. All that would change is that there would be some “openness” in how the world could evolve from one time to another. But given that it wouldn’t be up to us to decide which way it does evolve, intuitively this is not something that could give us any more control over what happens. In sum, although we may not know if (2) or (3) is true, perhaps we may know this: the truth of (2) wouldn’t make us more in control than the truth of (3). Thus, the truth of determinism wouldn’t deprive us of any significant form of control over our acts, one that we could only have if determinism were false. As a result, this may lead us to think that determinism is not a threat to our free will and may move us even closer to compatibilism.

Part 2: Two ways of being a compatibilist: leeway and source compatibilism In this section I discuss two different ways of being a compatibilist. The two ways correspond to two rival ways of understanding free will (or, again, the type of freedom or control required for responsibility). According to the first conception, leeway freedom, free will requires leeway: it requires options that we could have chosen from (“alternative possibilities” of action, or the ability to do otherwise). According to the second conception, source 234

Free will and the case for compatibilism

freedom, free will doesn’t require any such thing; instead, on this view, all that matters is the actual sources, motivations, or causes of our behavior. When coupled with compatibilism, leeway freedom and source freedom give rise to two varieties of compatibilism: leeway compatibilism and source compatibilism (the same goes for incompatibilism, as the distinctions crosscut each other, but I won’t focus on this here because what we’re interested in now are varieties of compatibilism, not of incompatibilism). Leeway compatibilism is the view that determinism is compatible with free will because it’s compatible with leeway freedom, and source compatibilism is the view that determinism is compatible with free will because it’s compatible with source freedom. For examples of leeway compatibilism, see Ayer 1954; Lewis 1981; Vihvelin 2013. And, for examples of source compatibilism, see Frankfurt 1971; Fischer 1994; Fischer & Ravizza 1998; McKenna 2013; Sartorio 2016, forthcoming. What are the motivations for embracing each conception of free will? I’ll discuss leeway freedom first, and then source freedom. The motivation for leeway freedom is quite straightforward, for it seems very natural to think of freedom in terms of having a choice among different options or alternatives. The leeway model can be visualized as a “branching paths” model, where the branching paths represent different options available to agents. And it can be motivated by the idea that, if there’s just one single path, or only one thing that you could have done at a time, then it seems that you’re being forced into that single path, and thus that you can’t act freely. How could you act freely, if that was the only path available to you? To illustrate, imagine that your friend Ed is a compulsive hand-washer: Ed washes his hands driven by an irresistible compulsion, even when he recognizes that he has no good reason to wash them. Ed doesn’t act freely when he washes her hands. Intuitively, the reason he doesn’t act freely seems to be that he couldn’t have helped it (he had no alternatives; there is only one thing he could have done at the time). In contrast, people who are not compulsive hand-washers can refrain from washing their hands (they have a different path open to them). This arguably explains why they act freely when they wash their hands. Thanks to its intuitive appeal, the leeway model of free will was widely embraced (by both compatibilists and incompatibilists about the determinism and free will problem) up until the philosopher Harry Frankfurt cast serious doubt on it in a groundbreaking article (Frankfurt 1969). Frankfurt argued that the intuitive appeal of the leeway model is just an illusion that results in a confusion (an understandable confusion, perhaps, but a confusion nevertheless). The illusion arises from the fact that, in typical cases where we can’t do otherwise, we also don’t act freely. But this is only because the causes that bring about our behavior in those typical cases are also the factors that make the behavior inevitable (in the case of Ed, the compulsive hand-washer, for example, the compulsion is both what causes Ed’s behavior and what makes it inevitable). As a result, our behavior ends up having the wrong kinds of causes in those cases. According to Frankfurt, that results in the illusion that the branching paths are needed for free will, when in fact they are not. To disarm the illusion, Frankfurt notes that the factors that explain a behavior and those that render it inevitable can conceivably come apart from each other: although usually they tend to go hand in hand, we can imagine scenarios where they don’t. And those scenarios can help us uncover the truth about the matter, which is that we don’t really need alternative possibilities to have free will. Of course, those scenarios are necessarily going to be quite artificial since, again, in typical or normal cases where we can’t do otherwise, we tend 235

Carolina Sartorio

to not act freely. Those artificial scenarios, designed to show what really matters to free will, have been called “Frankfurt-style” cases. Here is an example of a Frankfurt-style case. Imagine, now, that a non-compulsive agent (imagine it’s Fred again) decides to wash his hands because they’re dirty. The process leading to that decision is completely ordinary – in particular, it is based on good reasons and not on a compulsive desire. However, imagine that, as it turns out, Fred couldn’t have made any other decision. For a sneaky and resourceful (and very clean) neuroscientist would have made sure that he made that same decision if he hadn’t made it on his own. Imagine, for example, that the neuroscientist had been secretly monitoring Fred’s thoughts and could reliably predict what Fred was going to do by reading some of his brain patterns. Imagine that, if the neuroscientist could predict that Fred wasn’t going to make the decision to wash his hands on his own, he would have intervened by making certain neurons in Fred’s brain fire, which would have reliably resulted in Fred’s forming the intention to wash his hands. In this case, Fred lacks the ability to make any other choice. Still, given that the neuroscientist never intervened, and Fred does it for his own reasons, he seems to wash his hands freely, despite not being able to do otherwise. If so, this means that acting freely doesn’t require the ability to do otherwise. At the same time, it motivates the idea that all that matters to freedom is the actual causes or sources of our behavior. This is, again, the source freedom idea. When I described leeway freedom, I noted that the basic idea behind that model is very intuitive. But it’s important to see that the basic source idea is also quite intuitive. For, as Frankfurt pointed out, it’s very natural to think that factors that play no role whatsoever in the explanation of an agent’s behavior are simply irrelevant to the agent’s freedom. Again, a Frankfurt-style case illustrates this nicely: the presence of the neuroscientist seems to be irrelevant to the agent’s freedom, given that he never intervenes, and thus is not part of the explanation of the agent’s behavior. The fact that we tend to think like this is particularly evident in cases where agents act in morally significant ways. Imagine, for example, that Fred now decides to cheat on his taxes, and makes that decision completely on his own, on the basis of his own selfish reasons. Intuitively, Fred acts freely and is blameworthy for what he does, even if, unbeknownst to him, the neuroscientist had been spying on him and wouldn’t have let him do anything else. Fred seems just as blameworthy as if he would have been if the neuroscientist hadn’t been present. Again, this motivates the source freedom idea. Now that we have introduced the two rival conceptions of freedom, let’s turn to comparing them. Obviously, here I cannot offer a full analysis of their relative merits and demerits. So, instead, I’ll focus on what I see as the main advantages of the source conception (my preferred view of free will) over the leeway conception. I’ll start by drawing attention to one advantage that is particularly relevant to the main topic of this chapter: compatibilism. It’s the fact that it’s a lot easier to be a compatibilist about the problem of determinism and free will if you’re a source compatibilist than if you’re a leeway compatibilist. This is because, if you’re a source compatibilist, and thus you believe that freedom doesn’t require the ability to do otherwise, you don’t then need to worry about whether the truth of determinism would rule out the ability to do otherwise: even if determinism ruled out the ability to do otherwise, we could still act freely in a deterministic world. In contrast, if you’re a leeway compatibilist, you need to be able to defend the claim that determinism doesn’t rule out the ability to do otherwise. And you would have to defend 236

Free will and the case for compatibilism

this claim against important arguments to the contrary (most notably, the “consequence argument” put forth, among others, by van Inwagen 1983; Ginet 1990). In other words, you need to be able to claim that we can still act freely in a deterministic world because, in the relevant sense, we can retain the ability to do otherwise even if our acts are determined. Several compatibilists have attempted this, including “classical compatibilists” (such as Ayer 1954) as well as the “new dispositionalists” (such as Vihvelin 2013). But the point is simply that it’s an extra burden to take on. This is a well-known advantage of the source view of freedom, one that source theorists have emphasized in their work (see, for example, the defense of this view in Fischer & Ravizza 1998, which exploits this fact explicitly). But there are other advantages of the source view of freedom that tend to be overlooked in the free will literature. I’ll focus on one main such advantage. It is this: the source view stands out as the simplest or most elegant conception of free will. And this means that, unless given good reason to think otherwise, we should embrace the source view over the leeway view. I realize that this claim flies in the face of tradition (the leeway model is typically regarded in the literature as the “classical” or “traditional” view of freedom). So, let me explain why I think this. The reason is that the source model appeals to factors that everybody already agrees, or should agree, are relevant to our free will. These factors are actual causes or actual explanations. As a result, an account in terms of just actual causes is simpler, and more ecumenical or neutral, than an account that (also) appeals to alternative possibilities. Why do I claim that everybody accepts, or should accept, the relevance of actual causes? Because actual causes are, clearly, at least part of what matters: why we do what we do obviously matters to whether we act freely and whether we are morally responsible (blameworthy or praiseworthy) for what we do. For example, it clearly matters if you act as a result of an ordinary process of deliberation, or if you act from an irresistible compulsion. It also clearly matters if our acts have any causes at all: if one day you discover that you’ve been acting on the basis of intentions formed completely at random, or purely as a result of some indeterministic quantum processes that bypass your rational capacities, then you’ll probably think that you weren’t in control of any of those actions and thus you can’t be held accountable (blameworthy or praiseworthy) for what you do. Control is, at least partly, a causal notion: a notion that has to do with actual causes. Everybody should agree with that much. Notably, those who embrace the leeway conception of freedom typically do agree with that much. For they in fact tend to understand free will as a dual form of control, one that requires both actual and counterfactual control (see, e.g., van Inwagen 1983: 8). In particular, the leeway idea is typically expressed in terms of a principle, the “Principle of Alternative Possibilities,” which is formulated as only a necessary condition on free will: as the claim that we can act freely or responsibly only if we’re able to do otherwise (see Frankfurt 1969 for the classical formulation). Thus, the leeway view is perfectly consistent with the claim that actual causes are also relevant, and many leeway theorists in fact believe that they are. Now, imagine that we could explain everything that needs to be explained with just actual causes, without appealing to anything else (including alternative possibilities). Then that’s arguably what we should do. In other words, the default view of free will is one in terms of just actual causes, not one in terms of alternative possibilities. This turns the tables on proponents of the leeway model. For many seem to think that, because of the intuitive appeal of the Principle of Alternative Possibilities, it’s on critics of 237

Carolina Sartorio

that principle to show that the leeway model should be abandoned in favor of the source model (hence the lively debate about Frankfurt-style cases). But, on the basis of the previous discussion, I dare say it’s the other way around: once it becomes clear that we may be able to explain everything in need of explanation just by appealing to actual causes, it’s on leeway theorists to show that we also need alternative possibilities. Thus, we are justified in embracing the source freedom view, at least as a starting point.

Acknowledgements I am grateful to Maximilian Kiener for inviting me to contributing to this volume as well as for providing helpful feedback on an earlier version of this chapter. Thanks also to the participants at the October 2022 Routledge series of workshops where drafts of the papers for this volume were presented.

Further reading M. McKenna and D. Pereboom’s Free Will: A  Contemporary Introduction (Routledge, 2016) is a helpful introduction to the free will problem, and it includes a discussion of the main contemporary compatibilist views, including the main leeway and source views. R. Kane and C. Sartorio’s Do We Have Free Will? A Debate (Little Debates about Big Questions, Routledge, 2021) (written in the form of a debate between a compatibilist and an incompatibilist) includes a more in-depth discussion of the motivations for both compatibilism and incompatibilism. For more on source v. leeway approaches, see K. Timpe’s “Leeway vs. Sourcehood Conceptions of Free Will” (in K. Timpe, M. Griffith, and N. Levy (eds.), Routledge Companion to Free Will, Routledge, 2017, chapter  19). For more on Frankfurt-style cases, see C. Sartorio’s “Frankfurt-Style Examples (in Timpe, Griffith, and Levy, op. cit., chapter 16).

References Ayer, A. J. (1954). “Freedom and Necessity,” in Philosophical Essays (pp.  271–284). London: Macmillan. Fischer, J. (1994). The Metaphysics of Free Will: An Essay on Control. Oxford: Blackwell. Fischer, J., & Ravizza, M. (1998). Responsibility and Control: An Essay on Moral Responsibility. Cambridge: Cambridge University Press. Frankfurt, H. (1969). “Alternate Possibilities and Moral Responsibility,” Journal of Philosophy, 66(23), 829–839. ———. (1971). “Freedom of the Will and the Concept of a Person,” Journal of Philosophy, 68, 5–20. Ginet, C. (1990). On Action. Cambridge: Cambridge University Press. Griffith, M. (2013). Free Will: The Basics. Abingdon: Routledge. Kane, R., & Sartorio, C. (2021). Do We Have Free Will? A Debate. New York: Routledge. Lewis, D. (1981). “Are We Free to Break the Laws?” Theoria, 47, 113–121. McKenna, M. (2013). “Reasons-Responsiveness, Agents and Mechanisms,” in D. Shoemaker (ed.), Oxford Studies in Agency and Responsibility, vol. 1 (pp. 151–183). Oxford: Oxford University Press. McKenna, M.,  & Pereboom, D. (2016). Free Will: A  Contemporary Introduction. New York: Routledge. Mele, A. (2006). Free Will and Luck. New York: Oxford University Press. Sartorio, C. (2016). Causation and Free Will. Oxford: Oxford University Press. ———. (Forthcoming). Causalism: Unifying Action and Free Action. Oxford: Oxford University Press. van Inwagen, P. (1983). An Essay on Free Will. New York: Oxford University Press. Vihvelin, K. (2013). Causes, Laws, and Free Will: Why Determinism Doesn’t Matter. New York: Oxford University Press.

238

18 DELIBERATION AND THE POSSIBILITY OF SKEPTICISM* Simon-Pierre Chevarie-Cossette

1 Introduction Some views are too far-fetched to be true. This might for instance include cultural relativism, the view that every moral truth is relative to one’s group; and external-world skepticism, the view that knowledge of the external world is impossible to attain. Philosophers still attempt to find decisive refutations of these views, which would show that they are incoherent or groundless. Sometimes, however, the far-fetched views are coherent and supported by arguments. In these cases, philosophers might adopt a more dogmatic attitude to them. They might say, I know that I have hands. If I know that I have hands, I know there is an external world. So, I know there is an external world. . . . and so there must be a mistake in skeptical arguments.1 or, Genital mutilation is wrong, whatever others might think. If something is wrong whatever others might think, cultural relativism is false. So, cultural relativism is false. . . . and so there must be a mistake in relativist arguments.2 These responses are dogmatic since they are clearly not intended at convincing the other side. In fact, it would be unreasonable for a skeptic or a relativist to be moved by them, unless they had previously misunderstood the consequences of their views. These responses are also dogmatic insofar as their proponent refuse to seriously entertain the possibility that relativism or skepticism be true: she who would entertain this possibility would also (temporarily) suspend her belief about her hands or about mutilation. We’d rather refute a view than take a dogmatic line against it. But sometimes at the end of the day (that is, after having shown why the view we are targeting rests on resistible 239

DOI: 10.4324/9781003282242-27

Simon-Pierre Chevarie-Cossette

arguments), we must be dogmatic in this way. If we don’t, we run the risk of leading two lives (like Hume the naturalist and Hume the skeptic): one practical life where we rely on our knowledge and moral beliefs that we take to be objective; and one theoretical life where we refuse to reject any view that we cannot prove to be false.3 Still, it’s a cautious policy to keep the dogmatic response for truly far-fetched philosophical views. It is hard to say precisely what they are, but we can give a tentative answer. A  view might be truly far-fetched because it completely runs contrary to common sense or observation. But a view might also be far-fetched because it cannot be stably adhered to; we might have to presume that it is false when engaging in valuable and inescapable human activities.4 It seems that one cannot, for instance, be an external-world skeptic while deliberating about what to do next: when we deliberate, we try to make the best of what we know. Similarly, it seems impossible to adhere to cultural relativism while deliberating about moral issues with people who don’t share our values: by the relativist’s lights, they are not part of our community, and so are governed by different moral truths.5 This chapter concerns a different form of skepticism, but one that may have similar problems, namely responsibility septicism. This is the view that we cannot be morally responsible (henceforth “responsible”) for our conduct. Accordingly, holding others responsible, for instance by punishing them, is always unfitting (but perhaps not unjustified since it might have great benefits). For the responsibility skeptic, it is also unfitting to display reactive attitudes such as resentment and guilt. Responsibility skepticism – by contrast with external-world skepticism, and not unlike cultural relativism – has been defended by several contemporary philosophers and popular figures. It is endorsed for many different reasons, which we should not confuse. Some believe that there are no moral properties or relations and so no moral responsibility. Others think that there are no persons and so that there is no one to be responsible for anything. But most serious versions of responsibility skepticism – the ones it would be less appropriate to dogmatically reject – are premised on doubts about free will, the power to choose freely.6 Our question is then the following. Given that responsibility skepticism is a coherent doctrine, can it guide the life of human beings in their numerous activities such as moral engagement and deliberation? Differently put, is the truth of the skeptical doctrine not just a live possibility, but one that we can live with? If not, we might expect the non-skeptic to say something like: The Myanmar Junta is responsible for atrocities against civilians. If the Myanmar Junta is morally responsible for atrocities against civilians, responsibility exists. So, responsibility exists.7 . . . and so there must be a mistake in skeptical arguments. That is, if the responsibility skeptic cannot show that her doctrine is livable, she can expect a dogmatic response. How could the skeptic protest? She cannot simply complain that this is dogmatic, for, as we have said, it is reasonable to be dogmatic against far-fetched views. To lodge a complaint against dogmatism, the skeptic must show that her doctrine is not far-fetched, that it can be stably adhered to. She does not need to establish that she can be an engaged skeptic at every instant, but it won’t be enough to show that she can be a skeptic in the classroom.8 As things stand, most prominent skeptics have claimed that their doctrine is livable. For instance, over the years, the Derk Pereboom has argued that we could “live without free will”, by explaining in remarkable detail how to excise responsibility from our morality 240

Deliberation and the possibility of skepticism

system without damaging it, how to replace some of our emotional reactions like blame while preserving our interpersonal relationships, and how to reform our criminal justice system on the model of quarantine. This is the kind of skeptic that I have in mind in this paper. Our question has an important psychological aspect – could we be happy skeptics? – which I shall not attend to here. But it has other aspects that require the attention of the philosopher of action. Just like the external-world skeptic and the cultural relativist, the responsibility skeptic has a difficult challenge to meet which concerns deliberation. I turn to it now.

2  Free will skepticism and practical deliberation There is something strange with someone who claims, at once, that she can never decide freely and yet who engages extensively in decision-making. There seems to be a tension between skepticism and deliberation. But what skepticism exactly are we talking about? Responsibility skepticism comes, as we have briefly seen, in different flavours. The view that concerns us in this chapter says that we are not responsible because we lack the power to choose freely, namely free will. Responsibility skepticism per se is not in tension with deliberation. A child might, without any incoherence, ask herself what to do even if she realizes that she cannot be held responsible. Instead, deliberation is in tension with free will skepticism. More specifically, deliberation is in tension with the view that we don’t have true alternatives or options – that we lack, in philosophical speak, the ability to do otherwise or leeway.9 The skeptic might believe that she has no leeway on the grounds that determinism is true, but determinism is not the crux of the matter.10 What matters is that our skeptic thinks that she always has at most one option and yet that she deliberates as if she had several. What sort of deliberation is problematic for the free will skeptic? The literature on the question carefully distinguishes two kinds.11 Theoretical deliberation is the activity of trying to figure out the answer to a factual question, by reasoning (rather than by guessing or trying to recall). At the end of successful theoretical deliberation, we identify an answer to our question and form a belief about it. Practical deliberation is the activity of trying to figure out what to do, by reasoning. At the end of successful practical deliberation, we identify a course of action and make a decision to act. Consider some difficult cases, to clarify the distinction. To decide what to do next (including what to investigate next) is practical, whereas to try to predict what one will do next is theoretical. To decide whether to write down an answer on an exam sheet is practical, whereas to wonder what answer is the right one is theoretical. These examples show that practical and theoretical deliberations are mutually dependent activities: to determine what to do, we must answer some factual questions; and often in the course of answering a factual question, we deliberate about what to investigate first.12 The present challenge concerns practical deliberation, not theoretical deliberation. There is no tension in believing that whatever we do is unavoidable and still trying to figure out the truth about a subject matter. Suppose that we ask whether people have duties to themselves, or whether seaside towns in Britain have a bright future, or whether modus ponens is a valid rule of inference. To deliberate about such matters, we don’t need to presume that the answers that we consider are open to us – that each is believable. We might know that, as soon as we understand our options, we will see that all are incoherent but one. By contrast, when we deliberate about what to do, it seems that we must presume that each option that we consider is open to us. If, on my way to the cinema, I deliberate about whether to go see The Fabelmans or a retrospective of Blade Runner, I cannot at the same time believe that one 241

Simon-Pierre Chevarie-Cossette

of them is not showing. I can at best deliberate (theoretically) about which I should or would go see if they were both available13; I can deliberate about which of the movies is the best. Before we assess whether skeptics can engage in practical deliberation, we must consider an objection. What if all our deliberation consists in trying to figure out the answer to a factual question? When I ask whether to go see Blade Runner or The Fabelmans, I do ask myself questions such as: “which is more entertaining?” “do I  feel like being thrilled or touched?” etc. But perhaps that is all I do. Or rather, as Aristotle seems to have thought, maybe all I ask when I conduct practical deliberation is what is the best means to reach my end? But if that is the case, practical deliberation is just a subset of what I have called “theoretical deliberation”: practical deliberation ends with a belief about what the best means to my end is. There is, in a word, no deep distinction between practical and theoretical deliberation; since skeptics have no problem with theoretical deliberation, they have no problem with practical deliberation either. So goes the objection.14 Surely, it might be replied, there are moments where theoretical deliberation is over – I have found the best means to my end – and yet I deliberate. I am, as we might say, on the fence, trying to decide. So, practical deliberation cannot be reduced to theoretical deliberation; it must include an executive part. As J. L. Austin put it, “ways and means are a matter for the planning staff; decision is a matter for the commander” (1979: 286).15 I am no longer convinced that this answer decisively establishes that practical deliberation is a distinct activity. For, when I’m on the fence, I might simply (1) be rerunning the theoretical deliberation in the hope that I find a better answer; or (2) display indecisiveness, which is not a rational activity worthy of the name “practical deliberation.” True, one can deliberate about what to do without arriving at a decision – only the “planning staff” does its job – but it does not follow that there is a kind of deliberation, distinct from theoretical deliberation, whose aim is to form a decision. I think the right answer to the objection is instead that even if practical deliberation consisted in identifying the best means to our end, it would be closed to the skeptic. This is because to identify the best means to our end, we must restrict these means that we consider to those available to us – or to those that we believe to be such. Otherwise, it seems that we are engaged in a different activity. Suppose, for instance, that my aim is to reach Australia from Europe next week and that I deliberate about the best means to get there. If I seriously consider an option like “flying a military jet,” I am not truly engaging in practical deliberation. The assertion “the best way to get to Australia is to fly a military jet” might be true, but it is irrelevant to me. Inaccessible means to reach my end should not be part of my practical deliberation, although they might be part of other activities, such as daydreaming or what-iffing – deliberation about what I should do if things were different (for instance, if I were a member of the Royal Airforce). So, it seems that whatever conception of practical deliberation we accept,16 we must presume that our options – or the means to our end that we consider – are open to us.17 (To follow the literature on the question, in what follows I shall stick to the view that practical deliberation is the activity of deciding what to do). A last comment before we assess whether the skeptic can meet the challenge. She might of course insist that her view is true but that she can’t adhere to it while she deliberates.18 But remember: that is not the kind of skeptic that we are focused on; nor is it the kind of skeptic that is so influential in the free will literature. Our skeptic is a highly practical person who exhorts us to reform our ways by enacting her doctrine: we must abandon some attitudes (blame) and practices (punishment). Our skeptic carefully tries to save a livable system of morality that is compatible with the excision of responsibility from it.19 She 242

Deliberation and the possibility of skepticism

would not suggest either that we eliminate practical deliberation and just rely on old habits since she wants us to change many of these habits.20

3  Can the skeptic deliberate qua skeptic? Having specified the sort of skeptic and deliberation that are in tension, let us examine this tension. Can skeptics deliberate? Of course they can, for they do. Can skeptics deliberate rationally? Again, they can, for they can abandon their doctrine (perhaps irrationally), and then deliberate rationally.21 What is less clear is whether skeptics can deliberate rationally while holding on to their doctrine. In a word, can skeptics deliberate qua skeptics? To make up our mind, we’ll need to consider two cases. Take: 1: I want to go see The Fabelmans or Blade Runner. Yet, I know that Blade Runner is not showing.

cinema

In this situation, I can deliberate about which film I should see if they were both showing; or I could deliberate about which film is the best. But I cannot deliberate about which film to go see (at least not rationally). Why? One simple explanation is: To deliberate about whether to Φ or Ψ (where Φ and Ψ are incompatible options), one must presume that Φing is an available option and that Ψing is an available option.

deliberative openness:

It seems quite natural to say in fact that if I deliberate about two options, I must presume, though not necessarily believe,22 that each is open. Yet, if Deliberative Openness is true, skeptics cannot deliberate qua skeptics – they must (silently) renege their doctrine whenever they deliberate. For their view entails that, for any two incompatible options, one is unavailable. When they deliberate, they then presume the truth a proposition which is incoherent with their doctrine. The skeptic must then find an alternative principle, one that explains why deliberation is ruled out in cinema 1, but not in the skeptic’s life. She might accept: To deliberate about whether to Φ or Ψ, one’s evidence must not strongly suggest that Φing is unavailable and one’s evidence must not strongly suggest that Ψing is unavailable.23

deliberative epistemic openness:

This principle explains why deliberation is ruled out in cinema 1 (I know that Blade Runner is not showing), but not in the skeptic’s life (the skeptic typically does not know the result of her deliberation in advance24). How should we decide between our two explanatory principles? We need to consider a second case: 2: I want to go see The Fabelmans or Blade Runner. I know that one of them is not showing but I don’t know which.

cinema

Again, I cannot deliberate about which film to go see. Yet, if I deliberate, I do not infringe deliberative epistemic openness: my evidence does not suggest that The Fabelmans is 243

Simon-Pierre Chevarie-Cossette

not showing, and it does not suggest that Blade Runner is not showing.25 By contrast, if I deliberate, I infringe deliberative openness. I can’t presume that each film is showing and yet I deliberate. A case like cinema 2 therefore shows the superiority of deliberative openness over its epistemicised counterpart. It then makes a strong case for the view that skeptics cannot deliberate qua skeptics.26 The skeptic must explain what sets her apart from me when I deliberate in cinema 2. She might insist that her deliberation is, contrary to mine, efficacious. I know that my deliberation is not efficacious because I know that whatever I decide, I’ll end up going to see the movie that is in fact showing. By contrast, the skeptic will execute whichever option she opts for. That is, if she decides to Φ, she will Φ; and if she decides to Ψ, she will Ψ. Thus, the skeptic has a substitute for deliberative openness: To deliberate about whether to Φ or Ψ, one must presume that the deliberation is efficacious, that is if one decided to Φ, one would Φ and if one decided to Ψ, one would Ψ.27

deliberative efficacy:

Of course, both the movie that I will go see and the option that the skeptic will decide on (if her doctrine is right) are already settled before the deliberation. But that’s the end of the analogy, says the skeptic. That some event is settled does not rule out the efficacy of a process causing it. That the extinction of dinosaurs was already settled during the big bang did not preclude the collision of the meteorite from being efficacious. That the skeptic’s decision was already settled during the big bang did not preclude her deliberation from being efficacious. But that Blade Runner was not showing did preclude my deliberation from being efficacious in the cinema cases.

4  Skeptics cannot deliberate qua skeptics The deliberative efficacy response might seem decisive: it seems to allow skeptics to identify cases of bad deliberation without inculpating themselves. Yet I believe the response fails. Consider first some problematic cases for deliberative efficacy. The principle is too strong – it wrongly excludes some fine deliberation. Suppose that while playing golf, I deliberate about whether to hole it from the left or from the right. Typically, I’d presume that I can do each. To be a coherent skeptic, the suggestion goes, I should instead presume that if I decided to hole it from the right, I’d do it, and the same for the left. But that’s too strong a requirement. Even if I am a good golf player, I might think that there is a high chance that I fail. It might be precisely because I don’t presume that I will succeed that I don’t engage in betting, for instance. One28 might object that it would be irrational to deliberate in this way: if I can’t presume that I will hole the ball from the right if I decide to, I should only deliberate about whether to hit it from the right. But what if I am trying to hit it so hard that I can’t even presume that if I decide to hit the ball from the right I will, since I will probably miss the ball? Presumably, the objector would say that I should only deliberate about whether to try to hit the ball from the right. This seems too strict. But regardless, the problem will reappear in cases where one doubts whether one will try. I can clearly deliberate about whether to call my mother tomorrow, even if – forgetful as I am – I can’t even presume that if I decide to, I will try. A similar problem applies to predicted weakness of will. If I deliberate about whether to stop smoking, so long as I presume that I can stop (that I have the ability and opportunity 244

Deliberation and the possibility of skepticism

to stop), my deliberation can be rational. I should not have to presume that were I to decide to stop smoking, I would. deliberative efficacy is also too weak. Suppose that I deliberate about whether to eat some red sugar balls, but that I have a phobia (of which I am aware) about everything that conjures the image of red cells. I know I cannot decide to eat a sugar ball, and yet I deliberate about whether to eat it or not. My deliberation respects deliberative efficacy (since I presume that if I decided to eat a sugar ball I would), and yet it seems irrational.29 The skeptic needs to integrate in deliberative efficacy the requirement that the deliberator presume that she can decide to Φ and that she can decide to Ψ. Of course, this precise formulation is not available to the skeptic, since skepticism implies that we can never do, and a fortiori decide to do, otherwise. Finding the right formulation is a tall order. We don’t want to say that a deliberator must presume that if she decided to decide, she would. That’s absurd: we very rarely decide to decide. We don’t want to say that she must presume that if she found a conclusive reason in favor of Φ, she would decide to Φ, for this does not take care of the numerous cases where one decides to Φ in the absence of a decisive reason. At this point, the reader might have a strong sensation of déjà vu. That’s because the previous cases are exactly the ones we find in a different debate in the free will literature, namely the debate about whether we can reduce “cans” to “ifs”. The idea is simple. Some philosophers (such as Moore) have claimed that to say that I can do A just is to say that if I want to do A, I will do A. If this conditional analysis of “can” is correct (or something close to it), then free will is saved. For, often, had I chosen to do otherwise, I would have. That’s uncontroversial, whether determinism is true or not. And so, if the conditional analysis is correct, I often could have done otherwise, pace skepticism. Now, skeptics and non-skeptics have challenged the conditional analysis of “can” with the cases that we saw previously. The conditional analysis of “can” is too strong: “Consider the case where I miss a very short putt and kick myself because I could have holed it. It is not that I should have holed it if I had tried: I did try, and missed” (Austin 1979: 218). And the conditional analysis is too weak: Suppose that I am offered a bowl of candy and in the bowl are small round red sugar balls. I do not choose to take one of the red sugar balls because I have a pathological aversion to such candy. . . . [I]f I had chosen to take the red sugar ball, I would have taken one, but, not so choosing, I am utterly unable to touch one. (Lehrer 1968: 32) It is because of examples like these that the conditional analysis has often been taken to fail. If the conditional analysis is instead successful (or if it is repaired30), then we can sometimes do otherwise: free will skepticism is therefore false. So, it is important for the free will skeptic that the conditional analysis fails. And yet, for the skeptic to be able to secure deliberative efficacy, it seems that, in effect, the analysis must succeed; a response to problematic cases for deliberative efficacy provides a response to counterexamples to the conditional analysis and vice versa. Suppose, for instance, that we try to immune deliberative efficacy from the golf example by asking that a deliberator presume that she would Φ in several close possible worlds where she decides to Φ. This amendment will immune the conditional analysis from the golf example as well: the fact that Austin missed his putt is compatible with the fact that he would succeed in several close possible worlds where he decides to putt. 245

Simon-Pierre Chevarie-Cossette

In a word, the skeptic is trapped. Either we can capture “can” with conditionals or we cannot. If we can, then free will skepticism is false (and responsibility skepticism is groundless): the truth of the conditional analysis rules out skepticism. If we cannot capture “can” with conditionals, then the skeptic loses deliberative efficacy: the cases that threaten the analysis threaten the principle. But this means that the skeptic has no alternative to deliberative openness and must concede that she cannot deliberate qua skeptic. We should not be surprised. We should have expected that the concept of “can” involved in free will (the power to choose) be the same as the one involved in deliberation (the activity of deciding). Our skeptic might look for alternatives to deliberative efficacy, but this will be difficult.31 Take for instance the principle that one must rule out courses of action that, in all likelihood, won’t be pursued. This rules out some cases of good deliberation. Suppose that Mike is akratic about bribery; he has accepted bribes his whole career as a police officer. So when he once again deliberates about whether to wet his beak, he thinks with good reason that he is unlikely to do the right thing. But suppose he ends up refusing a bribe. According to the considered principle, Mike’s deliberation was problematic. This is wrong. Since Mike can refuse the bribe, refusing it can feature in the list of options to consider. That’s what the much better deliberative openness tells us. If our skeptic cannot find a way out (for instance, by changing the reason why she rejects responsibility), she might insist that her view is true even if it is impossible to maintain when deliberating. Or she might argue that the irrationality involved in her deliberation is unimportant.32 Both these responses make skepticism less appealing as a doctrine. And they make it more reasonable to take a dogmatic line against skepticism.

5 Conclusion Is responsibility skepticism a livable doctrine? The skeptic must meet several challenges before we can deliver a positive answer. One says that if we lose responsibility, we lose obligations and rights. Another says that if we lose responsibility, we lose deep relationships and love. Skeptics have spent considerable time dealing with these challenges. By some lights, they have succeeded.33 It is a different challenge that I  have examined in this chapter, one that has received insufficient attention.34 Can responsibility skeptics (those who premise their view on free will skepticism) deliberate qua skeptics? To decide, we have looked at principles that might govern deliberation. We need them to capture the irrationality or impossibility of deliberating about options that we believe are out of our reach. The simplest way to do so is to maintain that when we deliberate, we must presume our options to be open (deliberative openness). If that’s right, the free will skeptic who engages in rational deliberation must presume that she has several options – a presumption that is incoherent with her doctrine. There are alternatives to this principle, in particular the principle according to which when we deliberate, we must presume that, for each option, if we chose it, we would act accordingly (deliberative efficacy). But this leads the skeptic to a dilemma. For this new principle makes a dangerous use of conditionals to capture something like freedom. Can we define freedom with conditionals? If we can’t, the deliberative principle looks bad: cases that undermine the conditional analysis of freedom will also undermine the deliberative principle. If, however, we can define freedom with conditionals, and overcome the troubling cases, what looks bad is not the deliberative principle but skepticism itself. For if freedom can be defined with conditionals, there is no question that we are free. 246

Deliberation and the possibility of skepticism

This result gives rise to a rather unfortunate dialectic. The responsibility skeptic is unlikely to change her doctrine in response. I have not, after all, given her any reason for thinking that we are sometimes free and responsible for our actions. And yet, I have given some reasons to the responsibility believer not to take skepticism too seriously. He could say for instance: The Myanmar Junta is morally responsible for atrocities against civilians. If the Myanmar Junta is morally responsible for some atrocities against civilians, responsibility exists. So, responsibility exists. . . . and so there must be a mistake in the skeptical arguments. When the skeptic accuses him of being dogmatic, the believer can now respond that dogmatism is acceptable on some occasions. It is acceptable when launched against views that are unlivable. Perhaps in the end the moral responsibility skeptic should not so heavily premise her view on troublesome claims about free will.

Notes * I wish to thank John Hyman and the Roots of Responsibility group for their fruitful comments at the Special Lecture at UCL on 02.12.2022. Many thanks also to Olivier Massin, Kathrin Koslicki, Robin Bianchi, Marco Marabello, Antoine Taillard, Salim Hirèche, Maximilian Kiener, and Tanya Goodchild. 1 See, e.g., Moore (1939). 2 See, e.g., Dworkin (1996). 3 Kenta Sekine rightly points out that flat-earthers find themselves in this sort of situation. 4 See Strawson (1962) for this kind of consideration against responsibility skepticism. 5 See Williams (1972: 34–39), for an important discussion. 6 See, e.g., Strawson (1994), Caruso (2013), and Pereboom (2014). A notable exception is Waller (2011, chap. 3), who argues that we lack responsibility but have free will. 7 For a similar argument about free will, see Lycan (2019, chap. 4). 8 Otherwise, as Kenta Sekine remarks, the requirement would discredit too many doctrines such as anticapitalism or Christianity. 9 Although they often think that we cannot do otherwise (see, e.g., Pereboom 2014, chap. 6), many contemporary skeptics typically refrain from explicitly premising their denial of responsibility on this. Why? They often take their main compatibilist adversary to be the semi-compatibilist, who believes that we can be responsible even if we lack leeway. If my argument in the next section is right, the skeptic and semi-compatibilist should be careful not to endorse, whether as a basis or a consequence of their theory the view that we lack leeway. 10 Contrary to what many have assumed (Cohen 2018: 87; Henden 2010; Nelkin 2004: 215–218; Nielsen 2011: 283–284; Pereboom 2008: 288–289; Taylor 1966: 181–184; Waller 1985: 48), the deliberation challenge should not concern the determinist in general. For the soft determinist, who believes that determinism is compatible with the ability to do otherwise, has no problem regarding deliberation: she can accept deliberative openness (see § 3). 11 See, e.g., Taylor (1966: 168), Clarke (1992: 107), Nelkin (2011: 168). 12 Clarke (1992: 107–108), Pereboom (2008: 291), and Watson (2003: 176) include the identification or weighing of reasons in practical deliberation. 13 Note that whereas we can deliberate about what to do, it is infelicitous to say that we deliberate about what to do if things were different. 14 Nielsen (2011) forcefully argues in this way. I thank John Hyman and the Roots of responsibility Group for presenting it to me afresh. 15 This kind of consideration has led some, including my past self (Chevarie-Cossette 2020), to accept the distinction between practical and theoretical deliberation.

247

Simon-Pierre Chevarie-Cossette 16 Might there be better ways to separate practical deliberation from daydreaming? Might one say that that practical deliberation concerns options which are probably available to one? No: a skeptic could then not deliberate when too many options are considered. See also § 4. 17 This is what Aristotle seems to have endorsed in fact. See NE VI 5, 1140a. 18 She would be in the company of some error theorists (e.g., Streumer 2017) and, to some extent, some consequentialists (e.g., Parfit 1986, chap. 1). 19 See, e.g., Pereboom (2001, 2014, 2021), Waller (2011, 2014). 20 One exception to this is Blackmore (2013). 21 Contra what I defend in (Chevarie-Cossette 2020). 22 I can deliberate about whether to take Oxford Street or Divinity Street even if I don’t believe that each is available – I might have no firm opinion on the matter. Still, I must presume that they are available. The principle could also be phrased in terms of the absence of a belief that the options are unavailable (see Clarke 1992; Chevarie-Cossette 2020: 387–389). 23 See Kapitan (1996: 437) and Dennett (2015: 118–126). 24 Clarke (1992) and Cowan (1969) argue against deliberative epistemic openness: we can sometimes deliberate even when we know what we will decide because we want to decide for the right reason or autonomously (as when we deliberate after receiving a recommendation from a trusted friend). If that is right, the skeptic might be in hot water in these cases – for she deliberates with the full knowledge that she cannot pick a particular option. 25 This is explicitly recognized by several foes of deliberative openness (McKenna & Pereboom 2016: 297–298; Nelkin 2011: 130). 26 The original case was presented by Taylor (1966: 177–178). An early discussion can also be found in van Inwagen (1983: 154). 27 See Nelkin (2004: 223, 2011, chaps. 6–7) and Pereboom (2008, 2014, chap. 5). Clarke (1992: 103) presents a similar principle, but one that features the notion of ability, and as such is not skeptic friendly. 28 I thank Patrick Todd for raising this issue. 29 True, in this case I don’t respect deliberative epistemic openness (which can be combined with deliberative efficacy). But we can change the case: I  know that I  have a mental block that makes me incapable of deciding to Φ or Ψ, not knowing which. 30 Vihvelin (2013, chap. 6) proposes an interesting attempt. 31 Skeptics might also look for a non-conditional analysis of efficacy (see Nelkin 2004: 226–232). But a similar problem emerges: if, for instance, we can capture good deliberation with the idea of a causal nexus, why can’t we use this notion to capture free will, as compatibilists sometimes do? 32 Nielsen (2011: 413–414) discusses a case where it seems rational to decide to do something we think we can’t do: when Aron Ralston cut his own hand stuck under a boulder to save himself. This case is a good reminder that it can be instrumentally rational to engage in irrational deliberation (see Chevarie-Cossette 2020: 402–404). 33 For the challenges, see respectively Haji (2012) and Strawson (1962). For the responses, see Waller (2011, chap. 10), Pereboom (2014, chaps. 7–8), Sommers (2007), Caruso (2018, sec. 3.2–3.3). 34 Or perhaps it has received the wrong kind of attention, as it was wrongly seen as a challenge for the determinist. See note 10.

Further reading For early occurrences of the deliberation challenge (and other practical challenges to free will skepticism), see Richard Taylor’s Action and Purpose (Prentice-Hall), 168ff. and Peter van Inwagen’s Essay on Free Will (Oxford: Oxford University Press, 1983) 153ff. Randolph Clarke’s “Deliberation and Beliefs About One’s Abilities” (1992), who champions the opposite view, is a model of clarity. Dana Nelkin’s excellent book Making Sense of Freedom and Responsibility (Oxford: Oxford University Press, 2011) contains two further important chapters on the matter. See also Derk Pereboom, Free Will, Agency, and Meaning in Life (Oxford: Oxford University Press, 2014) for a remarkable attempt to show that free will skepticism is not in fact far-fetched and that it is compatible with good deliberation.

248

Deliberation and the possibility of skepticism

References Austin, J. L. (1979). Philosophical Papers. Oxford: Oxford University Press. Blackmore, S. (2013). “Living Without Free Will,” in G. D. Caruso (ed.), Exploring the Illusion of Free Will. Plymouth: Lexington Books. Caruso, G. D. (2013). Exploring the Illusion of Free Will and Moral Responsibility. Plymouth: Lexington Books. ———. (2018). “Skepticism About Moral Responsibility,” in E. N. Zalta (ed.), The Stanford Encyclopedia of Philosophy (Spring 2018). Metaphysics Research Lab, Stanford University. https://plato. stanford.edu/entries/skepticism-moral-responsibility/, https://plato.stanford.edu/cite.html. Chevarie-Cossette, S.-P. (2020). “Sceptical Deliberations,” Pacific Philosophical Quarterly, 101(3), 383–408. Clarke, R. K. (1992). “Deliberation and Beliefs About One’s Abilities,” Pacific Philosophical Quarterly, 73(2), 101–113. Cohen, Y. (2018). “Deliberating in the Presence of Manipulation,” Canadian Journal of Philosophy, 48(1), 85–105. Cowan, J. L. (1969). “Deliberation and Determinism,” American Philosophical Quarterly, 6(1), 53–61. Dennett, D. C. (2015). Elbow Room: The Varieties of Free Will Worth Wanting. Cambridge: MIT Press. Dworkin, R. (1996). “Objectivity and Truth: You’d Better Believe It,” Philosophy & Public Affairs, 25(2), 87–139. Haji, I. (2012). Reason’s Debt to Freedom: Normative Appraisals, Reasons, and Free Will. Oxford: Oxford University Press. Henden, E. (2010). “Deliberation Incompatibilism,” Dialectica, 64(3), 313–333. Kapitan, T. (1996). “Modal Principles in the Metaphysics of Free Will,” Philosophical Perspectives, 10, 419–445. Lehrer, K. (1968). “Cans Without Ifs,” Analysis, 29(1), 29–32. Lycan, W. G. (2019). On Evidence in Philosophy. Oxford: Oxford University Press. McKenna, M.,  & Pereboom, D. (2016). Free Will: A  Contemporary Introduction. New York and Abingdon: Routledge. Moore, G. E. (1939). “Proof of an External World,” Proceedings of the British Academy, 25(5), 273–300. Nelkin, D. K. (2004). “Deliberative Alternatives,” Philosophical Topics, 32(1/2), 215–240. ———. (2011). Making Sense of Freedom and Responsibility. Oxford: Oxford University Press. Nielsen, K. M. (2011). “Deliberation as Inquiry: Aristotle’s Alternative to the Presumption of Open Alternatives,” The Philosophical Review, 120(3), 383–421. Parfit, D. (1986). Reasons and Persons. Oxford: Oxford University Press. Pereboom, D. (2001). Living Without Free Will. Cambridge: Cambridge University Press. ———. (2008). “A Compatibilist Account of the Epistemic Conditions on Rational Deliberation,” The Journal of Ethics, 12(3–4), 287–306. ———. (2014). Free Will, Agency, and Meaning in Life. Oxford: Oxford University Press. ———. (2021). Wrongdoing and the Moral Emotions. Oxford: Oxford University Press. Sommers, T. (2007). “The Objective Attitude,” The Philosophical Quarterly, 57(228), 321–341. Strawson, G. (1994). “The Impossibility of Moral Responsibility,” Philosophical Studies, 75(1/2), 5–24. Strawson, P. F. (1962). “Freedom and Resentment,” Proceedings of the British Academy, 48, 1–25. Streumer, B. (2017). Unbelievable Errors: An Error Theory About All Normative Judgements. Oxford: Oxford University Press. Taylor, R. (1966). Action and Purpose. New York: Humanities Press. van Inwagen, P. (1983). An Essay on Free Will. Oxford: Oxford University Press. Vihvelin, K. (2013). Causes, Laws, and Free Will: Why Determinism Doesn’t Matter. Oxford: Oxford University Press. Waller, B. N. (1985). “Deliberating About the Inevitable,” Analysis, 45(1), 48–52. ———. (2011). Against Moral Responsibility. Cambridge: MIT Press. ———. (2014). The Stubborn System of Moral Responsibility. Cambridge: MIT Press. Watson, G. (2003). “The Work of the Will,” in S. Stroud & C. Tappolet (eds.), Weakness of Will and Practical Irrationality (pp. 172–200). Oxford: Oxford University Press. Williams, B. (1972). Morality: An Introduction to Ethics. Cambridge: Cambridge University Press.

249

19 RESPONSIBILITY AND MANIPULATION Massimo Renzo

How should we think about the relationship between manipulation and moral responsibility? The view that responsibility is undermined by manipulation is so widespread that it is often treated as a premise in philosophical arguments. For example, a prominent strategy to defend the view that moral responsibility is incompatible with determinism is the so-called manipulation argument (Pereboom 2014). The argument has the following structure: Pick any compatibilist conditions you might be tempted to invoke as sufficient to make us responsible for our conduct, despite the fact that it was causally determined by our remote past in combination with the laws of nature. For any of these conditions (Ayer 1954; Frankfurt 1988; Fischer & Ravizza 1998; Wallace 1994), we can imagine a case in which we’ve been manipulated by someone to fulfil it. But in that case, the argument goes, surely we cannot be considered responsible for our conduct. Being manipulated is incompatible with being responsible! So, unless we can point at some morally significant difference between fulfilling the compatibilist condition because we’ve been manipulated and doing so because we’re determined by the laws of nature and the past, we should conclude that being determined by the laws of nature and the past is incompatible with being responsible. So entrenched is the idea that manipulation and moral responsibility are incompatible that proponents of this argument expect to be able to rely on it to convince opponents holding very different beliefs about the compatibility of responsibility and determinism. Perhaps because the idea is widely regarded as uncontroversial, it has received little attention in the philosophical debate. This is regrettable. Not only because it’s unclear that the idea can play the role attributed to it in the manipulation argument (Mckenna 2014), but also – and more importantly– because there are a number of scenarios in which being the victims of manipulation clearly does not undermine responsibility for our conduct. Othello, for example, is intuitively blameworthy for killing Desdemona, despite the fact that he is the victim of Iago’s manipulation. If so, the simple idea that manipulation undermines responsibility cannot be correct. Leave blame aside now and consider the following question: Wouldn’t Othello be liable to be harmed in order to protect Desdemona from his attack? Liability to defensive harm, according to many, is grounded in the fact that those who are responsible for waging an unjust attack lose their right not to be harmed defensively. If Othello is not responsible, DOI: 10.4324/9781003282242-28 250

Responsibility and manipulation

it looks as if we should conclude that he’s not liable to be defensively harmed to prevent Desdemona’s death. That also seems implausible. So, the relationship between manipulation and responsibility is more complex than we might think at first. Indeed, the very framing of the question in terms of determining what “the” relationship between the two is can be misleading, since there are different notions we might have in mind when we talk of responsibility as well as of manipulation. Our first order of business, then, is to disentangle these notions and make clear which of them we have in mind when we make claims about the relationship between manipulation and responsibility. Let’s start with responsibility.

Varieties of responsibility and manipulation It is common to distinguish at least two senses of responsibility (Watson 1996). We can be responsible for an action in the sense that it expresses certain evaluative commitments of ours. When this is the case, the action can be attributed to us in the sense that it says something (good or bad) about us – something that that the action in question would not say about us if it could not be attributed to us. But being responsible in this sense is not yet to say that that we are responsible for the action in the sense that we are blameworthy or praiseworthy. Further conditions need to be in place for us to be accountable in this way. A famous case Gary Watson uses to make this point is the following: If I  steal your books, it’s perfectly appropriate to say that I behaved shoddily. The conduct in question is shoddy because it expresses the adoption of bad values, or at least a certain indifference to moral norms, on my part. The conduct is thus attributable to me insofar as it says something bad about me that it would not say if I had taken the books because I was a kleptomaniac, or under hypnosis, or because I reasonably believed that they were a gift (Wolf 2019: 233–234). Still, I might not be blameworthy if, given my upbringing (or perhaps given the fact that we live in a deterministic world) I lacked a sufficiently robust opportunity to adopt different values. If so, I am responsible for stealing the books in the sense that the act can be attributed to me, but not in the sense that I’m accountable for it. Consider now a third notion of responsibility that, unlike accountability, is not concerned with issues of praise and blame. We can be responsible for acting in ways that impose costs on others, where this raises the question of who should bear these costs. Now, blameworthiness, if present, might obviously play a role in how we answer this question. But the question itself is not a question of accountability. Here we are in the business of allocating costs, rather than calling people to account for their behavior. And there will be plenty of cases in which we are not blameworthy for creating those costs (Sartorio 2021). For instance, I might be responsible for threatening your life if I suddenly lose control of my car, even if drive conscientiously and the accident is caused by a mechanical fault that could not be foreseen. This notion of responsibility is important because many believe that when we are responsible in this sense we can become liable to bear the costs associated with neutralizing, or at least reducing, the harm we have produced (McMahan 2009; Otsuka 1994; Tadros 2021). For example, many believe that I am liable to bear significant costs, potentially even losing my life, in order not to harm you with my car; costs that a bystander would not be required to bear. These costs are grounded in my responsibility for the harm caused by my conduct, even if I’m not blameworthy for causing it (since one is not blameworthy for driving conscientiously.) 251

Massimo Renzo

So, we have at least three notions of responsibility: responsibility as attributability, responsibility as accountability and responsibility as a ground for liability. Let’s now look at the notion of manipulation. While responsibility has been widely discussed in the philosophical literature, and philosophers largely converge on the understanding of the three notions discussed previously, manipulation has received not nearly as much attention. A  popular view is that manipulation involves “tricking someone to form faulty mental states” or “making them fall short of certain ideals” that should govern the manipulee’s belies, desires and emotions (Noggle 1996, 2020). Another is that manipulation involves exercising a particular kind of pressure on the manipulee, one that does not raise to the level of coercion, but that falls short of the ideal of convincing others by using persuasive arguments (Baron 2003; Feinberg 1986).1 Both these strategies capture important aspects of manipulation but are ultimately problematic because they fail to distinguish between cases of manipulation that are structurally very different. To see this, consider four different ways in which I can try to influence your behavior in ways that we might be tempted to call “manipulative”. First, I could shape your circumstances so as to create new reasons for you to act as I wish. For example, I could create a new reason for you to please me by dangling the possibility of introducing you to someone who can facilitate your career. Second, I could shape your perception of your circumstances so as to get you to act as I wish because doing so will seem to you what you have reason to do, when in fact it isn’t. For example, I could induce you to not get a vaccine by presenting you with evidence that mistakenly, but purposively, suggests that the vaccine has dangerous side effects. Third, I could get you to act as I wish by implanting in you ideals and goals in ways that bypass your capacity to assess their merits. This is what happens in cases of brainwashing or indoctrination.2 Finally, I could exploit certain irrationalities we are all prone to in order to trigger in you psychological states that are not responsive in any way to what you have reason to do. For example, I could try to entice you to buy the house I’m showing you by lighting cookie-scented candles during your viewing.3 These strategies are structurally very different. The problem with the view that manipulation involves “tricking someone to form faulty mental states” or the view that manipulation involves exercising a type of pressure on the manipulee is that they fail to capture these differences. The latter view, for example, fails to discriminate between cases in which one is pressured to do x because the manipulator has created new reasons for her to do so (an instance of the first strategy) and cases in which the manipulator has compelled her to do x by exploiting certain phobias or other psychological states that are not responsive to reasons (an instance of the fourth strategy). The former view is so generic that it’s oblivious to the distinction between the second, third and fourth strategy. I think we should be suspicious of any attempt to provide a unified account of manipulation. Any such account would have to be so broad that its philosophical interest would be necessarily limited. We could say, for example, that manipulation involves trying to influence other people’s behavior by interfering with their effort to exercise their practical agency and respond to what they have reason to do. This is true of all the previous strategies. But it’s not very informative. If we want to understand the wrong of manipulation, we’ll need to provide a separate account for each of the four strategies identified previously. For reasons of space, here I shall only consider the second one. Let’s look at it in more detail.

252

Responsibility and manipulation

Manipulation and practical reasoning We are interested here in cases of manipulation in which A shapes B’s perception of her circumstances so as to get B to act as A wishes because doing so will seem to B what B has reason to do, when in fact it isn’t. Let’s start by getting clearer about what happens in cases that have this structure, which will help us explain the nature of the wrong we suffer when we are manipulated in this particular way. We will then rely on this account to assess how our responsibility is affected when we are the victim of this type of manipulation. Start with a simple thought: Exercising our practical agency is a matter of adequately responding to the reasons for action that apply to us. If so, there are two ways in which we can fall short, qua practical agents. One is obvious: We can fail to do what we have reason to do. A second, less obvious, way consists in doing what we have reason to do for the wrong reason. Suppose for example that I  have reason to get a vaccine because this will reduce the risk of contracting a dangerous virus, but I instead get the vaccine because I believe that this will improve my IQ. In this case, although I end up doing what I have reason to do, there is a sense in which I fall short in responding to the reasons for action that apply to me. I fail to be moved by those reasons in the way they call for, even if I end up conforming to what they require.4 The sort of manipulation we are interested in involves intentionally causing this sort of failure in how we exercise our practical agency. So, I might manipulate you either by getting you to fail to do what you have reason to do or by getting you to do what you have reason to do for the wrong reasons. Because of my intervention, the actual reasons for action that apply to you and the apparent reasons that I have created for you are misaligned so that the more you try to respond to your actual reasons for action, the harder it will be for you to do so.5 Importantly, this misalignment between your actual and apparent reasons for action needs to be produced in a particular way, if it is to count as an act of manipulation. To begin with, I must be intervening with the intention of producing this effect.6 If the same effect is produced accidentally, I might mislead you but not manipulate you. Second, I must produce this effect by exploiting your capacity to deliberate about how to respond to what you have reason to do. To see why the intention to produce the relevant effect is necessary, imagine that, unaware of the fact that a crime has been committed in the seminar room of my department, I  place in the room the library card you’ve forgotten in my office, expecting you to find it the following day. A  detective finds the library card and forms the belief that you have something to do with the murder. Clearly in this case, the detective hasn’t been manipulated by me. It is only if I leave your library card with the intention of inducing the detective to form the belief that you are connected to the murder that they have been manipulated by me. The motivation for the second condition is that I  could generate the misalignment between your apparent reasons and the actual ones in all sorts of ways that intuitively do not count as manipulative: I could put some LSD in your drink or stun you by hitting you in the head with a wrench. These two conditions are important to understand why we are wronged by (this kind of)7 manipulation. To begin with, when we are manipulated, our interest in adequately responding to the reasons for action that apply to us is set back. But the mere fact that one

253

Massimo Renzo

of our interests is set back is not enough to explain why we are wronged. It is only when the set back is produced intentionally or recklessly (or, possibly, negligently), that we are wronged by those who caused it. Secondly, when we are manipulated, we are wronged because the manipulator exploits our very capacity to respond to our reasons for action to induce us to fail to do so. Typically, the best way to respond to our actual reasons for action is to act as our apparent reasons for action call for. Aware of this fact, the manipulator ensures that our apparent reasons fail to track our actual ones, so that the more we strive to respond to what we have reason to do, the less likely we are to do so. In this sense, when we are manipulated, our practical agency is not only impaired but, in a sense, turned against itself. It’s precisely by doing our best to respond to our reasons for action that we condemn ourselves to fail to do so, because the manipulator’s interference ensures either that we end up acting against what we have reason to do or that we do what we have reason to for the wrong reasons. What makes manipulation distinctively objectionable is the fact that it co-opts the capacity of the manipulee to respond to their reasons for action and uses it to induce them to fail to do so. In this way, the manipulee becomes an accomplice in the very wrong they suffer.8

Attributability and accountability Now that we have a clearer picture of what manipulation involves, at least in one of its forms, we are in a position to consider the question of whether our responsibility is affected when we are manipulated. Start with the notions of responsibility as attributability and responsibility as accountability. A relatively simple story can be offered about both. Consider, for example: Bomb: Al wants to kill Clint, whom he dislikes. To do so, he provides Bianca with overwhelming false evidence of the fact that Clint is about to detonate a bomb that will kill a large number of innocent people. Bianca can kill Clint at a relatively small cost to herself, and thus reasonably believes that she has a duty to do so. Bianca attacks Clint. There is little doubt that Bianca’s conduct wrongs Clint. But can the wrongful conduct be attributed to Bianca? Not if doing so presupposes that the conduct expresses Bianca’s adoption of bad values or her indifference to moral norms. After all, Bianca is doing her best to respond to the actual reasons for action that apply to her. She fails to do so only because Al’s manipulative interference has created a mismatch between Bianca’s actual reasons for action and her apparent one. Insofar as her action reveals anything about Bianca, it’s that she’s someone ready to act as morality requires. In this sense, responsibility as attributability is undermined by manipulation, at least in cases like Bomb. The wrong suffered by Clint cannot be attributed to Bianca in the way it would had she not been manipulated. What about responsibility as accountability? The relationship between responsibility as attributability and responsibility as accountability can be understood in different ways (Watson 1996; Shoemaker 2015). Generally, however, we can be accountable only for conduct that can be attributed to us. The conditions of responsibility as accountability are those that make it legitimate for others to blame us and, possibly, demand an apology or compensation for our conduct. But blame, apology and compensation are normally inappropriate in relation to conduct that cannot be attributed to us. Thus, insofar as manipulation

254

Responsibility and manipulation

undermines responsibility as attributability, it undermines responsibility as accountability as well. Here we need to tread carefully, however. By saying that responsibility as attributability and responsibility as accountability are undermined in cases like Bomb, I do not mean to suggest that Bianca is not a legitimate target for moral praise and moral blame merely in virtue of the fact that she’s been manipulated. This is because how praiseworthy or blameworthy we are is normally a function of how we respond to the apparent reasons for action that apply to us; and we can certainly assess how someone responds to her apparent reasons for action even when, as in Bomb, those reasons fail to track the reasons for action they in fact have. Suppose, for example, that killing Clint would be dangerous, and thus Bianca reasonably believed that doing so was supererogatory, rather than mandatory. In that case, we might think that Bianca would be praiseworthy for attacking Clint, even if in fact she had no reason to do so and Clint was wronged by her. And we’ve already seen how Othello can be blameworthy for killing Desdemona, despite the fact that he’s been manipulated by Iago. Once again, this is because in assigning blame to Othello, we only need to assess how he responds to his apparent reasons. So, the reason why Bianca is not blameworthy in Bomb is not that the victims of manipulation are not legitimate target of moral praise or blame, but rather that, given the way in which Bianca responded to the apparent reason for action that apply to her, the wrong suffered by Clint cannot be attributed to her. This is what excuses Bianca’s behavior. Bianca is excused because the way in which she responds to her apparent reasons for action does not reflect poorly on her or her character.9

Liability to defensive harm Time now to consider responsibility as a ground for liability to defensive harm. Assessing the impact of manipulation here is less straightforward. Consider again Bomb and imagine that the only way for Clint to defend himself is to inflict defensive harm on Bianca. Does the fact that Bianca is not blameworthy mean that she’s not liable to be harmed defensively by Clint? For simplicity, let’s keep the amount of harm equal, so imagine that Bianca reasonably believes that she can only stop Clint by killing him, and that the only way in which Clint can prevent that is by killing Bianca in turn. Is Bianca not liable to be killed defensively by Clint in virtue of the fact that she’s not responsible in the sense of the notion that grounds blameworthiness? That does not seem right. To see this better, contrast Bomb with the following two cases: Canyon: Al wants to kill Clint, whom he dislikes. To do so, he grabs Bianca and throws her down a canyon, knowing that her body will crush Clint, who is sleeping at the bottom of the canyon. Sleepwalking: For the first time in her life, Bianca has an episode of sleepwalking, during which she attacks Clint. Clearly Bianca is responsible for the threat she poses to Clint in Bomb in a way in which she is not for the threat she poses to him in the other two cases. Although there is a sense in which Bianca is exercising her agency in Sleepwalking, this is not an instance of responsible agency. Bianca lacks the sort of control over her conduct that is required to consider

255

Massimo Renzo

her morally responsible for it. This is even clearer in Canyon, as in that case Bianca is not exercising her agency at all (and thus, a fortiori, cannot be considered a responsible agent). She might be causally, but not morally responsible for the threat to Clint. Bomb is different. In this case, Bianca is responsible in the sense that she voluntarily imposes a risk of harm on Clint after freely deliberating about what to do. If, as many believe, this sort of responsibility is enough to make one liable to defensive harm (assuming, of course, that the harm is necessary and proportionate), it looks as if we should conclude that in Bomb Bianca is liable to be killed to defend Clint.10 One might be tempted to challenge this view by objecting to the claim that Bianca “freely deliberates” about what to do. Doesn’t Al’s interference preclude precisely that, since it prevents Bianca from understanding that she is attacking an innocent person? Korsgaard, for example, writes that someone who’s the victim of deception “doesn’t know what the real end of the action is, and is therefore not in a position to make it his own – to choose, freely, to contribute to its realization” (Korsgaard 1996: 347). If she’s right, it looks as if we should conclude that in Bomb, like in Sleepwalking, Bianca exercises her agency but not her responsible agency. That, however, cannot be right. We often act without an adequate understanding of what the consequences of our conduct will be, whether or not our misunderstanding is the product of someone’s manipulation. If that were enough to render us non responsible, the class of instances of non-responsible agency would be too large. For example, Bianca would not be responsible in the following two variations of Bomb: Bomb (Mistake): Bianca has overwhelming evidence that Clint is about to detonate a bomb that will kill a large number of innocent people. Bianca’s belief is not unreasonable, and she has not been negligent in collecting evidence. However, Bianca’s belief is mistaken. In fact, Clint is not about to detonate a bomb. Bianca’s evidence is the product of natural causes. (Say, an optical illusion produced by the odd lighting in the building.) Bianca can kill Clint at no significant cost to herself, and thus reasonably believes that she has a duty to do so. Bomb (Persuasion): Like Bomb, except here Al genuinely believes that Clint is about to detonate the bomb. Acting in good faith, Al persuades Bianca that this is the case. Al’s belief is not unreasonable, and he has not been negligent in collecting evidence. However, Al is mistaken, and Clint is innocent. It’s true not only in Bomb, but also in Bomb (Mistake) and Bomb (Persuasion), that Bianca would act differently if she had a better grasp of her circumstances. And yet, Bianca seems clearly responsible (in the sense of responsibility that grounds liability to defensive harm) in Bomb (Mistake) and Bomb (Persuasion). But if her imperfect knowledge does not undermine this type of responsibility in Bomb (Mistake) and Bomb (Persuasion), there’s no reason to think it does so in Bomb. The takeaway here is that a responsible agent cannot be someone who acts with a correct understanding of the implications of her conduct, or we would be responsible all too rarely for what we do.11 Korsgaard’s line of argument should thus be rejected. Now, if Bianca is responsible in Bomb, and if responsibility is what grounds liability to defensive harm, it looks as if we should conclude that in Bomb Bianca is liable to be killed to defend Clint. Still, it would be implausible if Al’s interference made no difference. Intuitively, the fact that Bianca has been manipulated by Al ought to affect what she is 256

Responsibility and manipulation

liable to. But how can that be, if a) what grounds her liability to be harmed is the fact that she’s responsible for the harm she’s inflicting on Clint, and b) she’s no less responsible for the harm she inflicts on him in Bomb than she is for the harm she inflicts on him in Bomb (Mistake)? The answer, of course, is that in Bomb, Al plays a distinctive role in producing the harm Clint is threatened with. This matters for several reasons. To begin with, unlike in Bomb (Mistake), Bianca is not the only agent responsible for the threat Clint is facing. Al also is. At the very least, this raises the question of how the costs associated with defending Clint ought to be distributed between Bianca and Al. Secondly, unlike in Bomb (Persuasion), Al is not only responsible, but culpable (i.e., blameworthy) for the harm Clint is threatened with. This matters because, all things equal, those who acts culpably in posing a threat are more liable to bear the costs associated with neutralizing the threat than those who are merely responsible for posing the threat. This means not only that culpable parties are liable to bear larger degrees of harm than merely responsible ones (McMahan 2009; Lazar 2010), but also that, all things equal, their duty to internalize the costs of defensive harm should take precedence over the duty to do so held by responsible parties. Thirdly, and more importantly for our purposes here, Al’s interference matters because it wrongs Bianca. The reason why it would be wrong to impose the defensive costs associated with defending Clint on Bianca rather than on Al is not simply that Al is a culpable agent, whereas Bianca is merely a responsible one. It’s that Al wrongs Bianca, in addition to wronging Clint. Al wrongs Bianca by a) inducing her to fail to adequately respond to what she has reason to do; and b) exploiting her capacity to exercise her practical agency in order to induce her to harm an innocent person. For this reason, he owes a reparative duty not only to Clint, but also to Bianca. Now, the remedial duty Al incurs vis-à-vis Clint is clear. Al ought to bear the costs associated with defending Clint from the harm imposed on him. But what is the remedial duty Al owes Bianca? If being responsible for posing a threat of harm can be sufficient to incur liability to defensive harm, we would have to conclude that Bianca is liable to be killed defensively to protect Clint in Bomb for the same reason why she is liable to be killed defensively in Bomb (Mistake) or Bomb (Persuasion). I submit that Al’s remedial duty toward Bianca is to shield her from such liability, as this is what it takes to offset the costs he has imposed on Bianca by manipulating her. This explains the difference between Bianca’s liability in Bomb and Bomb (Mistake). Bianca is liable to defensive harm in Bomb (Mistake) – assuming that responsibility-based accounts of liability to defensive harm are correct – but not in Bomb, because in Bomb she’s owed a remedial duty by Al, whose liability to bear the cost of defending Clint ­pre-empts Bianca. We should then adopt the following principle: “Pre-emption principle”: When B φ-s and the following conditions are in place: i. B φ-s in response to apparent reasons that are misaligned with her actual reasons; ii. the misalignment between the actual reasons and the apparent ones is intentionally produced by A’s interference with B’s practical deliberation in the way described earlier; iii. A acts culpably or recklessly; 257

Massimo Renzo

A, rather than B, is liable to bear (at least some of) the costs associated with B φ-ing, provided that doing so will secure whichever aims justify the imposition of those costs.12 The pre-emption principle explains a number of things we want to say about cases like Bomb. For example, suppose that Clint’s only way to defend himself against Bianca’s attack would be to kill her. While it might be impermissible for Bianca to redirect Clint’s defensive attack toward an innocent party, it’s not impermissible for her to redirect it toward Al. And while an innocent party could permissibly defend themselves if Bianca were to attempt to redirect the defensive harm necessary to neutralize the threat to Clint toward them, Al would not be allowed to do the same. This is because Al’s liability to defensive harm takes precedence over Bianca’s. Of course, Bianca’s liability is pre-empted by Al’s only insofar as Clint can be effectively defended by imposing the costs of his defence on Al. When that’s not the case, Bianca is liable to bear those costs for the same reasons why she would be liable to do so in Bomb (Mistake) and Bomb (Persuasion). If Clint cannot be defended by harming defensively Al, Al cannot discharge his remedial duty toward Bianca, and therefore her liability is not pre-empted. This is not to say that whenever an agent is manipulated, they are not liable to bear any of the costs associated with neutralizing the threat that they have created. The role of the pre-emption principle is to prevent the victims of manipulation from having to bear certain unfair costs: the costs produced by the manipulator’s effort to ensure that the manipulee’s apparent reasons for action do not align with their actual one. But there are costs that it is not unfair to impose on a manipulated agent: Those that it is appropriate to impose in light of how the agent responds to her apparent reasons for action. This explains why Othello is liable to be killed to defend Desdemona, despite the fact that he’s the victim of Iago’s manipulation. Attacking one’s spouse because of her infidelity is enough to make one liable to defensive killing, so in this case, the fact that Othello has been manipulated does not relieve him of his liability when he attacks Desdemona for that reason.13

Notes 1 On manipulation, see also (Greenspan 2003; Buss 2005; Gorin 2014; Rudinow 1978; and the essays collected in Coons & Weber 2014). 2 The cases of manipulation discussed in the literature on free-will and responsibility (including the various versions of the manipulation argument) tend to be of this kind. 3 This is not meant to be an exhaustive list. 4 On the distinction between being motivated by a reason for action (or “complying” with it) and merely conforming to it, see (Raz 1999: 178–182). 5 On the difference between actual and apparent reasons, see (Parfit 2011). 6 Or at least acting recklessly with respect to the production of this effect (Baron 2003). 7 Henceforth, I will omit this qualification. 8 I develop this account in (Renzo 2023). 9 For an influential account of excuses along these lines, see (Gardner 2007). For Gardner, “excuses are what the person acting upon them takes to be valid reasons, and justifiably so” (110). 10 Prominent defences of this view include (McMahan 2009; Otsuka 1994). Appealing to this notion of responsibility is not necessary to justify liability to defensive harm. A number of further grounds, most obviously culpability, can also be successfully invoked (Tadros 2021). But since this chapter is meant to explore the relationship between manipulation and responsibility, I will focus here on responsibility-based accounts.

258

Responsibility and manipulation 11 See (Buss 2005) for a version of this argument. Buss, however, is interested in the question of whether Bianca acts autonomously, rather than on the question of whether she is responsible for her conduct. 12 I defend this principle at length in (Renzo 2021), on which this section draws. 13 An early version of this work was presented at a workshop on responsibility at the University of Oxford. I’m grateful to the audience for the helpful discussion. I’m especially grateful to Max Kiener for excellent comments on a later version of the paper. The research for this essay was financially supported by a Research Fellowship of the Leverhulme Trust.

Further reading Mills, Claudia. (1995). “Politics and Manipulation,” Social Theory and Practice, 21(1), 97–112 offers an account of manipulation in politics focusing on the relationship between manipulation and reasons for action; Cave, Eric M. (2007). “What’s Wrong with Motive Manipulation?” Ethical Theory and Moral Practice, 10(2), 129–144 discusses a particular kind of manipulation that operates by affecting the motives for our behavior; Blumenthal-Barby, J. S. (2012). “Between Reason and Coercion: Ethically Permissible Influence in Health Care and Health Policy Contexts,” Kennedy Institute of Ethics Journal, 22(4), 345–366 contrasts manipulation with persuasion and coercion, focusing on the debate in bioethics; Baron, Marcia. (2014). “The Mens Rea and Moral Status of Manipulation,” in Christian Coons & Michael Weber (eds.), Manipulation: Theory And Practice (pp.  98–120). Oxford: Oxford University Press provides a sophisticated discussion of whether manipulation needs to be intentional and of whether manipulation is always pro tanto wrong; Gorin, Moti. “Towards a Theory of Interpersonal Manipulation,” in Manipulation: Theory And Practice, 73–977 rejects a number of common claims about the morality of manipulation and sketches an alternative account of why manipulation is wrong; Anne Barnhill, ‘What Is Manipulation?’, in Manipulation: Theory and Practice, 51–72 is a helpful critical discussion of a number of features attributed to manipulation; Wood, Allen W. “Coercion, Manipulation, Exploitation,” in Manipulation: Theory and Practice, 17–50 discusses coercion, manipulation, and exploitation as three different ways in which the value of freedom (understood as independence of one’s choices from the will of others) can be undermined; Manne, Kate. “Non-Machiavellian Manipulation and the Opacity of Motive,” in Manipulation: Theory and Practice, 221–45 is a nuanced discussion of how manipulation need not be intentional or the product of evil motivations; Hanna, Jason. (2015). “Libertarian Paternalism, Manipulation, and the Shaping of Preferences,” Social Theory and Practice, 41(4), 618–643 discusses the relationship between manipulation and practical deliberation, focusing on its implication for the morality of nudging; Sunstein, Cass R. (2016). The Ethics of Influence: Government in the Age of Behavioral Science. Cambridge: Cambridge University Press discusses, among other things, the charge that nudges are morally problematic because manipulative.

References Ayer, A. J. (1954). “Freedom and Necessity,” in Exploring Philosophy: An Introductory Anthology (pp. 271–284). Palgrave Macmillan. Baron, Marcia. (2003). “Manipulativeness,” Proceedings and Addresses of the American Philosophical Association, 77(2), 37–54. Buss, Sarah. (2005). “Valuing Autonomy and Respecting Persons: Manipulation, Seduction, and the Basis of Moral Constraints,” Ethics, 115(2), 195–235. Coons, Christian,  & Weber, Michael (eds.). (2014). Manipulation: Theory and Practice. Oxford: Oxford University Press. Feinberg, Joel. (1986). Harm to Self. Oxford: Oxford University Press. Fischer, John Martin,  & Ravizza, Mark. (1998). Responsibility and Control: A  Theory of Moral Responsibility. Cambridge: Cambridge University Press. Frankfurt, Harry G. (1988). “Freedom of the Will and the Concept of the Person,” in The Importance of What We Care About: Philosophical Essays (pp.  11–25). Cambridge: Cambridge University Press.

259

Massimo Renzo Gardner, John. (2007). “The Gist of Excuses,” in Offences and Defences: Selected Essays in the Philosophy of Criminal Law (pp. 121–139). Oxford: Oxford University Press. Gorin, Moti. (2014). “Do Manipulators Always Threaten Rationality?” American Philosophical Quarterly, 51(1), 51–61. Greenspan, Patricia. (2003). “The Problem with Manipulation,” American Philosophical Quarterly, 40(2), 155–164. Korsgaard, Christine M. (1996). Creating the Kingdom of Ends. Cambridge: Cambridge University Press. Lazar, Seth. (2010). “The Responsibility Dilemma for Killing in War: A Review Essay,” Philosophy & Public Affairs, 38(2), 180–213. Mckenna, Michael. (2014). “Resisting the Manipulation Argument: A  Hard-Liner Takes It on the Chin,” Philosophy and Phenomenological Research, 89(2), 467–484. McMahan, Jeff. (2009). Killing in War. Oxford: Oxford University Press. Noggle, Robert. (1996). “Manipulative Actions: A Conceptual and Moral Analysis,” American Philosophical Quarterly, 33(1), 43–55. ———. (2020). “Pressure, Trickery, and A Unified Account of Manipulation,” American Philosophical Quarterly, 57(3), 241–252. Otsuka, Michael. (1994). “Killing the Innocent in Self-Defense,” Philosophy & Public Affairs, 23(1), 74–94. Parfit, Derek. (2011). On What Matters. Volume 1. Oxford: Oxford University Press. Pereboom, Derk. (2014). Free Will, Agency, and Meaning in Life. Oxford and New York: Oxford University Press. Raz, Joseph. (1999). Practical Reason and Norms. Oxford: Oxford University Press. Renzo, Massimo. (2021). “Manipulation and Liability to Defensive Harm,” Philosophical Studies, 178(11), 3483–3501. ———. (2023). “Manipulation and Practical Reasoning,” manuscript. Rudinow, Joel. (1978). “Manipulation,” Ethics, 88(4), 338–347. Sartorio, Carolina. (2021). “The Concept of Responsibility in the Ethics of Self-Defense and War,” Philosophical Studies, 178(11), 3561–3577. Shoemaker, David. (2015). Responsibility from the Margins. New York: Oxford University Press. Tadros, Victor. (2021). “Two Grounds of Liability,” Philosophical Studies, 178(11), 3503–3522. Wallace, R. Jay. (1994). Responsibility and the Moral Sentiments. Cambridge, MA: Harvard University Press. Watson, Gary. (1996). “Two Faces of Responsibility,” Philosophical Topics, 24(2), 227–248. Wolf, Susan. (2019). “Attributability and the Self,” in D. Justin Coates & Neal A. Tognazzini (eds.), Oxford Studies in Agency and Responsibility Volume 5: Themes from the Philosophy of Gary Watson. New York: Oxford University Press.

260

20 RESPONSIBILITY AND COERCION Carla Bagnoli

1 Introduction In law and morality, legal and ethical practices, coercion is taken to affect the agent’s responsibility for action, but there are disagreements about how and why this is so. These disagreements depend on the ways the key concepts of coercion, agency, and responsibility are characterized and related.1 This chapter departs from the current distinction between physical and volitional coercion by refocusing on the relation of dominance that coercion aims to establish. Coercion is the result of a struggle of independent wills, in which one is forced to submit to another, but the modes of subjugation vary, and some build upon oppressive social, cultural, and economic structures. The key notion of responsibility is mutual accountability, which differs from blameworthiness, and responds to the breakdown of relations based on mutual respect and recognition. The chapter privileges the perspective of the victim and considers the issue whether the coercee can legitimately disclaim, take, or claim responsibility for submissive action.2 This perspective shows that the implications of coercion for moral responsibility and accountability are far more complex than practices of third-person attribution assume.3 The failure to account for this complexity leads one to underappreciate the varieties of coercion, the distinctive wrong done and suffered in such varieties, and the kinds of responsibilities that arise from them. I discuss cases in which the coerced agents can disclaim responsibility for action on grounds different from incapacitation, and others in which they can be held accountable and take or claim (moral) responsibility for their submissive action because and insofar as they exercise some agential powers in deliberating about what to do. The analysis of these cases highlights the multiple and distinctive ways in which victims of coercion are wronged, and calls attention to the conditions under which coercion excuses and blocks punitive and retaliatory attitudes while favoring other pro-active and prospective attitudes. Finally, it raises questions regarding the grounds for resisting dominance, as well as the normative expectations, and moral responsibilities associated with the capacity for refusal and resistance, in a prospective and interpersonal dimension.

261

DOI: 10.4324/9781003282242-29

Carla Bagnoli

2  Disclaiming responsibility under physical coercion In the case of physical coercion, the means of subjugation are physical violence and force. By and large, victims of physical coercion are not accountable for their submissive action, and they are justified in disclaiming responsibility for what they did under coercion, because they are prevented from exercising their agency.4 Physical detention wholly exempts from responsibility because it does not only block the agent’s efficacy but prevents the exercise of full agency. For instance, prisoners cannot be held accountable for not fighting in support of their companions because they were detained, physically prevented to act; they can disclaim responsibility for the omission of help [example 1].5 Agents tortured can be excused for releasing information that damages third parties; they can disclaim responsibility for a breach of confidentiality [example 2]. Enslaved prostitutes are victims of objectification, abducted and sold on the black market, treated like mere objects, and subjected to physical and emotional abuse can disclaim responsibility for what they do [example 3]. They can be excused on the ground that they were subjected to violence designed to break the will, and akin to incapacitation (Austin, 1961). In all these cases, coercion disables agential capacities that agents normally have. Like obstacles, impediments, and hindrances, physical coercion obstructs the agent’s capacity for action and, a fortiori, it impedes ordinary agency (understood as expressive of oneself and directed to chosen ends). Unlike obstacles and impediments, however, the coercer superimposes on the agent’s, orients his efficacy, and thus undercuts his authority of action. This interference is wrongful, though different theories disagree about what makes it wrong, and how it wrongs the agent.6 I argue that it is wrong because and insofar as it establishes a relation of dominance. The previous cases exemplify different degrees of severity. When coercion annuls the agent’s efficacy, the submissive action is not even attributable.7 In other cases, coercion undermines but does not sever the relation between agency and the will. As we shall see, some paradigmatic examples of coercion via conditional threat are such that the victims are treated as mere means and become instruments in the hands of the coercers. While different forms of coercion challenge different dimensions of the normative status of persons, all of them challenge the agential authority of the agent and generate a form of alienation. For this reason, many definitions of coercion by threat refer to violation of autonomy or freedom.8 However, not all forms of coercion obtain by disabling the agent’s capacity for action. In the remaining part of the chapter, I  discuss varieties of coercion, which do not operate by denying the basis of agential responsibility for action. Rather, they function by implicating the victims in action against their will. The case in point is volitional coercion.

3  Taking responsibility under volitional coercion The category of volitional coercion differs from physical coercion in that victims of volitional coercion are not undermined in their agential efficacy but challenged in their normative authority of action by altering their reasons for action. That is, coercion affects the way agents relate to their action, by impacting on the reasons why they undertake action and do what they do. These victims are agents. In some of these cases, which closely resemble some of the cases of physical coercion – the agents are treated as mere means: they are instruments to achieve the coercer’s ends. However, they are not treated as 262

Responsibility and coercion

objects; they are not denied their status of independent wills, capable of setting ends and attributing values. In fact, their instrumentalization depends entirely on the recognition of such capacities for setting ends and valuing. Their normative authority is challenged in distinctive and complex ways. Importantly, volitional coercion does not leave the agent’s will intact because the coercer’s role is not merely obstructive or constrictive. The coercer imposes himself as an external source of normative authority that reconfigures the agent’s will and channels its expression into pre-ordered venues, hence contributing to the formation of reasons for action. In a trailblazing article, Robert Nozick captures the latter feature of volitional coercion by taking conditional threats to provide the agent with normative considerations for action that are only partly exogenous, since they are firmly anchored to the agents’ own interests. Nozick’s definition of coercion provides necessary and sufficient conditions in terms of proposals that exert (normative) influence on the agent, because they furnish him reasons for action.9 Nozick’s schema (1969: 441–445) is as follows: 1. A aims to keep B from choosing to perform action x; 2. A (successfully) communicates a claim to B; 3. A’s claim indicates that if B performs x, then A will bring about some unwelcome consequence that would make B’s x-ing less desirable to B than B’s not x-ing; 4. A’s claim is credible to B; 5. B does not do x. Part of B’s reason for not doing x is to lessen the likelihood that A will bring about the consequence announced in (3). The novel feature of Nozick’s analysis is that it focuses on the coercee as a rational agent susceptible to conditional threats because capable of rational bargaining, an independent will that reacts and engages a fight with the coercer, rather than a passive victim, suffering a breakdown of the will, or “wholly dependent on the coercer for his needs” as suggested by alternative construals of coercion (cf. Frankfurt 1973).10 Coercive threats do not operate by disabling and curbing the agent’s agential power. On the contrary, coercion is successful only if such power is properly stimulated by a powerful and credible threat. Coercion does not destroy the coercee’s capacity for rational choice but exploits it. A conditional threat is a claim that the coercer makes to force the agent to act to his advantage by relating the prospective action to adverse consequences that will occur if the agent does not comply with the coercer: e.g., If you tell anybody I stole the classified doc, I will hurt you [Example 4]. The threat has leverage because the coercer depicts a possible scenario that the agent has reason to prevent from happening. On Nozick’s construal of the case, the agent alone has the power to avert the threatened scenario by submitting to the coercer’s request; if he does not, the realization of the threatened scenario is his doing. Thus, the threat addresses the agent as the one with (i) the interest, (ii) the power and opportunity, and (iii) reasons to comply – all which point to the responsibility to act. The coercer engages the victim’s rational skills by threatening a scenario that is conditional on his submission. The potential victim is somebody for whom a threat is intelligible. She can reason about what to do, and sensitive to normative considerations. To appropriately respond to the coercive threat, victims need to exercise rational skills, assess the credibility and the normative significance of the threat, assessing the prospects of action, and ranking the forced options of choice. 263

Carla Bagnoli

On this account, victims can claim or take responsibility, and be held accountable for their submissive action even when protesting that the rational bargain is no genuine choice because the options are artificially constrained. This claim is not paradoxical, in contrast to other definitions of coercion as violation of autonomy. The victim is wronged precisely by being forced to take responsibility for an action that she would not have reason to choose prior to and independently of the coercive threat. She is not made into an instrument but forced to enter in a shared deliberation with her coercer, and thus implicated in the wrong done to her. Correspondingly, the agent operating under conditional threat might be expected to take responsibility for the submissive action, since this is the result of his deliberation. But this is not the same as expecting him to take the blame for it; nor does it necessarily license the third-party blame. While agents deliberating under threat can take responsibility for their submissive action, they may still justifiably refuse to take the blame for it. Taking responsibility and taking the blame are not synonymous. Furthermore, while submission is blameworthy there are cases in which the submissive agent is not (see section 8).

4  Claiming responsibility under coercive threat Consider the conditional threat: If you do not betray your country, I will kill your family [example 5]. In this case, the third concerns third parties. To obtain a mendacious testimony, the coercer exploits the victim’s moral sense, close ties, and possibly a division within the moral domain – between partial and impartial reasons. He exploits the fact that the victim is autonomous, i.e., capable of setting intrinsic and unconditional ends.11 While the victims of threats like 5 have their agential authority challenged, they may orient their choice in a way expressive of their values. They may have moral qualms betraying their country and yet be convinced that there is a right answer to their tragic choice. In justifying their response to the coercive threat, they may claim responsibility for submitting to it. I introduce a distinction between taking and claiming responsibility for action to emphasize two agential modes open under coercion. The former verb has a more passive overtone: the agent takes responsibility for aspects of action that were not willed, that exorbitated one’s sphere of intentional agency while they followed an intentional act. The latter emphasizes a more active posture of the agent toward her action: by claiming responsibility for action under coercive threat, the agent stresses that despite the wrongful constraints, he expressed his agency in ways that he does not think arbitrary and unreasonable. He chose to submit not based on the duty to protect, which allegedly overrides the duty to tell the truth, but rather in conformity with his cherished values and in ways that are expressive of his character. This distinction captures modes of response to coercion internal to the victim’s perspective on action, which do not translate in the third-person perspective. In the third-person perspective, it suffices to say that the agent is responsible for submissive action; but the first-person mode tells us how the agent relates to his action and is relevant to assess the nature and measure and measure of alienation of the wrong she suffers.

5  The power to refuse Victims of coercion by threat retain the power to refuse (Nozick 1969, 1981). This is a basic agential capacity that victims of physical coercion cannot exercise. In the most serious cases of long-term exposure to emotional abuse, torture, and violence, it is arguable 264

Responsibility and coercion

that such capacity is not merely obstructed but severely impaired. As argued in section 2, this is a decisive reason for exempting excusing or victims depending on the significance of the impairment, and for victims of physical coercion to disclaim responsibility for action. Instead, the examples of volitional coercion discussed in sections 3 and 4 show that the coercer alters the agent’s normative reasons, by introducing normative considerations. Nozick holds that unless the coercee acquiesces, there is no coercion, and he has the capacity to refuse. However, based on Nozick’s description of the normative pressure exerted by the threat, victims have strong instrumental (example 4) or moral (example 5) to act to avert the threatened scenario and represent themselves as bound by such (rational and moral) reasons. In some tragic cases, the victim represents himself as both instrumentally and morally obligated to comply with the coercer. Thus, on which grounds does refusal rest? The standard approach to coercion by conditional threat encourages a disturbing similarity between instrumentally rational actions and submissive actions under coercive threats. Either the threat is miscalculated, or the victim acts irrationally when he refuses the coercive deal. This problem partly arises because of the narrow scope of Nozick’s analysis.12 A plausible starting point to correct this tendency is to consider that agent deliberating under coercive is bound by the requirements of due deliberation, which are established by a normative theory. However, before entering any dispute in normative theory, there are two strategies for rejecting the claim that morality and rationality inevitably offer divergent grounds for resisting coercion. These two strategies may eventually lead to mutually incompatible normative theories, but at this point in the argument, they are both open. The first strategy is to enhance the instrumentalist approach by expanding the scope of due deliberation beyond the context of choice at the time of action. The instrumentalist should argue that the effects and wrongness of coercion should be calculated in a temporally and socially expanded perspective. This strategy allows the instrumentalist to broaden the scope of deliberation and refocus on a larger set of reasons for rejecting submission that presently avoids the worse scenario. It does so by re-describing the worse scenario in a way that includes prospective reasons and potential victims. This temporal and social expansion significantly complicates matters: It commits the agent to deliberate about possible interpersonal and inter-temporal conflicts, and well beyond his own self-interest at the time of choice. The second strategy is based on the view that only a non-instrumentalist approach can vindicate independent reasons for refusing a coercive deal, that is, reasons that cannot be balanced and weighted against others. This argumentative strategy institutes a difference in kind between ordinary action and action under coercion; second, it categorizes coercion as an intrinsic disvalue; and third, it disputes that the reasons to submit can be measured up and override reasons of autonomy – however the latter concept is defined. On both these views  – the enhanced instrumentalism, and non-instrumentalism  – the agent deliberating under coercive threat may decide to refuse to submission and do so in conformity with the requirements of practical reason. However, only on the latter view can the agent refuse to submit solely in response to the attack on her normative status and entitlements as an autonomous agent – independently of the comparative assessment of the consequences. This reasoning excludes the possibility that coercion is morally acceptable if it promotes ‘good ends’, but it does not make submission irrational or immoral. Thus, this characterization explains that the tragic dimension of the victim’s choice is under threat, which produces a loss in value and leaves a remainder. Furthermore, while appreciating 265

Carla Bagnoli

the deliberative burden sustained by the coercee, this view also explains how she may be expected to take responsibility for the consequences of refusal to submit. This is an explanatory advantage in contrast the view that the coercee is relieved of the responsibility for the consequences because they depend wholly on the will of the coercer (cf. Williams 1973; Frankfurt 1973). The comparison between these strategies brings to the fore a feature of coercion that does not emerge from Nozick’s analysis, that is, the possibility that coercion may be rejected as an attempt to challenge the normative status of the coercee and establishes a normative relation of domination. In examples 4 and 5, coercers dominate by exploiting their victims as valuers and rational choosers capable of reasoning and disposed to enter rational bargains to protect what they value and love. However, coercers dominate the submissive agents not only by turning them into means, or by presenting them with irrefusable deals (Fowler 1982) or irresistible motives (Frankfurt 1973), but also by abusing them emotionally, exploiting the emotional dimension of their personal relationship. In such cases, coercers may use the victims to his advantage but, primarily, he wants to impose a relation of power over them. As an example, consider the conditional threat [6]: If you do not eat a saucer of mud, I  will break your leg. Example 6 shows that coercion is not always intended to the coercer’s advantage, in the sense of furthering her own ends. The threat is meant to degrade and mortify the victim, rather than to achieve ulterior ends: humiliation is the very end of coercion. The humiliating effect is amplified by the fact that the demand is arbitrary and absurd. Coercion is not exerted to achieve ends whose realization depends on the victim’s submissive action, but only to ensure her submission, that is, to establish domination.

6  The dynamics of domination In the examples examined, the coercive threat creates an asymmetry of power and, where successful, establishes a relation of dominance. However, coercive threats may build on preexisting asymmetries in power, grounded in economic structures, social schemes, and practices, enacted by individual or groups.13 I call ‘oppressive coercion’ a category of coercive acts that enact social schemas that distribute normative roles and functions in a given social network (e.g., in relation to class, race, or gender).14 Oppressive schemas and practices are not fully transparent, and thus coercers may unreflexively add to or wilfully exploit social schemas, thereby perpetuating oppression while presenting the coercive bargaining situation as a genuine exchange of proposals. In the victim, opacity may induce or mix with the internalization of oppression.15 Let us examine the conditional threat: [7] If you file for divorce, you will lose parental custody. In this case, the coercer is not so much interested in making the victim do something useful to him as in preventing them from doing what she wants. This form of volitional coercion is based on emotional blackmail typically belongs in compromised networks of social relations. Coercion occurs rather pervasively in structured social environments, and its most damaging aspect resides in the fact that it depends on and results from ordinary schemes of social interactions, often contiguous and continuous with morally appropriate normative practices. Refocusing on the recurrent normative relation between coercer and coerced has a significant comparative advantage in this regard.

266

Responsibility and coercion

Understood as a normative relation of dominance, coercion is more fruitfully approached as arising from and developing within specific existing social contexts and networks that are internally organized by normative structures or schemas. However, the latter provides opportunities for exercising powers enacted by persons; thus, personal relations are the real locus of the struggle. This methodological choice is alternative to both individualist and structural approaches and recommends an investigation of moral responsibility in its dynamic dimension, bringing to light differentiated normative grounds for sustaining refusal. In one scenario, the wife decides to reject the husband’s deal out of concerns for maintaining a respectful relationship with her husband and children rather than a worry that her moral integrity will be disfigured. She considers it better for her children to be raised by separated parents who maintain a relationship of mutual respect rather than by parents bound together by emotional blackmail [7a]. Or, she may decide to fight for custody on grounds that go beyond worries about the moral deterioration of her household, as a matter of principle and to oppose the emergence and spread of coercive patterns. Her refusal to submit under coercive threat can be rooted in a prospective concern: the victim may be concerned with the future and be motivated by the conviction that by resisting the present coercive threat, she actively contributes to the establishment of fair relations. Refusing to trade on affects not only defuses the threat in the actual case but also works as a deterrent against future violations of respect.16 It breaks an unfair pattern and serves as an incentive for future rebellion. In a second scenario, the wife justifies and takes responsibility for her choice to stand by her husband on the ground that she does not want to lose custody of the children [7b]. She explains that her choice not to file for divorce is based on her heightened sensitivity to the consequences that divorce may have on herself as well as the children. In this case, taking responsibility for submissive action counts as an act of reassessment and re-appropriation of one’s agency, rather than a compromise. Finally, coercive relations evolve over time and transform the agents involved. In a distant future, the wife reveals that the husband’s threat made her realise how important it was for her to raise their children together [7c]. This awareness makes the husband’s threat secondary and circumstantial. Retrospectively, the fact that she made this decision under threat has become entirely irrelevant, almost inessential. In fact, she does not even regard her choice as a negotiation. This narrative explains how she has come to terms with the consequences of the choice that she made under threat, but it does not diminish the wrong done to her. In fact, taking responsibility for a coerced action without resenting the coercive threat is morally problematic because it may indicate that the agent has internalized a relation of dominance. Nonetheless, at least in some cases, reclaiming responsibility for an action performed under threat (whether of submission or resistance) qualifies as a normative response to the coercer’s attack, a response by which the agent reaffirms his normative agentive power. An advantage of the relational approach is that it attends at the evolution of the normative relation between coercer and coerced across time by tracking the exchange of justifications and the dynamics of coercive relations over time. The judgment of moral responsibility responds to a breakdown in the normative relation between the coercer and the coercee. For both coercer and coercee, identifying and locating (mutual) responsibilities may work towards understanding the direction and repairing a personal relationship. To this extent, the relational model has a practical and prospective

267

Carla Bagnoli

application as it focuses on the acts of taking and claiming responsibility as modes of counteracting and overcoming the effects of dominance.

7  Coercive offers Insofar as it operates through social structures and targets members of a group,17 oppressive coercion is a collective phenomenon. The social and economic structure may be taken to organize and be expressive of a collective, through codified norms that operate at the level of the interpretation and organization of the social world and, thus, arrange personal relations into patterns of dominance. Consider the offer made to an undocumented immigrant who cannot be vaccinated through the national health system [8]: “I offer you a job for which you need no proof of vaccination, just not with a regular salary”.18 This offer is coercive because the undocumented immigrant is forced to accept the proposal as the best alternative available to him, although the offer results from unfair restraining of other alternatives through the exercise of economic and social power.19 The victims of oppressive coercion have their normative status as individual agents and members of the relevant normative community seriously undermined in various interlocked and cumulative ways. B’s agency is constricted in deep ways but not in ways that emerge via the analogy with physical coercion. First, oppressive coercion not only systematically limit the victims’ opportunities but also undercuts the victims’ sense of their opportunities for action: thus, creating external obstacles but also internal hindrances that block their agency and diminish their efficacy. Second, it deprives them of real bargaining power within specific social settings and endangers their capacity to participate in cooperative interactions. Third, it introduces exogenous ends that undermine the agents’ motivational stability and, more importantly, distract them from the pursuit of their own ends, obstructing or delaying the achievements internal to their conception of a good life. Fourth, oppressive conditions weaken the speculative and epistemic capacities of practical agents. Inasmuch as these agents are constantly subjected to external interference, they have limited predictive abilities. Finally, these agents perceive the social world that they inhabit as inhospitable, such that their agency is forcefully channelled in unpredictable ways. Victims of oppressive coercion are likely to experience a severe discordance between their normative and empirical expectations, that is, between what they should expect and demand from their social and institutional world and what they do expect. This is a form of alienation that thrives on epistemic opacity, that is, lack of knowledge of the full details of one’s own place in the social and institutional world, but its effects are not exclusively epistemic because this condition affects the victims’ perceived reliability in interpersonal contexts and thus also severely undermines their prospects in strategic and cooperative interactions. The fact that submissive agents do not have the strength to refuse coercive deals sometimes exposes a volitional weakness or a vulnerability to others that may be found morally suspicious. Conversely, the woman who stands up to a coercive boss under threat of being fired attracts admiration and praise as a display of the virtue of courage or willpower. Is the normative expectation of refusal associated with a duty? A strong argument for treating refusal to submit as a strict duty is based on the duty of self-respect (respect for humanity) and applies in cases such as 4.20 A second argument concerns the negative effects on possible victims, and descends from the duty of respect toward humanity, and applies to 268

Responsibility and coercion

cases such as 5. In cases such as 6, a third argument derived from respect toward humanity applies, which is that submissive actions support oppressive structures.21 The submissive victim becomes involved in the operation of oppressive institutions and, therefore, helps to keep unfair practices alive. The victim can be held accountable for the negative testimonial effects, insofar their submission sets a bad example for other members of the same group who are likely to find themselves in similar situations in the future. The submissive attitude reinforces the stereotypical division into oppressed and oppressors, political actors, and instruments of power. Correspondingly, submission may undercut potential victims’ motivation to endure the power struggle. Based on these considerations, there is a moral and political duty to resist, justified by the consideration that this is the primary way in which the agent may contribute to changing or uprooting unfair practices and inspire others to do the same.

8  Responsibility to resist In the case of massive wrongs, however, it is arguable that the normative expectation to resist oppressive coercion is cognitively and morally inappropriate. Thus, for instance, it may seem utterly unfair to blame women who conform to coercive demands in a patriarchal society.22 In conditions of massive wrongdoing, victims cannot be expected to have developed an overall vision of social conditions radically different from the one in which they are immersed. Furthermore, as an effect of prolonged abuse, they may not have the strength and resilience to put up a fight. Thus, to impose on them a duty to resist is unappreciative of the condition of dominance that they suffer. It constitutes unfair overburdening. These are certainly important considerations, but how do they affect the practices of taking, claiming or disclaiming responsibility for failure to resist under oppressive coercion? First, judgments ascribing responsibility are not tantamount to judgments of blameworthiness. Blame is but one among many reactive attitudes that respond to others as capable of responsible agency and resistance. Second, oppressive coercion can be an excusing condition, which not only suspends blame but weakens the judgment attributing responsibility. It can be also the ground for disclaiming responsibility for particular actions, or omissions, including responsibility for failure to resist the coercer. However, it can be invoked also in defence of the oppressors on analogous grounds.23 A significant drawback of these arguments is that it may underestimate the victims’ resilience and their capacity for driving social change.24 Likewise, the argument exculpating coercers in oppressive networks may underestimate their understanding, involvement, and capacity for change. While the circumstances of massive and normalized injustice corrode agential capacities, neglecting the agential roles of the coercers and victims can be counterproductive and have further undermining effects on them. Third, the corrosion of the capacity to resist is counteracted through engagement with others. If agents no longer have sufficient self-respect to consider themselves worthy of fighting, we can conclude that the oppressive condition to which they have been subjected has fatally undermined their status as agents. Agential capacities can be recovered thanks to supporting and enhancing relations. In this perspective, investing somebody with normative expectations that seem unrealistic at the moment may be an effective form of ethical and political empowerment. Reactive attitudes play a key role in the restoring dignity to victims. Even blame of submission may have self-enhancing and empowering effects rather than be used as a retributive and punitive measure. 269

Carla Bagnoli

Finally, since oppressive coercion is a collective phenomenon, the duty to resist does not fall on the agents as isolated individuals but also as members of organised collectives. All members of the organised collective have a moral and political responsibility to sustain agents under oppressive coercion in their process of resisting and reclaiming responsibility for action. Thus, the next step is to identify and distribute the different kinds of responsibilities generated by coercion.

9 Conclusion This chapter argues that coercion names a cluster of diversified phenomena, whose conditions of moral responsibility vary, and whose unifying feature is that they are marked by a personal relation of dominance, which exploit and build upon various mechanisms, including physical violence, emotional blackmail, threats, and cultural and social structures. This approach provides diversified but related diagnoses of the wrongs done via coercion. In some cases, the moral offense in coercion is that the victim’s agency is disabled or frustrated; in others, she is forced to take responsibility for deliberating, planning, and executing the submissive action. Some forms of coercion exploit the fact that agents are capable and willing to entertain non-instrumental ends. These varieties of coercion produce different kinds of responsibility for action or alienation from one’s own agency. Second, this discussion uncovers strong rational and moral grounds for the view that there are duties to resist the coercion, in contrast to the general view that resistance may be rationally unjustified as futile or instrumentally ineffective, or even morally self-indulgent, when third parties suffer the consequences of it. The dynamic approach highlights the complexities of coercive relations behind resistance but also calls attention to empowering effort to rebuild corroded agential and moral competences. The dynamics of disclaiming, taking, and claiming responsibility for action under coercion may work towards repairing personal relationships and reconstruct a just social network.

Notes 1 See, e.g., Bayles 1972, Pallikkathayil 2011; Anderson 2021. I leave aside the discussion of coercion relative to the state’s enforcement of law; but see Raz (1975) for the claim that law does not require coercion to count as law, although coercion may be merely instrumentally needed. 2 In ordinary language, there is no sharp distinction between “taking” and “claiming” responsibility. However, in taking responsibility the agent may have a more passive attitude toward the action, e.g., taking a responsibility for something because she is expected or required to take responsibility for it, and independently of her agential role. In contrast, the agent might claim responsibility independently of whether she is expected or required to do so, and with the positive aim of vindicating her agential role and authorizing one’s action. The distinction is thematized in section 4. 3 I use the term moral responsibility in a broad sense, which includes but does not fully coincide with accountability. One may claim moral responsibility for something, even when one is not accountable to others for it (e.g., for one’s hidden thoughts or attitudes). In the legal domain, the third-person perspective prevails. Coercion is generally taken to affect the coercee’s responsibility for action by way of exempting, excusing, or justifying him, depending on the extent that the submissive action (or its consequences) was beyond his control or willpower to prevent, or unintentional. In moral philosophy coercion is often regarded as a form of incapacitation since the agent is literally unable to do otherwise, because she is impeded, overwhelmed, or compelled (Frankfurt 1973), or lacks rational autonomy (Korsgaard 1996; O’Neill 1991, 2000). 4 On the use of force, see Bayles 1972; Gunderson 1979; Lamond 1996, 2000.

270

Responsibility and coercion 5 Imprisonment is often considered the paradigm form of coercion, Lucas 1966: 60. 6 Fowler (1982), cf. Kroon (1996), DeRose (1992). 7 The coercee is not an agent insofar as he is “acted upon”; see McCloskey 1980. 8 It is disputable that coercion is antithetical to freedom; see Carr 1988: 59; Murray & Dudrick 1995. Different forms of coercion target different modes of being free or autonomous: coercion as compulsion violates personal autonomy (Frankfurt 1973); coercion as instrumentalization, and coercion as domination violates moral autonomy, which is based on negative and positive freedom of choice (e.g., O’Neill 1991; Korsgaard 1996, cf. Bagnoli 2018); coercion enacted by an oppressive schema, indoctrination or authority is a challenge also to political autonomy, see Yaffe 2003; Ripstein 2004. 9 Nozick distinguishes threats from other kinds of conditional proposals such as offers, by reference to the negative impact of threatened scenario, which is worse for the coercee than it would be in the “normal and expected course of events” (1969: 447). However, offers can be used coercively, see Vinit 1976; Zimmerman 1981; Frankfurt 1988 [1973]; Held 1972; Lyons 1975; Feinberg 1986; Stevens 1988. 10 This emphasis explains that coercion suspends the coercee’s moral responsibility for action, Frankfurt 1988: 75–76. On the related issue of freedom, see also Korsgaard 1996, 159–187, 188–222. 11 The relation of coercion is not necessarily a relation of dependence, exploiting the coercee’s needs, contra Frankfurt 1988: 71. 12 Nozick anticipates that further complexities may require revisions, although not in relation to the problem I formulated previously; see Nozick 1969, 445–447. 13 There is a divide between intention- vs. structure-based accounts of phenomena of coercion; see, e.g, Zimmerman (1981) vs. Cudd (2006: 130–136). 14 Coercion here consists in a differential bargaining power (McGregor 1988–89, 25); as such, it is a constitutive feature of oppression; see Cudd 2006 (1994: 22–24). 15 Cudd 1994. 16 See Korsgaard 1996: 155–156. 17 Grouping is not necessarily a voluntary act and may be involuntary even when it conveys social identity. 18 The request of vaccination is not coercive; a separate debate concerns the coercive nature of the measures and controls that apply to immigrants; see, e.g., Miller 2010. It may prove difficult to discern voluntary cooperation from domination. 19 The main strategy to differentiate threats and offers is by comparison to the baseline: threats worsen the coercee’s situation, while offers do not. Nozick defines the baseline in terms of “normal or natural or expected course of events” (Nozick 1969: 447). However, there are cases in which normative/moral and descriptive expectations lend different results; see Nozick 1969: 450–451. Non-normative accounts promise to determine whether coercion has occurred without relying on more contentious normative principles; see Zimmerman 2002; Rhodes 2002. By contrast, normative accounts make judgments of coercion dependent on prior normative judgments, and especially moralized-baseline accounts, Wertheimer 1987; Raz 1986; Carr 1988; Ryan 1980; Hetherington 1999; Anderson 2021: §2.3.   Contra Nozick (1969), many argue that offers can be coercive, as is the case with wage offers in the market; see Zimmerman 1981: 145, Cudd (2006: 125–131), Held (1972), and McGregor (1988–89). For Nozick whether the constrained actor acts voluntarily “depends upon whether these others had the right to act as they did” (1974: 262). But the objection is that capitalist exploitation is never clearly distinct from coercion, given that the coercee’s alternatives are severely constrained. On the discussion whether capitalist employment relations are exploitative rather than coercive, see, e.g., Wood 1995; Zwolinski 2007. 20 There is a dispute about the nature of this duty, whether it is strict or broad, absolute or prima facie, based on respect for humanity, human fellowship; see Hay 2011, footnote 37. 21 Benson 1991; Cudd 2006. 22 This is the subject of Superson 2005. 23 Being able to resist is conditioned by factors external to individual efforts, Boxill 2010: 10–11. In “normal contexts” participation in discriminatory practices is not excusable; in severely compromised contexts such practices may be excusable; see Calhoun 1989: 396; Moody-Adams 1994: 292–296. 24 See Buss 2010; Boxill 2010: 9; Hay 2011.

271

Carla Bagnoli

Further reading Nozick, Robert. (1969). “Coercion,” in S. Morgenbesser, P. Suppes, & M. White (eds.), Philosophy, Science, and Method (pp. 440–472). New York: St. Martin’s Press is a seminal article that conceptualizes coercion by conditional threat. O’Neill, Onora. (1991). “Which Are the Offers You Can’t Refuse?” in R. G. Frey & Christopher Morris (eds.), Violence, Terrorism, and Justice (pp. 170– 195). Cambridge: Cambridge University Press. Kavka, Gregory. (1987). Moral Paradoxes of Nuclear Deterrence. Cambridge: Cambridge University Press centers on ‘nuclear coercion’ and defends nuclear disarmament as a moral and realistic end. Bazargan, Saba. (2004). “Moral Coercion,”  Philosophers’ Imprint, 14 focuses on cases of coercion exploiting moral commitments. Mason, Elinor. (2012). “Coercion and Integrity,” Oxford Studies in Normative Ethics, 2 investigates the relation between coercion and integrity. Madhock, Sumi, Phillis, Anne,  & Wilson, Kalpana. (2013). Gender, Agency, and Coercion (Palgrave Macmillan), explores the relationship between agency and coercion in ethically problematic contexts.

References Anderson, Scott. (2021). “Coercion,” in Edward N. Zalta (ed.), The Stanford Encyclopedia of Philosophy (Summer 2021 ed.). https://plato.stanford.edu/archives/sum2021/entries/coercion/. Austin, John L. (1961). “A Plea for Excuses,” in J. O. Urmson & G. J. Warnock (eds.), Philosophical Papers. Oxford: Clarendon Press. Bagnoli, C. (2018). “Claiming Responsibility for Action Under Duress,” Ethical Theory and Moral Practice, 21, 4868. Bayles, Michael D. (1972). “A Concept of Coercion,” in Nomos: Coercion (XIV, pp. 16–29). Benson, Paul. (1991). “Autonomy and Oppressive Socialization,” Social Theory and Practice, 17(3), 385–408. Boxill, Bernard. (2010). “The Responsibility of the Oppressed to Resist Their Own Oppression,” Journal of Social Philosophy, 41(1), 1–12. Buss, Sarah. (2010). “Reflections on the Responsibility to Resist Oppression, with Comments on Essays by Boxill, Harvey, and Hill,” Journal of Social Philosophy, 41(1), 40–49. Calhoun, Cheshire. (1989). “Responsibility and Reproach,” Ethics, 99(2), 89–406. Carr, Craig L. (1988). “Coercion and Freedom,” American Philosophical Quarterly, 25, 59–67. Cudd, Ann E. (1994). “Oppression by Choice,” Journal of Social Philosophy, 25, 22–24. ———. (2006). Analyzing Oppression. New York: Oxford University Press. DeRose, Keith. (1992). “Deterrent Threats: What Can Matter,” Philosophical Studies, 67, 241–260. Feinberg, Joel. (1986). Harm to Self. New York: Oxford University Press, especially chs. 23–24. Fowler, Mark. (1982). “Coercion and Practical Reason,” Social Theory and Practice, 8, 329–355 Frankfurt, Harry. (1973). “Coercion and Moral Responsibility,” in Ted Honderich (ed.), Essays on Freedom of Action (pp. 65–86). London: Routledge & Kegan Paul. ———. (1988). The Importance of What We Care About: Philosophical Essays. New York: Cambridge University Press. Gunderson, Martin. (1979). “Threats and Coercion,” Canadian Journal of Philosophy, 9, 247–259. Haksar, Vinit. (1976). “Coercive Proposals,” Political Theory, 4, 65–79. Hay, Carol. (2011). “The Obligation to Resist Oppression,” Journal of Social Philosophy, 42, 21–45. Held, Virginia. (1972). “Coercion and Coercive Offers,” in J. Roland Pennock & John W. Chapman (eds.), Nomos XIV: Coercion, Chicago: Aldine-Atherton, Inc. Hetherington, Andrew. (1999). “The Real Distinction Between Threats and Offers,” Social Theory and Practice, 25, 211–242. Korsgaard, Christine M. (1996). Creating the Kingdom of Ends. Cambridge: Cambridge University Press. Kroon, Frederick. (1996). “Deterrence and the Fragility of Rationality,” Ethics, 106, 350–377. Lamond, Grant. (1996). “Coercion, Threats, and the Puzzle of Blackmail,” Chapter  10 in  A. P. Simester & A. T. H. Smith (eds.), Harm and Culpability (pp. 215–238). Oxford: Clarendon Press. ———. (2000). “The Coerciveness of Law,” Oxford Journal of Legal Studies, 20, 39–62. Lucas, J. R. (1966). The Principles of Politics. Oxford: Clarendon Press. Lyons, Daniel. (1975). “Welcome Threats and Coercive Offers,” Philosophy, 50, 425–436.

272

Responsibility and coercion McCloskey, H. J. (1980). “Coercion: Its Nature and Significance,” Southern Journal of Philosophy, 18, 335–352. McGregor, Joan. (1988–89). “Bargaining Advantages and Coercion in the Market,”  Philosophy Research Archives, 14, 23–50. Miller, David. (2010). “Why Immigration Controls Are Not Coercive: A  Reply to Arash Abizadeh,” Political Theory, 38, 111–120. Moody-Adams, M. M. (1994). “Culture, Responsibility, and Affected Ignorance,” Ethics, 104(2), 291–309. Murray, Michael J., & Dudrick, David F. (1995). “Are Coerced Acts Free?” American Philosophical Quarterly, 32, 118–123. Nozick, Robert. (1969). “Coercion,” in S. Morgenbesser, P. Suppes, & M. White (eds.), Philosophy, Science, and Method: Essays in Honor of Ernest Nagel (pp. 440–472). New York: St. Martin’s Press. ———. (1974). Anarchy, State, Utopia, New York: Basic Books. ———. (1981). Philosophical Explanations. Cambridge: Harvard University Press. O’Neill, Onora. (1991). “Which Are the Offers You Can’t Refuse?” Chapter 7 in R. G. Frey & Christopher Morris (eds.), Violence, Terrorism, and Justice (pp. 170–195). Cambridge: Cambridge University Press. ———. (2000). Bounds of Justice. Cambridge: Cambridge University Press. Pallikkathayil, Japa. (2011). “The Possibility of Choice: Three Accounts of the Problem with Coercion,” Philosophers’ Imprint, 11(16). Pennock, J. Roland, & Chapman, John W. (eds.). (1972). Nomos XIV: Coercion. Chicago: AldineAtherton, Inc. Raz, Joseph. (1975). Practical Reason and Norms. London: Hutchinson. ———. (1986). The Morality of Freedom. Oxford: Oxford University Press. Rhodes, Michael. (2002). Coercion: A Nonevaluative Approach. Amsterdam: Rodopi. Ripstein, Arthur. (2004). “Authority and Coercion,” Philosophy and Public Affairs, 32, 2–35. Ryan, Cheyney C. (1980). “The Normative Concept of Coercion,” Mind, 89, 481–498. Snyder, Jeremy. (2010). “Exploitation and Sweatshop Labor: Perspectives and Issues,” Business Ethics Quarterly, 20(2), 187–213. Stevens, Robert. (1988). “Coercive Offers,” Australasian Journal of Philosophy, 66, 83–95. Superson, Anita. (2005). “Deformed Desires and Informed Desire Tests,” Hypatia, 20, 109–126. Wertheimer, Alan. (1987). Coercion. Princeton: Princeton University Press. Williams, Bernard. (1973). “A Critique of Utilitarianism,” in J. J. C. Smart & Bernard Williams (eds.), Utilitarianism: For and Against. Cambridge: Cambridge University Press. Wood, A. (1995). “Exploitation,” Social Philosophy and Policy, 12, 136–158. Yaffe, Gideon. (2003). “Indoctrination, Coercion and Freedom of Will,” Philosophy and Phenomenological Research, 67, 335–356. Zimmerman, David. (1981). “Coercive Wage Offers,” Philosophy and Public Affairs, 10, 121–145. ———. (2002). “Taking Liberties: The Perils of ‘Moralizing’ Freedom and Coercion in Social Theory and Practice,” Social Theory and Practice, 28, 577–609. Zwolinski, M. (2007). “Sweatshops, Choice, and Exploitation,” Business Ethics Quarterly, 17(4), 689–727.

273

21 THE EPISTEMIC CONDITION OF MORAL RESPONSIBILITY Daniel J. Miller

0 Introduction While the contemporary philosophical literature is replete with discussion of the control or freedom required for moral responsibility, only more recently has substantial attention been devoted to the knowledge or awareness required, otherwise called the epistemic condition.1,2 This area of inquiry is rapidly expanding, as are the various positions within it. This chapter presents one way of carving up the territory and framing these positions, while highlighting advantages and challenges for each. It closes by sketching a novel approach that incorporates advantages of otherwise opposing positions on this topic. Although the epistemic condition is associated with awareness, it’s helpful to begin with an observation about ignorance. Competing explanations of this observation in turn have divergent implications for the epistemic condition.

1  When does ignorance excuse? Tamias is a bookkeeper for a seemingly innocuous shipping company, unaware that his work helps conceal the company’s human trafficking operation. Given his ignorance, is he to blame for his role in human trafficking? It depends. Compare two variants. Tamias1’s work behind the desk never reveals the slightest indication of the evils perpetrated by his employer. If so, it seems he’s off the hook. Tamias2 notices designations of items in his spreadsheets that seem suspicious. It occurs to him that his company may be smuggling something (though he wouldn’t dream of anything so horrific as the truth), but he pushes the thought out of his mind, preferring not to investigate any further. In that case, he’s plausibly blameworthy for complicity in human slavery – though perhaps not as blameworthy as if he were aware of the true nature of his work. Whether an agent is blameworthy for ignorant wrongdoing depends upon the status of their ignorance. Most theorists agree on the following: Blameless Ignorance Principle (BIP): If an agent is blameless for their ignorance then they are blameless for acting wrongly from that ignorance.3,4

DOI: 10.4324/9781003282242-30 274

The epistemic condition of moral responsibility

BIP implies that, since Tamias1 is blameless for his ignorance, he is blameless for his complicity in human trafficking. But since Tamias2 is (we may suppose) blameworthy for his ignorance, it doesn’t excuse him.5 BIP invites two questions. First, why does blameless ignorance excuse? Second, when is someone blameless for their ignorance? Some theorists answer these questions by appealing to the following principle: Reasonable Expectation (RE): An agent is blameworthy for something only if it was reasonable to expect the agent to avoid it.6,7 RE theorists standardly maintain that it’s reasonable to expect an agent to avoid something only if they have the capacities to recognize and respond to moral reasons as well as the opportunity to exercise these capacities to avoid the thing in question. Having this sort of opportunity in turn may require having access to evidence, exposure to alternative points of view, and adequate time to reflect upon the relevant features of one’s situation. A common motivation undergirding RE is that blameworthiness depends upon whether it would be fair to blame someone, and that it would be fair to blame someone for something only if they could have reasonably been expected to avoid it (Rosen 2003: 79; FitzPatrick 2008: 603). On RE, then, blameless ignorance excuses because it’s unreasonable to expect someone to have avoided behaving wrongly if they blamelessly believed that their behavior was permissible (Rosen 2003: 74–75). An agent is blameless for their ignorance if they lacked the capacity or opportunity to avoid being ignorant. This presents an explanation of the previous cases: while Tamias2 can reasonably be expected to have discovered the true nature of his work by following the evidence, this isn’t true of Tamias1. An alternative approach begins with the following principle: Quality of Will (QW): An agent is blameworthy for something if and only if it manifests a negative quality of will.8,9 QW theorists see blame as an appropriate response to an agent based on their quality of will, understood in terms of their regard or concern for others. Whereas good will involves moral concern for others (e.g., a desire to prevent harm, a belief that someone’s interests matter), a negative quality of will involves a lack of concern (e.g., indifference toward someone’s well-being, a desire to harm, a belief that someone’s interests are unimportant). On QW, blameless ignorance excuses because it precludes the manifestation of negative quality of will. Tamias1’s ignorance isn’t due to a lack of concern for the people trafficked by his company. Consequently, his ignorance and unwitting behavior cannot reflect any such attitude. In contrast, even if Tamias2 is unaware of his role in human slavery, his ignorance is due to a deficit of concern for possible wrongdoing (and those who might be harmed by it), and his continued work for the company reflects this. If so, QW implies that he’s blameworthy for his ignorance and unwitting wrongdoing. The fact that RE and QW often have overlapping implications can obscure their underlying differences. Sections 2 and 3 explore two dimensions of epistemic condition that serve as fault lines along which RE and QW diverge. The first concerns different kinds of ignorance. The second concerns what agents can be directly blameworthy for.

275

Daniel J. Miller

2  Two kinds of ignorance Sometimes ignorance of the wrongness of one’s behavior is due to ignorance of some feature that makes it wrong. This is called circumstantial ignorance. Tamias, for example, is ignorant that his work helps conceal human trafficking. A  doctor might fail to realize that prescribing a medication is wrong because they’re unaware of a contraindication. I might not realize that I shouldn’t share a story publicly because I wasn’t informed of its confidentiality. Other times agents are ignorant of the wrongness of their behavior because they are unaware that its wrong-making features (taken together) are wrong-making. This is called moral ignorance.10 Take an historical variant of our bookkeeper. Tamias3 is an ancient Greek who manages finances for a slave-trader. He’s fully aware of this, but given his cultural setting, he (like everyone else he knows) believes that slavery is morally permissible.11 Can his moral ignorance excuse him?

2.1  Moral ignorance and reasonable expectation If BIP is true, then agents like Tamias3 are excused for morally ignorant behavior if they’re blameless for their moral ignorance. And RE seems to imply that they can be. First, Tamias3 may lack opportunities to remedy his ignorance. He may, for example, be unlikely to encounter a dissenting opinion that would make him reconsider. Furthermore, even if he were to reflect upon his belief about slavery, his general framework of beliefs about rights and obligations may simply reinforce his ignorance. Similar considerations plausibly apply to countless racists, sexists, and warlords of the past. Given the ubiquitous and unquestioned acceptance of certain beliefs in their cultures, it’s unlikely such agents had the wherewithal or opportunity to avoid moral ignorance (Wolf 1987; Rosen 2003).12 If so, RE has the implication that Tamias3 and other historical agents are blameless both for their moral ignorance and unwitting wrongdoing. Contemporary instances of moral ignorance, however, are often not due to a lack of opportunities to avoid it. Take a 21st century racist who has regular exposure to articulate presentations of egalitarianism (and devastating critiques of racism) but who nevertheless refuses to reconsider his racist beliefs. Or consider a ruthless business executive who crushes her rivals through deception and force, rebuffing the efforts of those who try to confront her on the matter (“if you can’t stand the heat . . .” she snaps back).13 Lastly, consider complacent consumers who avoid watching persuasive documentaries on the evils of factory farming because they don’t think they need to listen to bleeding heart activists. The above agents arguably have the capacity and opportunities to eliminate their ignorance, even if they don’t believe they’re obligated to take these opportunities. And, if their failure to do so is the result of epistemic vices like arrogance, laziness, or stubbornness, this may hardly seem to excuse them. On these grounds, some argue that many morally ignorant agents can reasonably have been expected to have taken these opportunities, eliminated their moral ignorance, and thus avoided acting from it (FitzPatrick 2008: 605). Some RE theorists set a higher bar for reasonable expectation, maintaining that agents are blameworthy only if (at some point) they behaved contrary to their own belief about what they ought to do (Zimmerman 1997: 418; Rosen 2004: 307; Levy 2009: 735). Call this the akrasia requirement. One argument for this requirement is that it’s reasonable to expect an agent to φ only if the agent can φ rationally (rather than by chance or accident). 276

The epistemic condition of moral responsibility

Furthermore, the relevant sense of rationality concerning reasonable expectation is internalist: what’s rational for an agent is a function of their own beliefs and attitudes, rather than objective reasons. An agent can φ rationally in this sense only if they see a sufficient reason to. It follows that, if the ruthless business executive (or racist, or factory farming consumer) fails to see sufficient reason to reconsider their beliefs (and supposing they’re not blameworthy for this fact) then they cannot do so rationally, and thus cannot reasonably be expected to (Levy 2009: 735).14 If this argument is correct about the standards of reasonable expectation, then adopting RE may require a revision to our beliefs concerning how often agents are blameworthy for ignorant wrongdoing. We’ll revisit this concern in Section 3.

2.2  Moral ignorance and quality of will In contrast, QW theorists reject the view that moral ignorance can excuse in the way that circumstantial ignorance can (Harman 2011, 2022; Talbert 2013; Arpaly 2015). While circumstantial ignorance often precludes behavior from reflecting negative quality of will, moral ignorance seems to involve it.15 Consider that Tamias3 has no moral objection to the treatment of human beings as commodities, or that the business executive feels entitled to destroy the economic fortunes of her rivals. These agents’ attitudes themselves reflect a lack of concern for the rights and well-being of others. And if QW is correct, then it simply doesn’t matter whether they could have reasonably been expected to see any reason to refrain from their behavior – their negative quality of will is sufficient for blameworthiness for their moral ignorance and ignorant behavior. Indeed, QW might have a more sweeping implication. If an agent is aware of the wrongmaking features of their behavior (e.g., that it causes unnecessary suffering) yet fails to recognize that these features make their behavior morally wrong, this seems to reflect insufficient concern for what matters morally. If Tamias3 cared enough for the rights and interests of the slaves, then presumably he wouldn’t think slavery is permissible. If this can be generalized, then QW implies that morally ignorant wrongdoers of the past and present (racists, sexists, ruthless business executives, factory farming consumers, and more) are always blameworthy for their moral ignorance, and therefore that it cannot excuse them from morally ignorant wrongdoing (Harman 2011, 2022).16 As we’ll see more clearly in Section 3, adopting QW may force us to revise our beliefs about the requirements on blameworthiness for ignorance and ignorant wrongdoing.

3  Direct and derivative blameworthiness If Tamias2 is blameworthy for his unwitting role in human trafficking, that’s partly because he is blameworthy for something else: his ignorance. This introduces an important distinction. When an agent is blameworthy for something (at least partly) in virtue of being blameworthy for something else, their blameworthiness is derivative. An agent’s blameworthiness for something is direct when it is not derivative. Just as the issue of moral ignorance serves as a fault line in this literature, so too does the question of what agents can be directly blameworthy for. I take for granted the standard assumption that an agent is derivatively blameworthy for something only if they are directly blameworthy for something else (Rosen 2004: 299). Thus, the pervasiveness of blameworthiness for ignorant wrongdoing depends upon 277

Daniel J. Miller

whether agents can be directly blameworthy for it, and if not, what it takes to be derivatively blameworthy for ignorant wrongdoing.

3.1  Reasonable expectation 3.1.1  RE: actual awareness and the tracing strategy If the Blameless Ignorance Principle (BIP) is true, then blameworthiness for ignorant wrongdoing can’t be direct but instead derives from blameworthiness for the ignorance itself. Furthermore, many RE theorists hold that, since we lack direct control over what we believe, it’s unreasonable to expect someone simply to have or lack a belief. If so, RE implies that: NDBI: Agents are never directly blameworthy for their ignorance (Zimmerman 1997: 414– 415; Rosen 2004: 303; Clarke 2017: 236). If the reasoning behind NDBI is correct, agents can only be derivatively blameworthy for their ignorance, and in virtue of being directly blameworthy for something over which they have direct control: a prior action (or omission). It follows from the conjunction of BIP and NDBI that an agent can be directly blameworthy only for witting wrongdoing, and that blameworthiness for anything else must derive from instances of this.17 The “tracing strategy” aims to explain how this works. Recall Tamias2. He is blameworthy for his unwitting complicity in human trafficking because he is blameworthy for his ignorance, and he is in turn blameworthy for his ignorance only if it can be “traced back” to blameworthiness for some prior witting wrongdoing that resulted in it – in this case, his wittingly ignoring evidence of possible foul play.18 What Tamias2 is directly blameworthy for, then, is a witting failure to take the required steps to avoid ignorance. These “procedural epistemic obligations” can include gathering and considering evidence, reflecting upon the morally relevant features of one’s situation, asking for advice, and exposing oneself to alternative viewpoints (Rosen 2004: 301). Failures to fulfill these obligations are ubiquitous: a doctor failing to check their patient’s medical history, a parent ignoring safety instructions on their child’s toy, a spouse neglecting to set a reminder for their anniversary, a business executive stubbornly refusing to listen to objections to her ruthless behavior, and so on. A common feature of these examples is illustrative of a further epistemic requirement on the tracing strategy: in order to trace blameworthiness for ignorance and ignorant wrongdoing, they must have been reasonably foreseeable results of the prior wrongdoing. It’s reasonably foreseeable for a doctor, for example, that failing to check his patient’s chart might result in ignorance of a medical contraindication, and subsequently in a prescription that harms the patient. When something is not reasonably foreseeable for an agent, it’s unreasonable to expect them to foresee it, and thus they can’t be blameworthy for it (even if they are blameworthy for the wrongdoing that it results from).19,20 The foreseeability requirement is the target of an objection to the tracing strategy. Consider “Jeff the Jerk”, a teenager seeking social status who engages in behavior that gradually cultivates an insensitivity towards others and eventuates in a particular instance of unwitting rude conduct as an adult (Vargas 2005: 271). Though the particular instance of rude behavior was hardly foreseeable for teenage Jeff, adult Jeff still seems blameworthy for it. In response, some argue that what must be reasonably foreseeable are certain types of 278

The epistemic condition of moral responsibility

outcomes, not specific instances. Accordingly, teenage Jeff could have reasonably foreseen that behaving in certain ways might result in behaving insensitively on some later occasion. And though it wasn’t reasonably foreseeable for Tamias2 that neglecting evidence of possible wrongdoing might result in his unwittingly concealing human trafficking, it was reasonably foreseeable that it might result in unwittingly concealing wrongdoing more generally. If so, the tracing strategy can accommodate such cases (Fischer & Tognazzini 2009: 537–538). While the previous objection claims that reasonable foreseeability is too strong a requirement, a different objection claims that it’s too weak. Given the tracing theorist’s own commitments, reasonable foreseeability may be insufficient if it turns out that agents can be blameless for failing to foresee what was reasonably foreseeable for them (e.g., perhaps an agent can fulfill their procedural epistemic obligations without acquiring foresight, or be excused for failing to fulfill a procedural epistemic obligation). After all, if blameless ignorance excuses agents when acting from that ignorance, then it plausibly also excuses agents for bringing about consequences that they were blamelessly ignorant of. If so, then reasonable foreseeability is insufficient for satisfying the epistemic requirement on tracing, and actual foresight is required (Miller 2017). But since there are fewer cases in which ignorance (or ignorant wrongdoing) is actually foreseen, this stronger requirement may have the revisionary implication that we’re blameworthy for ignorant wrongdoing less often than we think.

3.1.2  A skeptical argument Tracing theorists standardly hold that all blameworthiness for ignorance or ignorant wrongdoing must be traced to an instance of witting wrongdoing. Suppose further, as Gideon Rosen (2004: 307) does, that it must be traced to an akratic behavior (recall the akrasia requirement from Section 2.1).21 If so, we can be justified in believing that an agent is blameworthy for unwitting wrongdoing only if we’re justified in believing that the agent has behaved akratically at some prior time. Rosen conjoins these commitments with a skeptical premise: one can never be justifiably confident on any particular occasion whether someone (including oneself) has in fact behaved akratically (307–310).22 To illustrate, suppose an individual believes that they shouldn’t lie to their partner, but in the moment convinces themselves that lying on this occasion is permissible (e.g., after all, it’s a fairly insignificant lie, their partner would be better off not knowing, etc.). When lying, then, they don’t act akratically. Or suppose I consider reasons for and against an action, but decide to act on the basis of considerations that favor it without ever coming to a conclusion about what the overall balance of considerations favored. If so, then when acting I  lack any belief about what I  ought to do, and so I don’t act akratically (Rosen 2004: 307). And sometimes I may only hold an implicit (non-occurrent) judgment about what I ought to do, and thus it’s opaque even to myself what I believe when acting. Even if I do act akratically, I’m not justified in believing I’ve done so. If considerations like these can apply to all cases of wrongdoing, then Rosen’s skeptical premise follows. But if so, and if all blameworthiness must ultimately be traced to akratic behavior, then a skeptical conclusion follows: we’re never justified on any particular occasion in believing that someone is blameworthy. Although QW theorists have responded to Rosen’s argument (see Section  3.2), some of the most influential responses are offered by RE theorists intent on showing that the 279

Daniel J. Miller

tracing strategy (or RE more generally) doesn’t yield such a radical conclusion. In what follows, I canvas three main avenues of response. First, an RE theorist might reject the skeptical premise. We often seem to have good evidence that we behave akratically. Feelings of guilt or shame during action, and attempts to conceal one’s behavior, seem to indicate akrasia (FitzPatrick 2008: 595–596). If so, we’re often justified in believing that the akrasia requirement has been met. Second, some RE theorists reject the akrasia requirement. Philip Robichaud (2014: 142ff) argues that some reasons are sufficient to make an action rational without making it obligatory. Thus, one can believe that it would be rational to listen to differing views without believing that one is obligated to do so. And, it’s plausible that agents can be reasonably expected to take such opportunities (thus fulfilling their procedural epistemic obligations) so long as they take themselves to have sufficient reason to do so.23 The view sketched in Section 2.1 offers another route to deny the akrasia requirement. While blameworthiness for certain instances of ignorant wrongdoing may derive from earlier acts, perhaps these earlier acts themselves needn’t be akratic. Perhaps the business executive can be blameworthy for her unwitting wrongdoing in virtue of being blameworthy for failing to take prior available opportunities to listen to others, even if these prior failures were also unwitting. If so, tracing may require a terminus that is either an instance of akratic wrongdoing or an unwitting failure to fulfill procedural epistemic obligations due to epistemic vices (FitzPatrick 2008: 609).24 Notably, both ways of rejecting the akrasia requirement increase the scope of prior acts we can plausibly trace to, and in turn increase instances of derivative blameworthiness for unwitting wrongdoing we can justifiably accept. The first two lines of response to Rosen’s skeptical argument accept some version of the tracing strategy. The third avenue of response rejects that ignorant wrongdoing must be traced to an earlier act at all, maintaining that agents can be directly blameworthy either for their ignorance or their ignorant wrongdoing. Both RE theorists and QW have developed versions of this response, which I discuss in the following two sections.

3.1.3  RE: capacitarianism Some RE theorists argue that ignorant wrongdoers often could have and should have been aware of some pertinent fact. They were capable, for example, of noticing some morally relevant feature of their behavior, or appreciating the application of some moral principle, or drawing an inference, even if they failed to exercise these capacities. On these grounds, such theorists argue, it’s reasonable to expect them to have known better and to have avoided unwitting wrongdoing (Sher 2009; Clarke 2017; Rudy-Hiller 2017). If so, then agents can be directly blameworthy either for their ignorance (contrary to NDBI) or ignorant wrongdoing (contrary to BIP).25 Consider George Sher’s case in which a distracted Alessandra forgets her dog, Sheba, in a hot car (2009: 24). To be blameworthy for this failure, Sher argues, Alessandra needn’t be at fault for any prior failure to ensure that she wouldn’t forget Sheba. What matters is rather that Alessandra could have and should have remembered Sheba (72ff).26 One clear advantage of this “capacitarian” view is that it captures our intuitions about blameworthiness in cases where it’s implausible that ignorance or ignorant wrongdoing is traceable to some prior action. However, our capacities are not completely reliable. Even conscientious agents sometimes fail to notice what they (in some sense) ought to, or forget a

280

The epistemic condition of moral responsibility

promise they should have remembered, and in general are subject to various mental “slips” (Rudy-Hiller 2019). Furthermore, failures to exercise these mental capacities needn’t reflect any underlying fault with the agent, and may be due primarily to bad luck. Alessandra’s forgetting Sheba needn’t be due to any prior wrongdoing, character flaw, or lack of care, but rather to an unexpected distraction. And it may seem unfair to blame someone for their ignorance or ignorant wrongdoing in such cases. If capacitarianism nevertheless implies that such agents are blameworthy, this may be problematic for the view.27

3.2  Quality of will If QW is correct, then an agent’s ignorance itself may reflect a negative quality of will, and therefore be something for which they are directly blameworthy. We’ve seen in Section 2.2 how this applies to moral ignorance. The same holds for circumstantial ignorance. If I forget my friend’s birthday, for example, this may reflect a lack of care for them (Smith 2005: 236–237). Supposing it does, QW implies that I’m directly blameworthy for my ignorance. Since QW allows that agents can be directly blameworthy for their ignorance, it also implies that blameworthiness for wrongdoing needn’t require awareness (either of the wrong-making features of an action or its moral status) at any point in its etiology. If so, then tracing will often be unnecessary, and Rosen’s skeptical conclusion is avoided. QW can also explain why some agents seem more blameworthy than others for similar behavior. Someone who bumps into us out of malice is intuitively more blameworthy than one who does so out of carelessness. Or, suppose that two doctors negligently fail to check their patient’s medical chart before prescribing medication. The first fails to check the chart while aware that this failure may result in the patient’s hospitalization or death, while the second fails to check the chart while only considering that this failure may result in some minor harm (e.g., a rash). While both doctors may be blameworthy for the patient’s subsequent hospitalization, it’s intuitive that the first is more blameworthy than the second (Miller 2019: 38–39). QW offers an explanation of these comparisons: the behavior of the first agent in each comparison manifests a worse quality of will, and therefore the first agent is plausibly more blameworthy for their behavior than the second. Perhaps the chief objection to QW is that it cannot accommodate the conviction that an agent is blameworthy only if it’s fair to blame them. To illustrate, consider four agents who believe that certain humans are fit for slavery. The first agent was raised in an egalitarian society but wittingly cultivated hateful and degrading beliefs about certain individuals that culminated in the belief that they belong in slavery. The second agent acquired the same belief, not by willful cultivation, but by being indoctrinated in a culture where the belief was unquestioned. The third agent began as an egalitarian but was altered via covert brain manipulation to hold this belief. The fourth agent was created instantaneously just a moment ago with this belief. QW implies that each of these agents are blameworthy for their belief. But this seems the wrong result. While agents who believe this are (in this respect) bad, it doesn’t immediately follow that it would be fair to blame them (Levy 2005; Miller 2014). But suppose we grant that they’re all blameworthy. Even so, QW has the further counterintuitive implication that they’re equally blameworthy for their ignorance, even though the last three had no reasonable opportunity to avoid their beliefs and (at least) the last two played no role whatsoever in acquiring them. Though QW captures our intuitions across a wide range of cases, these implications suggest that it may cast too wide a net.28

281

Daniel J. Miller

4  Sketching a hybrid view I’ll now put my cards on the table: I favor an RE approach that requires actual awareness for direct blameworthiness. But RE is more explanatorily powerful when conjoined with considerations taken from QW. On this hybrid view, reasonable expectation sets a threshold for blameworthiness. This preserves the connection between blameworthiness and fairness (and avoids the counterintuitive implications discussed in Section  3.2). An agent’s quality of will can then affect an agent’s degree of blameworthiness. This accommodates the intuition that an agent’s blameworthiness is sensitive to their regard for others. It’s reasonable to expect an agent to avoid wrongdoing only if they are aware of enough of its wrong-making features such that those features (taken together) make the behavior wrong (or if their ignorance of this is traceable to some prior failure with this awareness).29 But sometimes agents don’t realize just how wrong their behavior is. An agent might recognize, for example, that their behavior is impermissibly risky without realizing how much harm it risks (or how probable the harm is). If so, the agent may not be as blameworthy as they could be. Here quality of will comes into the picture: the greater the awareness that one behaves with, the greater potential for a worse quality of will manifested in one’s behavior. And the degree of quality of will manifested in one’s behavior can make a difference to their degree of blameworthiness.30 To illustrate, I offer first a case of direct blameworthiness, and then a case of derivative blameworthiness that involves tracing. When overhearing Bill mention his peanut allergy, Alan sees it as an opportunity to have a good laugh. He sneaks peanut butter into Bill’s sandwich, waiting to see if Bill’s face “blows up like a balloon” when he bites into it. Alan has no idea, however, that Bill’s allergy is deathly serious. Bill bites into the sandwich, has a severe reaction, and is rushed to the hospital (thankfully surviving).31 Alan is blameworthy for his action. But he’s less blameworthy than if he had known that the peanut butter might kill Bill and did so anyway. The hybrid view accounts for both facts. First, insofar as Alan is aware that his behavior will cause unjustified harm to Bill, the threshold condition is met, and (all else equal) Alan is directly blameworthy for harming Bill. Second, Alan’s behavior reflects a lack of concern for (and perhaps ill will toward) Bill, since he’s aware that his action will likely cause suffering and embarrassment. But if Alan did this while aware that it might kill Bill, that would manifest a more severely callous disregard for Bill’s life.32 Since Alan is only aware of a generic wrong-making feature of his behavior (that it will harm Bill in some way or other), his action cannot reflect a callous disregard for Bill with respect to the more specific and serious wrong-making feature (that it might kill Bill), and thus he’s less blameworthy for doing so than he could be. The hybrid view has a similar (though more complicated) application to cases of derivative blameworthiness. Return once more to Tamias2. Since he’s blameworthy for his ignorance, then (supposing he has no other excuse) he’s blameworthy for his ignorant wrongdoing. If the actual awareness version of RE is correct, then this is because he wittingly failed to fulfill his procedural epistemic obligation to follow the evidence. But it also seems clear that Tamias2 is less blameworthy than he would be if he continued to work with full awareness of the human trafficking. What explains this? On the hybrid view, Tamias’s greater awareness would have allowed for a greater manifestation of a negative quality of will – in particular, a lack of concern for those trafficked by his company. Since Tamias2 is in fact unaware of any wrongdoing, the scope of his awareness is limited to the vague possibility that something might be amiss (e.g., smuggling of some sort). Even so, he was aware 282

The epistemic condition of moral responsibility

at the time that ignoring the evidence might result in unwitting complicity in the company’s wrongdoing. In this way, his unwitting complicity still manifests a lack of concern, albeit indirectly (by way of his prior awareness of this possibility). But since at no point was he aware of the potential for complicity in the more specific and heinous wrong of human trafficking, his bookkeeping can’t manifest a degree of callous indifference with respect to this feature of his unwitting behavior.

5 Conclusion There are other ways to frame the literature on the epistemic condition.33 And certain views don’t fall neatly into the categories I’ve laid out.34 Furthermore, each of the views discussed here have more to be said in their defense, and the arguments canvassed here are targets of additional critiques I haven’t the space to discuss. For these reasons, I have provided references for further reading where it may be of interest. It’s my hope that this chapter offers readers one accessible pathway into this literature, and sufficient signposting to venture further in.35

Notes 1 The requisite awareness may not require knowledge (see note 17). 2 For historical treatments, see Aristotle (1999: III: 1) and Aquinas (2012: Q. 76). 3 I set aside cases of uncertainty about the moral permissibility of one’s behavior (Guerrero 2007). Agents who behave from ignorance (rather than simply while ignorant) would have behaved differently were they aware of the wrongness of their behavior (Rosen 2003: 62; Wieland 2017: 152, n. 8). 4 Theorists who accept BIP include Rosen (2003: 64, 2004: 300), Fischer and Tognazzini (2011: 390), Zimmerman (1997: 411), FitzPatrick (2008: 601–602), and Levy (2009: 741). Clarke (2017) offers reason to doubt BIP. 5 See Holly Smith 2017 for an argument that even blameworthy ignorance excuses. 6 Sometimes what’s reasonable to expect is that an agent avoid an omission to act by performing that action. Thus, it might be reasonable to expect Tamias2 to have avoided his failure to follow the evidence precisely by following that evidence. 7 RE theorists include Gideon Rosen (2003), William FitzPatrick (2008), Neil Levy (2009), John Fischer and Neal Tognazzini (2009), Rik Peels (2011), Randolph Clarke (2017), Philip Robichaud (2014), Elinor Mason (2015), Fernando Rudy-Hiller (2017), and Robert Hartman (2021). Some theorists may take reasonable expectation to also provide a sufficient condition for blameworthiness. One reason to doubt this concerns the possibility (discussed in Section 3.1.1) that reasonable foreseeability is insufficient for blameworthiness for action–consequences, a possibility that may also apply to reasonable expectation. 8 QW theorists include T M Scanlon (2008, 2015), Angela Smith (2005, 2008), Pamela Hieronymi (2008, 2014), Elizabeth Harman (2011, 2022), Matthew Talbert (2013, 2019), and Nomy Arpaly (2003, 2015). Gunnar Björnsson (2017) holds a related view. Theorists sometimes referred to as “attributionists” fall under QW insofar as an agent’s quality of will is attributable to them (Talbert 2019). Since QW is a necessary and sufficient condition, it may need to be amended to include a general capacity to recognize reasons; otherwise it may imply that non-human animals or very young children are often blameworthy (insofar as they manifest a lack of concern for the interests of others). 9 QW theorists typically maintain that agents can be blameworthy for their quality of will. To account for this, they could amend QW with a disjunction (“if and only if it is or manifests . . .”) or stipulate that an agent’s quality of will trivially manifests itself. 10 The terms “circumstantial ignorance” and “moral ignorance” may be misleading since they both involve ignorance that some behavior is morally wrong. Wieland (2017: 150) calls the latter kind “pure” moral ignorance.

283

Daniel J. Miller 1 This variation is modeled after a case offered by Rosen (2003: 64–65). 1 12 This claim is bolstered by Rudy-Hiller’s (2023) argument that the relevant capacities are socially constituted. See Moody-Adams (1994), however, for a challenge to the claim in question. 13 For a detailed discussion of such a case, see FitzPatrick (2008). 14 See Clarke (2017: 249–250) for a critique of this argument. 15 Hartford (2022) argues that, even on QW, epistemic difficulty can mitigate blameworthiness for moral ignorance. 16 See Wieland (2017), Johnson-King (2020), and Miller (2021) for arguments that moral ignorance doesn’t always reflect a negative quality of will, and thus that QW doesn’t have this implication. 17 By “witting wrongdoing” I  mean wrongdoing performed with awareness that the behavior is morally wrong. Some maintain that this awareness must be “occurrent” (i.e., consciously thought of) at the time of behavior (e.g., Zimmerman 1997: 421–422; Levy 2009: 736, n. 16), while others maintain that only dispositional awareness is required (Robichaud 2014: 148). Additionally, many theorists maintain that the relevant awareness needn’t involve knowledge (Rosen 2008: 595–596; Peels 2014: 493). Instead (justified) true beliefs may be sufficient to undergird reasonable expectation concerning behavior. 18 See King 2017 for an argument that the tracing strategy is often explanatorily unnecessary. 19 See Smith (1983: 505) for a case that motivates this requirement. 20 Most theorists understand this, not in terms of what a “reasonable person” would foresee, but instead what a particular agent (given their capacities and opportunities) can be reasonably expected to foresee (Sher 2009: 100ff; Miller 2017: 1565). 21 The akrasia requirement goes further because one might believe that an action is morally wrong and yet (for non-moral reasons) it is all-things-considered rationally permissible (Rosen 2004: 305–306). 22 One might reject the akrasia requirement and maintain that direct blameworthiness for some behavior only requires awareness that it’s morally wrong. But the skeptical argument could be amended accordingly in terms of witting wrongdoing instead of akrasia. 23 Even if this is correct, many morally ignorant agents might consider the possibility that they’re mistaken so unlikely as to not merit further reflection. 24 But consider the following challenge: if agents can be directly blameworthy for some unwitting failures (e.g., the business executive’s unwitting failure to listen to others), why can’t they be directly blameworthy for other unwitting failures (e.g., the business executive’s unwitting ruthless wrongdoing)? In other words, if this view is correct, why is tracing necessary at all in these cases? Thanks to Kyle Fritz for raising this concern. 25 Rudy-Hiller accepts BIP but rejects NDBI (2017: 413). Clarke is inclined to accept NDBI (2017: 236) but casts doubt upon BIP (2017: 238). 26 Sher adds a further condition: the agent’s ignorance (e.g., Alessandra’s failure to remember Sheba) is due to her own attitudes, dispositions, and traits (2009: 87). 27 Although capacitarians often maintain that such agents are blameworthy, Rudy-Hiller (2019) argues that capacitarianism needn’t have this implication. 28 The fundamental differences between RE and QW have led some to wonder whether these theorists are operating with different conceptions of responsibility (and blameworthiness) and are simply talking past each other. However, both RE theorists and QW theorists are concerned with the appropriateness of the reactive attitudes involved in blaming (e.g., resentment, indignation). Indeed, the debate between these theorists is less about what responses are appropriate for blameworthy agents and more about the requirements on the appropriateness of these responses (Nelkin 2016: 360; Talbert 2017: 18–19). 29 One might (plausibly) add that the agent must recognize these features as wrong-making and believe on this basis that the action is wrong. To be ecumenical, I leave this open. 30 This component of the hybrid view is developed in further in Miller (2019). 31 This is a variation on a sub-plot in an episode of Freaks and Geeks (“Chokin’ and Tokin’ ”). 32 Suppose that Alan would have behaved the same way even if he had been aware of this. The truth of this counterfactual might mean that Alan is particularly blameworthy for his quality of will. However, it arguably doesn’t make him more blameworthy for his action, since his ignorance (in the actual case) would preclude the manifestation of his quality of will in that action (Miller 2019: 35–36).

284

The epistemic condition of moral responsibility 3 For excellent alternatives see Talbert 2016 (Ch. 5) and Rudy-Hiller 2022. 3 34 Björnsson (2017) maintains that blameworthiness depends upon the quality of will that can be reasonably demanded of an agent, which in turn is sensitive to the agent’s capacities. 35 I am grateful to Randy Clarke, Gabriel De Marco, Kyle Fritz, Matt Talbert, and Max Kiener for helpful feedback on earlier drafts of this chapter. I would also like to thank participants who attended a workshop presentation of this chapter at Oxford University, and Jeffrey Hause for a thoughtful correspondence on that presentation. Finally, I  owe a special debt of gratitude to Heather, who cared for our June while I worked on this chapter.

Further reading For alternative introductions to this literature, see Fernando Rudy-Hiller’s article on the Epistemic Condition in the Stanford Encyclopedia of Philosophy, as well as Chapter 5 of Matthew Talbert’s Moral Responsibility. For two recent collections, see Philip Robichaud and Jan Willem Wieland, Responsibility: The Epistemic Condition, and Rik Peels, Perspectives on Ignorance from Moral and Social Philosophy.

References Aquinas, T. (2012). Summa Theologiae Prima Secundae, 71-114, trans Fr. L. Shapcote. Lander: The Aquinas Institute for the Study of Sacred Doctrine. Aristotle. (1999). Nichomachean Ethics, trans. and ed. T. Irwin. Indianapolis: Hackett Publishing Co. Arpaly, N. (2003). Unprincipled Virtue: An Inquiry into Moral Agency. New York: Oxford University Press. ———. (2015). “Huckleberry Finn Revisited: Inverse Akrasia and Moral Ignorance,” in R. Clarke, M. McKenna, & A. Smith (eds.), The Nature of Moral Responsibility: New Essays. New York: Oxford University Press. Björnsson, G. (2017). “Explaining (Away) the Epistemic Condition on Moral Responsibility,” in P. Robichaud & J. W. Wieland (eds.), Responsibility: The Epistemic Condition. Oxford: Oxford University Press. Clarke, R. (2017). “Ignorance, Revision, and Commonsense,” in P. Robichaud  & J. W. Wieland (eds.), Responsibility: The Epistemic Condition. Oxford: Oxford University Press. Fischer, J., & Tognazzini, N. (2009). “The Truth About Tracing,” Noûs, 43(3), 531–556. ———. (2011). “The Physiognomy of Responsibility,” Philosophy and Phenomenological Research, 82(2), 381–417. Fitzpatrick, W. J. (2008). “Moral Responsibility and Normative Ignorance: Answering a New Skeptical Challenge,” Ethics, 118(4), 589–613. Guerrero, A. A. (2007). “Don’t Know, Don’t Kill: Moral Ignorance, Culpability, and Caution,” Philosophical Studies, 136(1), 59–97. Harman, E. (2011). “Does Moral Ignorance Exculpate?” Ratio, 24(4), 443–468. ———. (2022). “Ethics Is Hard. What Follows?” in D. Nelkin & D. Pereboom (eds.), The Oxford Handbook of Moral Responsibility. Oxford: Oxford University Press. Hartford, A. (2022). “Difficulty and Quality of Will: Implications for Moral Ignorance,” Philosophical Explorations, 25(2), 141–158. Hartman, R. J. (2021). “Concomitant Ignorance Excuses from Moral Responsibility,” Thought, 10(1), 58–65. Hieronymi, P. (2008). “Responsibility for Believing,” Synthese, 161, 357–373. ———. (2014). “Reflection and Responsibility,” Philosophy and Public Affairs, 42(1), 3–41. Johnson-King, Z. (2020). “Don’t Know, Don’t Care?” Philosophical Studies, 177(2), 413–431. King, M. (2017). “Tracing the Epistemic Condition,” in P. Robichaud & J. W. Wieland (eds.), Responsibility: The Epistemic Condition. Oxford: Oxford University Press. Levy, N. (2005). “The Good the Bad and the Blameworthy,” Journal of Ethics and Social Philosophy, 1(2), 1–16. ———. (2009). “Culpable Ignorance and Moral Responsibility: A  Reply to FitzPatrick,” Ethics, 119(4), 729–741.

285

Daniel J. Miller Mason, E. (2015). “Moral Ignorance and Blameworthiness,” Philosophical Studies, 172(11), 3037–3057. Miller, D. J. (2014). “Answerability, Blameworthiness, and History,” Philosophia, 42(2), 469–486. ———. (2017). “Reasonable Foreseeability and Blameless Ignorance,” Philosophical Studies, 174(6), 1561–1581. ———. (2019). “Circumstantial Ignorance and Mitigated Blameworthiness,” Philosophical Explorations, 22(1), 33–43. ———. (2021). “Can Morally Ignorant Agents Care Enough?” Philosophical Explorations, 24(2), 155–173. Moody-Adams, M. M. (1994). “Culture, Responsibility, and Affected Ignorance,” Ethics, 104(2), 291–309. Nelkin, D. (2016). “Difficulty and Degrees of Moral Praiseworthiness and Blameworthiness,” Nous, 50(2), 356–378. Peels, R. (2011). “Tracing Culpable Ignorance,” Logos and Episteme, 2(4), 575–582. ———. (2014). “What Kind of Ignorance Excuses? Two Neglected Issues,” Philosophical Quarterly, 64(256), 478–496. Robichaud, P. (2014). “On Culpable Ignorance and Akrasia,” Ethics, 125(1), 137–151. Rosen, G. (2003). “Culpability and Ignorance,” Proceedings of the Aristotelian Society, 103(1), 61–84. ———. (2004). “Skepticism About Moral Responsibility,” Philosophical Perspectives, 18(1), 295–313. ———. (2008). “Kleinbart the Oblivious and Other Tales of Ignorance and Responsibility,” Journal of Philosophy, 105(10), 591–610. Rudy-Hiller, F. (2017). “A Capacitarian Account of Culpable Ignorance,” Pacific Philosophical Quarterly, 98, 398–426. ———. (2019). “Give People a Break: Slips and Moral Responsibility,” Philosophical Quarterly, 69(277), 721–740. ———. (2022). “The Epistemic Condition for Moral Responsibility,” The Stanford Encyclopedia of Philosophy (Winter 2022 ed.). https://plato.stanford.edu/archives/win2022/entries/ moral-responsibility-epistemic/. ———. (2023). “Moral Ignorance and the Social Nature of Responsible Agency,” Inquiry, 66, 821–848. Scanlon, T. M. (2008). Moral Dimensions: Permissibility, Meaning, Blame. Cambridge: Harvard University Press. ———. (2015). “Forms and Conditions of Responsibility,” in R. Clarke, M. McKenna, & A. M. Smith (eds.), The Nature of Moral Responsibility: New Essays. New York: Oxford University Press. Sher, G. (2009). Who Knew? Responsibility without Awareness. New York: Oxford University Press. Smith, A. (2005). “Responsibility for Attitudes: Activity and Passivity in Mental Life,” Ethics, 115, 236–271. ———. (2008). “Control, Responsibility, and Moral Assessment,” Philosophical Studies, 138, 367–392. Smith, H. (1983). “Culpable Ignorance,” Philosophical Review, 92(4), 543–571. ———. (2017). “Tracing Cases of Culpable Ignorance,” in R. Peels (ed.), Perspectives on Ignorance from Moral and Social Philosophy. New York: Routledge. Talbert, M. (2013). “Unwitting Wrongdoers and the Role of Moral Disagreement in Blame,” in D. Shoemaker (ed.), Oxford Studies in Agency and Responsibility: Volume 1. Oxford: Oxford University Press. ———. (2016). Moral Responsibility. Cambridge: Polity Press. ———. (2017). “Omission and Attribution Error,” in D. Nelkin & S. Rickless (eds.), The Ethics and Law of Omissions. New York: Oxford University Press. ———. (2019). “The Attributionist Approach to Moral Luck,” Midwest Studies in Philosophy, 43(1), 24–41. Vargas, M. (2005). “The Trouble with Tracing,” Midwest Studies in Philosophy, 29(1), 269–291. Wieland, J. W. (2017). “What’s Special About Moral Ignorance?” Ratio, 30(2), 149–164. 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. Zimmerman, M. (1997). “Moral Responsibility and Ignorance,” Ethics, 107(3), 410–426.

286

22 MORAL COMPETENCE AND MENTAL DISORDER Lubomira Radoilska

Introduction In this chapter, I explore moral competence as a central condition on moral responsibility. I distinguish two main conceptions. According to the first, a morally competent agent is someone who knows right from wrong. According to the second, a morally competent agent is someone who responds aptly to reasons. These two conceptions are not mutually exclusive; however, they merit separate treatment as they offer different insights on how and why moral competence might be compromised. This distinction is of particular relevance to us since the chapter critically examines a standard assumption, according to which whenever a mental disorder impacts moral competence, it decreases its scope (on the first conception) or precision (on the second). In close conversation with the memoir literatures on bipolar disorder, autism and schizophrenia, I argue that moral competence may also be affected by what looks like increases in scope or precision; moreover, neither impact  – decrease or increase  – necessarily undermines moral competence in and of itself; oftentimes, either could enhance it in a reliable way. Finally, by critically revisiting the ecological or scaffolding approach to moral responsibility, I show that moral competence is best understood as a practical way of knowing right from wrong embedded in daily routines and habits, and irreducible to propositional understanding or intellectual skills. This upshot bears on neighbouring debates about the epistemic condition on moral responsibility where the status of moral ignorance is hotly contested. As we shall see, the notion of moral competence can play the role of a heuristic device helping separate out instances where moral ignorance exculpates from others where, by contrast, it is the very object for which a person is called to account. The subsequent discussion assumes an interpersonal or socialised rather than a purely metaphysical approach to moral responsibility (e.g., Ceva  & Radoilska, 2018). On this approach, real-life practices of holding responsible and taking responsibility are interpreted as relevant to ascertaining the nature and scope of the concept.

287

DOI: 10.4324/9781003282242-31

Lubomira Radoilska

Why moral competence matters and how mental disorder might affect it One way to formulate the epistemic condition on moral responsibility is as follows: Moral responsibility requires knowledge of what one is doing. While this formulation has intuitive appeal, it also proves elusive on reflection. For instance, some authors have tried to fill it in by articulating the kinds of knowledge agents ought to have to be responsible for their actions. As Sliwa (2017: 127) puts it: “Our knowledge of the right and wrong thing to do is, in part, what determines whether we do the right or wrong thing intentionally. Moral responsibility inherits its epistemic condition from the epistemic condition on intentional action.” This line of reasoning ties closely together ignorance of what one is doing with ignorance of whether what one is doing is right or wrong. So, in paradigm cases, either kind of ignorance would disrupt – if not invalidate altogether – the ascription of moral responsibility for things done unawares. The so-called Parity Thesis (e.g., Rosen 2003; Levy 2011), according to which moral ignorance undermines moral responsibility in relevantly similar ways to factual ignorance is a logical conclusion of this approach. In other words, moral ignorance is an exculpating factor as much as factual ignorance is – provided that neither is the outcome of what Holly Smith (1983: 547) termed an initial ‘benighting act’: an agent chooses to not improve their knowledge or understanding when they should have done so, or even actively undertakes steps to avoid the required knowledge or understanding. Hence, any moral responsibility for actions done in ignorance, be it moral or factual, is ultimately reducible to that for prior intentional actions – making oneself ignorant, incompetent or incapacitated while knowing what one is doing. Most contributors to the discussion on the epistemic condition of responsibility would agree with this assessment of how factual ignorance works (cf. Robichaud  & Wieland 2017). Being unaware of relevant facts through no fault of one’s own  – as opposed to cases of negligence or recklessness where agents are under specific obligation to inform themselves of such facts – is a good ground for excuse (cf. Stark 2016). At the same time, however, the issue of whether and when agents can be responsible for their own moral ignorance is hotly contested. Unlike the initial approach we considered, there is another one which, by contrast, does not seek to align the epistemic condition of moral responsibility to that of intentional action. Instead, both knowing and ignoring whether what one is doing is right or wrong are taken to be core exercises of agency, for which moral responsibility can be aptly ascribed. On this alternative approach, our shared understanding of factual ignorance as a plausible ground for excuse would be a poor guide to ascertaining the impact of moral ignorance on moral responsibility. This view is known as the Asymmetry Thesis (e.g., Alvarez & Littlejohn 2017; Hartford 2019). Furthermore, recent developments in epistemology and, in particular, the epistemologies of ignorance (e.g., Sullivan & Tuana 2007) and vice epistemology (e.g., Tanesini 2021) have shed light on the various ways, in which ignorance can be culpable in and of itself, in the absence of a prior benighting act. And even though most contributors would not necessarily support the Asymmetry over the Parity Thesis, in regard to responsibility, moral ignorance is typically treated as a guide to factual ignorance rather than the other way around. In this context, the notion of moral competence becomes crucial. For it can offer a promising criterion for separating out instances of culpable, from instances of nonculpable and potentially exculpating moral ignorance. While the term ‘moral competence’ is not frequently used, the concept is right at the heart of current debates on both the nature and scope of moral knowledge and the conditions on moral responsibility. 288

Moral competence and mental disorder

A  closer look at the relevant literatures helps distinguish between two main conceptions. According to the first, a morally competent agent is someone who knows right from wrong (e.g., McGrath 2019). According to the second, a morally competent agent is someone who responds aptly to reasons (e.g., Ceva & Radoilska 2018). As outlined earlier, these two conceptions are not mutually exclusive; however, they merit separate treatment as they offer different insights on how and why moral competence might be compromised. For instance, if moral competence is about knowing right from wrong, the target notion is that of reliable detection. Getting it right here amounts to separating out morally acceptable courses of action from morally objectionable ones. Importantly, reliability in this respect is assessed without any reference to the reasons why an action is the right, as opposed to wrong, thing to do. For the relevant challenge is of breadth rather than depth: knowing right from wrong in whatever circumstances one might be called to action. By contrast, if moral competence is about responding appropriately to reasons, the target notion is that of adequate understanding. Getting it right here amounts to weighing reasons for and against eligible courses of action, and doing so in terms of potential justifications, explanations, or apologies that might be owed to others. For the relevant challenge is of depth rather than breadth: being able to give reasons for one’s actions that are reasons of the right kind (cf. Hieronymi 2005). So, while the first conception of moral competence is best understood in terms of scope, in response to the question “Is this a case in which one can tell right from wrong?” the second conception is best understood in terms of discernment, in response to the question “Could one grasp the reasons that make one’s actions right or wrong?” There is a long-standing trend in the literature on responsibility associating moral competence as discussed with sanity in the legal sense. Arguably, the best-known example of this trend can be found in Susan Wolf’s “Sanity and the Metaphysics of Responsibility” (1987). Recent discussions of the so-called insanity defence make the connection even more salient. Consider the following excerpts: The canonical case of legal insanity involves a defendant who forms the requisite mens rea for a crime but lack understanding of the legal and moral quality of her act – typically, that all things considered, the act is wrong. . .. Symptoms of a mental disorder can undermine a person’s capacity to be law-abiding at the time of a crime by causing a lack of moral knowledge; and the presence of a mental disorder signals to the court that the defendant is not culpable for this ignorance (Sifferd 2022: 113); [T]he rule of law presupposes creatures of a special kind; viz., those able to give and to comprehend reasons for their actions; persons able to defend their actions by supplying justifications which are, after all, reasons sanctioned by the law. The concept of mens rea is but the technical term for what is otherwise one of the core and common sense assumptions of legal responsibility and liability. (Robinson 2013: 19) These excerpts are noteworthy for a couple of reasons. First, each of them taps into a congenial conception of moral competence, the former, that of knowing right from wrong, the 289

Lubomira Radoilska

latter, that of responding aptly to reasons. And second, they both imply that insanity in the legal sense compromises the moral competence they consider as fundamental in virtue of a decrease in either scope or discernment below a required threshold. In some ways, the second feature might seem unsurprising. After all, mental disorders and psychological impairments more broadly are routinely treated as pathologies undermining of personal agency rather than mere obstacles to it (see, however, Radoilska 2022). Expanding on this underlying assumption, it would be tempting to argue that when mental disorders impede moral competence, the issue is either of reduced scope, for a person gets to tell right from wrong in fewer cases, if any, or of reduced discernment, for they do not fully, if at all, grasp the reasons which make their actions right or wrong. Could this be the full picture? In the following, we shall look into some fist-person narratives of people diagnosed with severe mental disorders and psychological impairments that demonstrate a more complex relationship to moral competence. To anticipate, in some cases, moral competence seems to be affected in virtue of what looks like an increase of scope or discernment: in the first instance, the difference between right and wrong becomes focal in almost every situation; in the second, the call of reasons for action, albeit of the right kind, is experienced as overwhelming. Taking up a helpful distinction first introduced in Petrolini (2020), I  shall refer to these kinds of increase as ‘hyper-agency’; conversely, I shall use the term ‘hypo-agency’ for the more familiar challenges to moral competence that derive from reductions in regard to scope or discernment.

Hypo- and hyper-agency in first-person narratives of people diagnosed with bipolar disorder, autism, and schizophrenia According to Petrolini (2020), both hypo- and hyper-agency have two dimensions: first, self-attribution, which is binary, and second, a sense of agency or control, which, by contrast, is scalar. For instance, the experience of auditory verbal hallucinations is a case of hypo-agency characterised by incorrect self-attribution and a diminished sense of agency. By contrast, the experience of pathological guilt is a case of hyper-agency where incorrect self-attribution comes along with an inflated sense of agency. While the epistemic focus of the original distinction is welcome as it turns the spotlight on the agent’s self-understanding and the impact of mental disorders on it, it is important to note that such conditions might also affect a person’s agential capacities more directly, independently of whether these effects are correctly assessed by the agent or not. For instance, autism might impede a person’s ability to ‘read the room’ by using non-verbal cues in a way that is often taken for granted by neuro-typical agents. Clearly, this is an instance of hypo-agency; yet, affected agents may correctly assess themselves both in terms of self-attribution and sense of control. As a young person with autism cited in Wing (1992: 131) reports: “People give themselves messages with their eyes but I do not know what they are saying.” To reflect this potential duality, the following discussion will distinguish between felt and manifest instances of hypo- and hyper-agency relevant to moral competence. The analysis will focus on three first-person narratives of people diagnosed with bipolar disorder, autism and schizophrenia: Kay R. Jamison, An Unquiet Mind (1996), Temple Grandin, Thinking in Pictures (2006) and Elyn Saks, The Centre Cannot Hold (2015); however, to address the worry that these memoirs are of exceptional personalities and, therefore, unrepresentative, some additional first-person accounts will also be drawn upon. 290

Moral competence and mental disorder

The rationale for this methodological choice is fourfold. First, as we saw earlier, knowing what one is doing is a defining feature of intentional actions. It also entails knowing that it is oneself doing it. Recent discussions have shown that this distinctive self-knowledge has a similar role to play beyond intentional actions, in wider exercises of personal and therefore responsible, as opposed to merely causal agency (cf. Radoilska 2017). So, if we are to explore the impact of hypo- and hyper-agency on personal agency and more specifically a person’s moral competence, this cannot be achieved without at the same time assessing the impact they have on agential self-knowledge. Second, heeding first-person testimonies of this kind is even more urgent when they come from marginalised or stigmatised groups and individuals as the case is with severe mental disorders and psychological impairments. As the growing literature on epistemic injustice forcefully demonstrates, ignoring the epistemic value of such testimonies impedes the creation of much-needed conceptual resources, depriving everyone – the epistemically disadvantaged as well as the epistemically privileged – from a chance to develop a more accurate self-understanding as agents (cf. Kidd et al. 2017). Third, all three conditions – bipolar disorder, autism and schizophrenia – might significantly affect moral competence. Bipolar disorder induces dramatic changes in one’s evaluative stance during episodes of mania or depression. Autism impacts the understanding of other people’s mental states thus making it difficult to distinguish approval from disapproval, consent from dissent. Schizophrenia is characterised by vivid and persistent hallucinations and delusions that are virtually impossible to distinguish from reality. Fourth and final, as the memoirs demonstrate, both felt and manifest hypo- and, respectively, hyper-agency are consistent with an intense commitment to maintaining one’s moral competence. Consider the following excerpts from Jamison’s memoir of what it is like to pursue a highly successful academic and clinical career, becoming a world-leading specialist of bipolar disorder, while at the same time living with this condition: The chaotic visual impact upon entering the room reflected the higgledy-piggledy, pixilated collection of electric lobes that only a few weeks earlier had constituted my manic brain. . . . There was a bill from a taxidermist in The Plains, Virginia, for example, for a stuffed fox that I for some reason had felt I desperately needed. I had loved animals all my life, had at one point wanted to be a veterinarian: How on earth could I have bought a dead animal? . . . How could I have so directly contributed to killing one? I was appalled at the grisly nature of my purchase, disgusted with myself (Jamison 1996: 75–76); The pharmacist, having just filled my first prescription for lithium, had smiled knowingly as he rang up the sale for my snakebite kits and other absurd, useless, and bizarre purchases. I knew what he was thinking and, in the benevolence of my expansive mood, could appreciate the humour. He, unlike me, however, appeared to be completely unaware of the life-threatening problem created by rattlesnakes in the San Fernando Valley. God had chosen me, apparently only me, to alert the world to the wild proliferation of killer snakes in the Promised Land. . . . I had also come up with a plan to alert the Lost Angeles Times to the danger. I was, however, far too manic to tie my thoughts together into a coherent plan. (ibid.) 291

Lubomira Radoilska

On these occasions, moral competence is clearly compromised. In the former excerpt, this is due to hypo-agency – first manifest (during the manic episode) and then felt (in its aftermath). It is best described as decrease in scope, in line with the first conception of moral competence we discussed. In the latter, by contrast, this is due to felt hyperagency, best described as superfluous increase of discernment, in line with the second conception. Yet, responsibility is ascribed both by the agent herself as evidenced by the earlier excerpt and wider society, e.g., Jamison (1996: 75): “money spent while manic doesn’t fit into the Internal Revenue Service concept of medical expense or business loss.” However, as Jamison’s memoir shows, the impact of hyper-agency on moral competence is not uniformly adverse. In fact, such positive contributions of manifest as opposed to felt hyper-agency in the context of bipolar disorder are extensively documented and widely acknowledged. To give an example outside of Jamison’s memoir: At our heights we may find ourselves capable of creating music, art, words, and inventions which touch people’s souls and shape the course of history. . . . We’re making the effort to stay balanced and grounded so we can use our powers to make the world we live on better, more beautiful, and way more interesting. (The Icarus Project 2013) Moral competence here is by no means compromised. Instead, it appears greatly enhanced by manifest hyper-agency both in terms of scope and discernment, coupled with lucid self-understanding. Let us now look into possible impacts of hypo- and hyper-agency in the context of autism. The following two excerpts come from a memoir by Temple Grandin, a successful academic and businessperson diagnosed with the condition: The work I do is emotionally difficult for many people, and I am often asked how I can care about animals and be involved in slaughtering them. Perhaps because I am less emotional than other people, it is easier for me to face the idea of death. I live each day as if I will die tomorrow. This motivates me to accomplish many worthwhile things, because I have learnt not to fear death and to accept my own mortality. . . . However, I am not just an unfeeling observer; I have a sensory empathy for the cattle. . . . My goal is to reduce suffering and to improve the way farm animals are treated (Grandin 2006: 94); There are situations where “normal” people have a horrific lack of empathy. Some of this lack of empathy is beyond my comprehension. Time after time I read about a company that is in financial trouble and they need to ask the workers to take a cut in pay. The workers agree to a pay cut, but the chairman of the board gives himself a bonus. . . . For me to have empathy is to visually put myself in the other person’s place. I can really empathise with a laid-off worker because I can visualise his family sitting at the dining room table trying to figure out how the bills will get paid. If the worker fails to pay the mortgage he will lose his house. (Grandin 2006: 97–99)

292

Moral competence and mental disorder

Arguably, the first excerpt presents a case of felt and manifest hypo-agency as decrease in scope with respect to emotions. The second, by contrast, demonstrates felt and manifest hyper-agency as increase of discernment in terms of empathy. This is particularly important as it belies a popular preconception that people with autism are unable to truly empathise with others due to emotional deficits in a way that affects their moral competence (cf. Sacks 1995). In a similar vein, Jim Sinclair (1992: 299) explains: “Being autistic does not mean being uncaring.” His testimony speaks of manifest hyper-agency as increase of discernment with respect to interpersonal relationships: Because I don’t need people in my life, I’m free, as nonautistic people can never be free, to want other people in my life. Because I don’t need relationships with anyone, I’m free to choose a relationship with someone – not because I need a relationship but because I like that person. . . . When I make a connection it’s special because I don’t have to do it, but I choose to do it. (Sinclair 1992: 300) In all three instances, both hypo- and hyper-agency contribute positively to moral competence instead of undermining it; moreover, they do so in a non-accidental way: hypoagency with respect to emotions operates as ‘noise reduction’, thus enabling a sustained constructive attitude to the suffering of fellow beings while hyper-agency with respect to interpersonal relationships supports authenticity and mutual respect (see also Sinclair 1992: 301–302). Non-accidental positive contributions to moral competence through felt and manifest hypo- and hyper-agency with respect to scope and discernment are also documented by Elyn Saks, a leading expert on mental competence in law, diagnosed with schizophrenia. For instance: I understood early on that going to the mental wards sometimes set me off emotionally – it probably aroused my own dependency needs, as well as my anger at how I had been treated when I was being held in the hospital. But I was convinced I understood more than most people did . . . about what it was like to be the helpless patient in that bed, or the terrified patient in four-point restraint. . . . While preparing my Note, I spoke to one mental health professional then on the Yale faculty. “Wouldn’t you agree that being restrained is incredibly degrading?” I asked. “Not to mention painful. And frightening.” The professor looked at me in a knowing way. “You don’t really understand,” he said kindly. “These people are different from you and me. It doesn’t affect them the way it would affect us.” If only he knew, I  thought to myself. . . . My work had made a difference. It helped another attorney and it helped patients who were no different from me. No different at all. (Saks 2015: 211–213) Taking stock of these consistently positive contributions is especially significant given what Saks (2015: 330–331) rightly refers to as ‘the mythology that fuels the stigma: that schizophrenics are violent and threatening’. Importantly, however, these contributions tag onto patently disturbing and disruptive instances of felt hypo- and hyper-agency that arguably undermine moral competence in terms of scope (first passage shortly) and discernment

293

Lubomira Radoilska

(second passage), both clearly echoed in the wider memoir literature on schizophrenia (cf. Caroll 2017; Weiner 2018): thoughts crashed into my mind like a fusillade of rocks someone (or something) was hurtling at me – fierce, angry, jagged around the edges, and uncontrollable. I could not bear them, I did not know how to defend myself against them (Saks 2015: 83); [T]here was a cluster of news stories about a workplace shooting. . . . Could I do that? Have I done that? Am I a mass murderer? Am I him? Did I shoot those people? Was the wrong person killed? It haunted me for weeks, worrying that somehow I’d had a hand in the carnage. Was the wrong person accused? Should I go to the police and confess? I’m evil. (Saks 2015: 119) To recap the discussion so far: we critically examined the standard assumption that severe mental disorders and psychological impairments undermine moral competence in terms of hypo-agency, a decrease in scope or precision below a minimum threshold. Looking at first-person accounts of people diagnosed with bipolar disorder, autism and schizophrenia, we discovered the following: 1. Moral competence can also be affected by ‘hyper-agency’, that is, increase in scope or precision above an expected range. 2. Both hypo- and hyper-agency can be manifest as opposed to merely felt. 3. There are instances where manifest hyper- as well as hypo-agency enhance moral competence in a non-accidental way. On the complex picture that emerges, neither felt nor manifest hypo- or hyper-agency undercut moral competence in and of themselves. The fact that more often than not they both operate and are experienced as impediments to moral competence in the context of severe mental disorders or psychological impairments may have to do with the interpersonal and institutional contexts in which this competence is exercised. This is the guiding hypothesis that we shall consider in the third and final section of this chapter.

Moral competence in context: scaffolding, un-scaffolding and partial re-scaffolding of one’s own agency At first blush, our hypothesis might look like a straightforward implication of the so-called ecological approach to moral responsibility (e.g., Vargas 2018; McGeer  & Pettit 2015). For this approach highlights the constitutive role that social interactions play for supporting individuals in their capacity of responsible agents. As McGeer and Pettit (2015: 175) put it: “Your capacity to respond to the reasons, then, will be fixed in place, not just by how you are in yourself, but by the audience-exposed environment in which you operate; it will have an ecological character.” In other words, moral competence is poorly understood if treated as a psychological capacity that individual agents possess independently of the interpersonal contexts in which they exercise it. For sensitivity to real-life audiences is a core dimension of it, as much as (direct) sensitivity to reasons: “the practice of holding one 294

Moral competence and mental disorder

another responsible is of immense importance in human life, providing for a sort of mutual scaffolding or capacitation and enabling us to lift our performance to a level we might not otherwise have attained” (ibid:187). Clearly, the notion of social scaffolding here is meant to drive home that interpersonal interactions where agents are challenged and called to answer for themselves help them maintain and hone their moral competence: sensitivity to audiences always strengthens sensitivity to reasons: no matter how sensitive I  take you to be to the reasons relevant in the situation, I  assume that you are also sensitive to me, as to an authorized audience: you are disposed in light of my manifest expectations as to how you will perform to become even more sensitive to the requirements of the relevant reasons . . . I speak as someone invested in your proving to be responsive and as someone who thinks I can help to make you responsive. (ibid: 178) This optimising picture is partly echoed by the memoirs we considered: “When you are really crazy, respect is like a lifeline someone’s throwing you. Catch this and maybe you won’t drown,” writes Saks (2015: 80) about being asked to give reasons for her actions and her preferences during episodes of severe psychosis. In a similar vein, Grandin (2006: 114) recounts multiple experiences of effective scaffolding, including the following: When a new manager took over the Arizona New Ranchman, I did not realise that he thought I was weird, and I was in danger of being fired. . . . My pal Susan saw the warning signs and she helped me assemble a portfolio of all my articles. After the manager saw how many good articles I had written, he gave me a raise. On other occasions, however, social scaffolding through sensitivity to real-life audiences proves disempowering and toxic, e.g., Sinclair (1992: 301): I met someone who offered to teach me what I needed to know. He was a doctoral student in special education who worked with developmentally disabled people. . . . He abused me. . . . He told me it was my fault. When I told his faculty adviser about it, the professor said that this was friendship, that it was something I needed. These occasions bear close analogy to cases of moral insecurity due to systemic deprivation and discrimination (Kennett & Wolfendale 2019). According to Kennett and Wolfendale, this is a widespread phenomenon where the sense of one’s agency and long-term perspective are adversely impacted since there is no reliable link between one’s efforts and outcomes. That is to say, planning and self-control do not typically pay off for morally insecure agents since they and their life projects are not treated as important enough by the wider society. At the same time, however, morally insecure agents are also subject to stigmatising moral condemnation, as, for instance, when extreme poverty is attributed to lack of initiative or poor work ethic. Drawing on the resources introduced earlier, we can see that moral insecurity points to a distorted and harmful social scaffolding which consists of inappropriate manifest hyperagency expectations – “If only you would try harder!” – coupled with unfair hypo-agency accusations – “You failed yet again, while others manage just fine!” 295

Lubomira Radoilska

Clearly, in such debilitating contexts, the social scaffolding available can only jeopardise one’s moral competence. Instead, deliberate un-scaffolding as insensitivity to real-life audiences appears to be the only chance – albeit precarious – for maintaining it. This experience – liberating and isolating at once – is widely documented in the memoirs we introduced earlier. In their context, un-scaffolding one’s agency often begins by questioning the limitations a diagnosis of schizophrenia, bipolar disorder or autism sets out in terms inability to, e.g., live independently (Saks 2015), have a professional career (Jamison 1996), build significant relationships and even learn to drive (Grandin 2006; Sinclair 1992). This resistant and creative dimension of moral competence is not easily accounted for by the ecological approach as it currently stands. For it interprets responsiveness to reasons, respectively moral competence as a set of skills that is strongly dependent on ongoing feedback from real-life audiences (cf. McGeer 2019: 311–313). Some authors, e.g., Kennett and Wolfendale (2019), take this difficulty as an indication that we should abandon the underlying project of socialising moral responsibility altogether and approach it instead through a metaphysical lens. This solution, however, would leave us with no account of how and why un-scaffolding one’s agency can be apt and even successful in debilitating contexts, yet subpar at best in more supportive settings. Expanding on Gilbert Ryle’s essay “On Forgetting the Difference between Right and Wrong” (1958), I would like to propose an alternative that retains the ambition of socialising moral responsibility but questions the complete and ongoing dependence of morally competent agents from their – at times inadequate – real-life audiences. Ryle’s starting point is the observation that such statements as: “I once knew the difference between right and wrong but have now forgotten it” should strike us as absurd. According to Ryle, this reaction tells us something important about the nature of moral knowledge, namely, its conceptual link to caring about right and wrong revealed in how we feel and act in light of the distinction between the two. Caring here is presented as a complex and embodied disposition, over and above the propositional grasp of moral norms at the root of the two initial conceptions of moral competence we considered or the skillbased responsiveness to reasons the ecological approach puts forward instead. So, on Ryle’s account, if one can no longer tell the difference between right and wrong, this can only be a case of ceasing to care and thus becoming a different person rather than forgetting: The use of ‘forget’ for the loss of information or technical abilities, and its nonuse for the secessions of caring, may go with another difference. If I have ceased to enjoy bridge, or come to admire Picasso, then I have changed. But, if I have forgotten a date or become rusty in my Latin, I do not think of this as a change in me, but rather as a diminution of my equipment. (Ryle 2009 [1958]: 401) Building on Ryle’s notion of caring, we can venture that some routines and habits of caring might be constitutive of moral competence over and above the skills highlighted by the ecological interpretation. Being irreducible to deliberate rules or policies, these routines and habits are meant to function as go-to modes of engaging with one’s context and its challenges, especially when environmental feedback from real-life audiences becomes unreliable. This makes them well-suited to the re-scaffolding, albeit partial and temporary, of one’s responsible agency in the aftermath of a morally competent social un-scaffolding. 296

Moral competence and mental disorder

Conclusion In this chapter, I  fleshed out the notion of moral competence with reference to the epistemic condition on moral responsibility. Thus, I  argued against a standard approach to the relationship between moral competence and severe mental disorders and psychological impairments, stating that whenever such conditions impact on moral competence, they do so by decreasing its scope or precision. Looking closely at first-person accounts of people diagnosed with bipolar disorder, autism and schizophrenia, I showed that moral competence can also be affected by increases in scope or precision; moreover, such increases as well as decreases do not always undermine moral competence; instead, they often enhance it in a reliable way. To make room for this resilient and creative dimension of moral competence emerging from the first-person accounts we considered, I revisited recent work on socialising moral responsibility. The upshot is a novel understanding of moral competence as practical knowledge embedded in personal habits and daily routines, beyond the capacity to merely tell right from wrong.

Acknowledgments I am grateful to the Handbook Editor, Maximilian Kiener, an anonymous reviewer, the participants in the Handbook Workshop on ‘The Conditions and Challenges of Responsibility’ and audiences at the Universities of Geneva and Roma Tre for their helpful feedback. Work towards this chapter was supported by a SNSF grant, Ref. IZSER0-216875/1.

Further reading Hutchison, K. et al. (eds.). (2018). Social Dimensions of Moral Responsibility. Oxford: Oxford University Press explores key challenges to theorising moral responsibility arising from non-ideal social contexts. King, M., & May, J. (eds.). (2022). Agency in Mental Disorder: Philosophical Perspectives. Oxford: Oxford University Press is a collection of new essays offering empirically informed philosophical discussions of responsibility in the context of mental disorder. Radoilska, L. (ed.). (2022). Autonomy and Mental Disorder. Oxford: Oxford University Press brings together interdisciplinary insights on the norms and value of autonomy in the context of mental disorder, with reference to moral responsibility and moral standing. Shoemaker, D. (2015). Responsibility from the Margins. Oxford: Oxford University Press develops a comprehensive theory of moral responsibility with particular emphasis on specific mental disorders and psychological impediments. Wolf, S. (1987). “Sanity and the Metaphysics of Responsibility,” in F. D. Schoeman (ed.), Responsibility, Character, and the Emotions: New Essays in Moral Psychology. Cambridge: Cambridge University Press is an important paper expanding on the implicit connection between moral incompetence and insanity in the legal sense.

References Alvarez, M., & Littlejohn, C. (2017). “When Ignorance Is No Excuse,” in P. Robichaud & W. Wieland (eds.), Responsibility: The Epistemic Condition. Oxford: Oxford University Press. Caroll, D. W. (2017). “Severely Schizophrenic and Successful? Yes, It’s Possible!” Schizophrenia Bulletin, 43, 1151–1152. Ceva, E., & Radoilska, L. (2018). “Responsibility for Reason-Giving: The Case of Individual Tainted Reasoning in Systemic Corruption,” Ethical Theory and Moral Practice, 21, 789–810. Grandin, T. (2006). Thinking in Pictures. New York: Vintage Books.

297

Lubomira Radoilska Hartford, A. (2019). “Moral and Factual Ignorance: A Quality of Will Parity,” Ethical Theory and Moral Practice, 22, 1087–1102. Hieronymi, P. (2005). “The Wrong Kind of Reason,” Journal of Philosophy, 102, 437–457. The Icarus Project. (2013). “Navigating the Space Between Brilliance an Madness,” accessed November 30, 2022, from www.theicarusproject.net. Jamison, K. R. (1996). An Unquiet Mind. London: Picador. Kennett, J.,  & Wolfendale, J. (2019). “Self-Control and Moral Security,” in D. Shoemaker (ed.), Oxford Studies in Agency and Responsibility, vol. 6. Oxford: Oxford University Press. Kidd, I. J., Medina, J.,  & Pohlhaus, G. Jr. (eds.). (2017). The Routledge Handbook of Epistemic Injustice. New York: Routledge. Levy, N. (2011). Hard Luck. Oxford: Oxford University Press. McGeer, V. (2019). “Scaffolding Agency: A Proleptic Account of the Reactive Attitudes,” European Journal of Philosophy, 27, 301–323. McGeer, V.,  & Pettit, P. (2015). “The Hard Problem of Responsibility,” in D. Shoemaker (ed.), Oxford Studies in Agency and Responsibility, vol. 3. Oxford: Oxford University Press. McGrath, S. (2019). Moral Knowledge. Oxford: Oxford University Press. Petrolini, V. (2020). “Too Much or Too Little? Disorders of Agency on a Spectrum,” European Journal of Analytic Philosophy, 16, 79–99. Radoilska, L. (2017). “Aiming at the Truth and Aiming at Success,” Philosophical Explorations 20(Supplement 1), 111–126. ———. (2022). “Pathologies of Agency,” in L. Ferrero (ed.), The Routledge Handbook of the Philosophy of Agency. London: Routledge. Robichaud, P.,  & Wieland, W. (eds.). (2017). Responsibility: The Epistemic Condition. Oxford: Oxford University Press. Robinson, D. (2013). “The Insanity Defence as a History of Mental Disorder,” in K. W. M. Fulford et  al. (eds.), The Oxford Handbook of Philosophy and Psychiatry. Oxford: Oxford University Press. Rosen, G. (2003). “Culpability and Ignorance,” Proceedings of the Aristotelian Society, 103(1), 61–84. Ryle, G. (2009 [1958]). “On Forgetting the Difference between Right and Wrong,” in G. Ryle (ed.), Collected Essays: 1929–1968, vol. 2. London: Routledge. Sacks, O. (1995). An Anthropologist on Mars, London: Picador. Saks, E. (2015). The Centre Cannot Hold. New York: Hachette. Sifferd, K. (2022). “Legal Insanity and Moral Knowledge: Why Is a Lack of Moral Knowledge Related to a Mental Illness Exculpatory?” in M. King & J. May (eds.), Agency in Mental Disorder: Philosophical Perspectives. Oxford: Oxford University Press. Sinclair, J. (1992). “Bridging the Gaps: An Inside-Out View of Autism (Or, Do You Know What I Don’t Know?),” in E. Shopler & G. Mesibov (eds.), High-Functioning Individuals with Autism. New York and London: Plenum Press. Sliwa, P. (2017). “On Knowing What’s Right and Being Responsible for It,” in P. Robichaud  & W. Wieland (eds.), Responsibility: The Epistemic Condition. Oxford: Oxford University Press. Smith, H. (1983). “Culpable Ignorance,” The Philosophical Review, 92, 543–571. Stark, F. (2016). Culpable Carelessness. Cambridge: Cambridge University Press. Sullivan, S., & Tuana, N. (eds.). (2007). Race and Epistemologies of Ignorance. Albany, NY: SUNY Press. Tanesini, A. (2021). The Mismeasure of the Self: A  Study in Vice Epistemology. Oxford: Oxford University Press. Vargas, M. (2018). “The Social Constitution of Agency and Responsibility: Oppression, Politics, and Moral Ecology,” in K. Hutchison et al. (eds.), Social Dimensions of Moral Responsibility. Oxford: Oxford University Press. Weiner, S. (2018). “The Details in Schizophrenia,” Schizophrenia Bulletin, 44, 707–709. Wing, L. (1992). “Manifestations of Social Problems in High-Functioning Autistic People,” in E. Shopler & G. Mesibov (eds.), High-Functioning Individuals with Autism. New York and London: Plenum Press. Wolf, S. (1987). “Sanity and the Metaphysics of Responsibility,” in F. D. Schoeman (ed.), Responsibility, Character, and the Emotions: New Essays in Moral Psychology, Cambridge: Cambridge University Press.

298

23 EXCUSE, CAPACITY, AND CONVENTION David Owens

Much attention has been paid to how the rules, mores or conventions of our society shape the content of our rights and obligations. It is obvious that social context exercises some influence over what we are obliged to do for others, as well as on what we can expect others to do for us. Legal conventions impose a duty to drive on the left in some countries and on the right in others; social rules ensure that children must be offered a place in the family home until marriage in Southern but not in Northern Europe. Here I shall be making two assumptions about these matters. First, I  assume that there are conventional rights and obligations, that at least some of our rights exist and our obligations bind because they are socially recognised. Second, I assume that no rights exist and no obligations bind simply because they are socially recognised. To be valid or binding, a socially recognised right or obligation must also have some value or serve some interest. How this all works will be left open.1 In this chapter I’m concerned with another way in which convention shapes our normative situation. Whether we are to blame for a wrongful act is a function not only of what we are obliged to do – of the applicable standard of conduct – but also of how culpable we are for failing to do it, i.e., of the applicable standard of culpability. There are considerations – excuses  – which reduce or eliminate culpability for wrongdoing and so shield us from blame without casting doubt on the wrongfulness of our conduct. Some of these excuses are epistemic, appealing to the agent’s ignorance of the character of their action. Others are motivational, highlighting the impact of things like fear, anger and exhaustion on our agency. Though I’m hopeful that what I say could be extended to excuses as such, here we’ll be discussing the motivational excuses alone. In the final section, I urge that convention helps to determine what constitutes a legitimate (motivational) excuse, at least where we are dealing with standards of conduct which are themselves conventional. Excuses for the violation of pre-conventional standards of conduct are a topic for another day. Some may find this idea surprising. For them responsibility is a function of how much control we have over an action, what we know about the likely consequences of that action, and how exactly our action brings the outcome about (e.g., whether it happened by omission). Don’t these conditions settle our culpability regardless of whether their significance is socially recognised and regardless of whether the standard of conduct at issue is 299

DOI: 10.4324/9781003282242-32

David Owens

conventional? I shall argue that the extent of our responsibility turns on how much one is obliged to know and how much control one is obliged to exercise over one’s action, requirements specified in the standard of culpability. And it would be surprising if convention had an impact on what we are required to do but no influence at all over the norms that determine whether we are culpable for their violation. As we shall see, convention matters in both ways and for related reasons. A standard of culpability governs what is often called ‘moral responsibility’. Responsibility as such involves being held to valid norms (social or otherwise) and praised or criticised for how well you conform to them. Moral responsibility (as I construe it) is concerned with deontic normativity and so our focus is on people’s vulnerability to a specific form of criticism, namely the blame reactions – guilt, indignation, resentment – and with how excuses shield you from them (though, as we shall see, an excuse need not block them entirely). ‘Moral responsibility’ should also be distinguished from legal responsibility, i.e., from liability to various forms of sanction or penalty. I’m sympathetic to the idea that punishment involves a public condemnation and I shall draw freely on the writings of criminal law theorists but the precise relationship between legal liability and blame for wrongdoing is left unexplored.

Excuse Let’s begin with a typology of considerations which shield you from blame. I shall distinguish exemptions, justifications and excuses.2 The standard examples of exemptions fall into at least two categories. There are cases of automatism where your general ability to act intentionally is compromised. Having kicked you in the shins I might explain that this was a reflex movement due to my suffering an epileptic seizure. Standards of conduct do not apply to my kick because I am incapable of being guided them. Epilepsy seems like a clear case, sleepwalking, hypnotic suggestion less so, but we shan’t try to draw the boundaries of automatism here. One’s general ability to act intentionally may also be compromised by infancy or insanity. The remaining class of alleged exemptions I’ll call compulsions. These include actions which manifest phobias, addictions, and obsessions. Compulsive action like obsessive handwashing is clearly intentional for the agent well knows what they are doing and why they are doing it (to lower the tiny risk of infection). Yet, it is said, they should not be blamed where they are incapable of doing otherwise. I mention compulsions mainly to distinguish them from the excuses that concern me here. Obsessives, addicts and so forth are deemed abnormal, occupying some diagnostic category defined by operational criteria and methods of treatment. That is not how it is with excuses like duress, provocation, and stress, excuses society makes available to all who find themselves in coercive, provocative, stressful, etc., situations and without the benefit of any medical diagnosis. We now turn to justification and excuse. In this chapter I shall focus on these forms of exculpation as they apply to intentional action, setting aside cases in which the agent commits a wrong unintentionally through ignorance or error. It is usually said that a justification demonstrates that what you did was the right thing to do because you had sufficient reason for doing it, as when you kill someone in self-defence. Killing as such is wrong but it is not wrong to kill someone in self-defence and so you are not to blame. I accept this characterization with two caveats. First, I  want to allow for the possibility of justified wrongdoing, of situations in which the right thing to do is to commit a wrong. Here (unlike 300

Excuse, capacity, and convention

self-defence) I can’t act without compunction because I do commit a wrong but I should not be blamed for my wrongdoing because I do the right thing, and so, in that regard, my justification exculpates.3 Second, I wish to leave it open what counts as ‘having’ a justification. Some would argue that being subject to a deadly assault does nothing to justify your killing the assailant unless you kill them because they are threatening you. For these writers, the only reasons that can justify your action are the reasons on which you act; the existence of further reasons which could have justified what you did had you acted on them is irrelevant. On a more liberal view, you are justified in killing provided a good case can be made for it and regardless of whether that was the basis on which you acted.4 I shall leave it open whether exculpation requires one to act on the justification but I do maintain that an excuse must motivate the act it excuses and must do so in a specific way. Rationalists about responsibility regard justification and exemption as the only exculpations. For them the way to shield yourself from blame is either to offer a justification or else to demonstrate that what you did was not under your control. Take the bank teller with a gun to their head. Likely they are not to blame for handing over the bank’s money to the robber because compliance is the lesser evil: the threat justifies the action. Alternatively, perhaps the sight of a gun was so terrifying that the teller automatically handed over the cash, like someone who reflexively leaps off the road and into a passing pedestrian to avoid an oncoming truck. In such cases of ‘diminished responsibility’ the teller is exempt.5 Other writers recognise a third shield against blame, neither justification nor exemption but excuse. Take fear. The chief engineer at a nuclear power station co-operates with a plot to extort money (or other concessions) from the city authorities by threatening to blow up the power station because the extortionists are holding her children hostage. If the engineer is shielded from blame, it is likely not in either of the ways so far considered. This isn’t a case of exemption: co-operating successfully with the extortionists requires careful planning and cool deliberation perhaps over an extended period, all of which is a paradigm of sane, intentional activity. Consequently, the engineer is still held to various standards of conduct (e.g., not doing more than the extortionists demand) and will be blamed for their violation. Furthermore, though it is perfectly reasonable for the engineer to fear for her family, the threats of the extortionists do not justify her cooperation. Unlike our bank teller, she should regard what she is being asked to do as gravely wrong and feel remorse for having done it. Nevertheless, the threats reduce her blameworthiness (Owens 2012: 41–42).6 Many would decline to blame her at all whilst others would modulate rather than extinguish their indignation. Our examples suggest (a) that a threat can function either as an excuse (duress) or as a justification (lesser evil) and (b) that where it functions as an excuse, such a threat can render blame inapt without justifying the action it excuses.7 What is required for a threat to function as an excuse? For one thing, our engineer must fear for her family. She need not be consumed by terror but the threat exculpates as it does only because of her presumed emotional connection with the people being held hostage. On the other hand, our engineer must not cooperate because she expects to be excused, because she won’t be vulnerable to blame. It mustn’t be the case that she gives in to the threats of the extortionists when her fear of blame would otherwise lead her to resist because she thinks those threats will render blame inappropriate. By thinking in this way, she deprives her fear of its power to excuse.8 When I say that you mustn’t give in because your excuse renders you invulnerable to blame, I am using ‘vulnerable’ in a normative rather than a non-normative sense. There is nothing paradoxical or self-undermining about considering the probability of actual blame 301

David Owens

in your deliberations. If the consequences of being blamed would be serious and the probability of actual blame is high, the fact that your excuse would significantly lower that probability might be highly relevant to what you ought to do. In this case the prospect of being blamed is operating as part of a justification and could enter into your deliberations alongside the object of fear (i.e., the threats of the extortionists). All this is certainly coherent and sometimes sensible.9 Not so being motivated by the idea that your excuse reduces your culpability, a culpability independent of the probability that you’ll actually be blamed. Here what supposedly moves you is not the thought that, given the excuse, blame (together with its adverse consequences) is less likely but rather the idea that, given the excuse, any blame that occurs would not be appropriate. We here encounter a crucial difference between excuses and justifications. The normative force of a reason for action is in no way undermined when your action is motivated by the thought that the reason will justify the action. Our bank clerk can get themselves to hand over the money by reminding themselves that life matters more than property. By contrast, the normative force of an excuse is sapped where your action is motivated by the thought that, though it doesn’t measure up as a justification, it is enough to excuse your action. To put the point another way, the idea that X constitutes a sufficient reason for A-ing can play – and on some views must play – a role in your practical deliberation, in your deliberations about what to do, whilst the idea that X constitutes a sufficient excuse for A-ing cannot. The excusatory force of your excuse must not move you.10 What holds for fear is equally true of anger and exhaustion. This insult won’t shield me from blame where I retaliate only because I imagine it will also excuse me and this remains so even if I feel a rage that would have excused me had I retaliated from anger alone.11 The standard of culpability which defines these and other excuses is meant to guide the assessment of action (and the conduct of the assessors) without guiding the actions assessed.12 Why do those seeking to shield themselves from blame generally prefer being justified to being excused? The answers may seem obvious – doing right is better than doing wrong – but there is more to it than that. Resorting to an excuse compromises one’s self-respect in a way that wrongdoing as such does not: it is shameful to offer an excuse rather than a justification. But why? One might well feel shame at a loss of self-control, yet unlike pleas of exemption, both justifications and excuses represent the wrongful conduct as intentional and therefore as, in that sense, under the agent’s control. Nevertheless there is a form of control which you possess only where your motives justify your actions and not when they excuse them. Most agents wish to be in control of whether they are vulnerable to apt criticism. Both justification and excuse establish that a deed is blameless in virtue of certain considerations. Where these considerations justify, the agent can control whether what they do is blameworthy by attending to the considerations that justify it and so make it acceptable in that way. Where these considerations excuse, the agent cannot control whether what they do is blameworthy by attending to the considerations which excuse it and so make it acceptable in that way.13 Since the standards of culpability are not meant to guide action, the acceptability of their actions qua excused escapes the agent’s control and with it their status as a decent agent.

Incapacity I have claimed that moral responsibility is a function of two variables, of standards of conduct which determine what we are obliged to do and of standards of culpability which 302

Excuse, capacity, and convention

determine when we are to blame for violating a standard of conduct. Why should there be two distinct standards here? Clearly, we need standards of conduct (i.e., rules) to guide our action. We also need to apply these standards only to those capable of being guided by such rules, exempting others from blame. But why acknowledge a further category of excuses, of considerations meant to influence our assessment of deliberate violations without guiding the action assessed? The most common answer is that excuses reflect a form of incapacity that prevents people from being guided by the relevant standard of conduct without preventing them from being guided at all. When helping the extortionists, our engineer deliberately violates her obligations yet, it is said, she shouldn’t be blamed for this because fear for her family makes her incapable of complying with the relevant standard of conduct.14 Similarly for those overcome by anger, etc.15 This capacity theory can offer to explain why attempts to deliberately take advantage of an excuse are self-undermining. An agent who does wrong because they think fear, anger, exhaustion or stress will excuse them is not entitled to the excuse because they are able to do the right thing and so lack the incapacity on which the excuse they appeal to depends. Still, the theory strikes me as untenable.16 First, the relevant notion of capacity is obscure. I am indeed literally incapable of controlling my epileptic seizure: I can’t I stop it however hard I try.17 But what does it mean for our engineer to be incapable of facing down the extortionists when she deliberately acts out of fear? It isn’t a matter of what has actually happened: some people don’t cooperate in similar circumstances and this engineer might have refused herself on a previous occasion. Nor can it be a matter of what she would have done in other circumstances (e.g., if only her own life had been at risk): how would that help us settle whether her giving into the current threat constitutes culpable weakness of will? Suppose we somehow knew everything there is to know about what our engineer (and people like her) actually did or would do in various circumstances and from what motives they would act. How does that fix what they are capable of doing in the sense relevant to culpability?18 Now suppose one’s incapacities can somehow be inferred from facts about what one has done or would do in various circumstances. Would such incapacities explain familiar standards of culpability? There are several grounds for doubt. First, those standards generally rule out motives like lust, greed and ambition – a thief who claimed to have been overcome by temptation would be laughed out of court  – yet I  see no reason to believe that such motives are generally weaker than those appearing on the standard list of excuses.19 Second, the notion of ‘incapacity’ draws a line between people unable to live up to the standard of conduct who should be excused and everyone else who should not but in both law and life, we recognise partial excuses that reduce blameworthiness without eliminating it altogether (Austin 1979: 177). One might accommodate partial excuses by supplementing all or nothing incapacities with a graded notion of difficulty, claiming that (within certain limits) we are excused for wrongdoing in proportion to the difficulty of compliance. It is true that standards of culpability are often sensitive to the difficulty of compliance, but the measure of difficulty here is not the motivational capacities of the agent before us. Our motivational capacities depend on our temperament, upbringing, environment and so will vary widely from one person to another. Any standard of culpability which tracked such peculiarities would be highly individualized (Gardner 2007: 136–137). Rather, convention creates a common standard of difficulty, often tied to a social role. Relative to that standard, weaker characters are just expected to try harder. 303

David Owens

The capacity theorist is right to distinguish the issue of whether we have done wrong from the issue of whether we are culpable for doing it. Yet, far from it being the case that our standards of culpability reflect our pre-normative psychological capacities, I propose it is the other way around: these elusive capacities are a projection of our standards of culpability.20 To put it another way, the relevant notion of capacity is normative. Normative claims are frequently expressed in the language of capacities. For example, ‘he couldn’t treat his own brother like that’ is one way saying that it would be wrong to do so. I propose that exculpating fear, anger, fatigue, etc., are those forms of fear, anger, and fatigue we are not expected to avoid or overcome. How would my proposal remedy the previously noted defects of the capacity theory? If the question of incapacity is a normative one, it should come as no surprise that facts about what we actually do or would do cannot settle it. And valuable norms of culpability are likely to discriminate between motives on grounds of content, facilitating the work of the good (e.g., feelings protective of your family or of your personal dignity), not the bad (greed, etc.).21 Furthermore, we’d expect those norms to reduce appropriate blame in some circumstances whilst blocking it altogether in others. Given that our standards of conduct recognise degrees of wrongfulness – registering how far people’s behavior departs from the relevant standard  – why shouldn’t our standards of culpability register not just whether but how far people fall short of the ideal of strength of character they set up? Finally, if the norms in question are social conventions, they won’t be tailored to the motivational peculiarities of each individual.

Convention We are looking to explain how an excuse can render it inappropriate to blame an agent without rendering what they did any the less wrongful. Were all of the previous objections overcome, I doubt that the capacity theory could account for the existence of excuses so understood. The incapacities postulated by the capacity theory resemble focused exemptions, depriving you of the ability to live up a given standard of conduct without depriving you of all control over what you do. But if our engineer literally can’t face down the extortionists, why think that she thereby does wrong and should feel any more compunction than an epileptic? Leaving the capacity theory behind us we need another account of why standards of conduct diverge from standards of culpability, an account which makes better sense of the idea that wrongdoing might not be blameworthy. Were blame no more than a judgement of wrongdoing, one might worry that excusing someone must involve a failure to adequately recognise the wrongfulness of what they did. Most writers agree that blaming someone means going beyond that judgement: blame has been interpreted as a distancing of the accuser from the wrongdoer, as a type of anger or even as a form of punishment.22 Whatever its nature, provided blame involves something more than a mere cognitive registration of wrongdoing, a gap opens up between the question of what wrong was done and the question of how we ought to react. It might for various reasons be inappropriate to get worked up about wrongdoing – perhaps you have better things to worry about given how long ago it all was or you’ve committed the same offense yourself or you should really be forgiving the wrongdoer to preserve your friendship or the wrongdoing is someone else’s business – and the point holds however one thinks of blame. Consequently, adequate recognition of wrongdoing need not involve blame (though it may require other reactions). 304

Excuse, capacity, and convention

Still many will feel that there is a categorical difference between all of the above ways of deflecting blame and the exculpation furnished by an excuse. Questions of standing, the need for forgiveness, the passage of time and so forth turn on the relation between a particular blamer and the object of their blame (Scanlon 2008: 145–147). By contrast, when I claim an excuse, aren’t I seeking to deflect blame for this act from anyone, however placed? It may be regarded as a strength of the capacity theory that it appeals to an intrinsic, context-independent feature of the agent to establish that they are blameless (Hart 2008: 42–43). Must any alternative account of culpability share this virtue? At least where the standards of conduct in question are conventional, it is not so clear. The bindingness of such a standard is a function of the social context and we must expect the same to be true of the associated standard of culpability. So just because the standard transcends the relation between specific individuals, we should not infer that it rests on a context-independent feature of the agent in question. Standards of conduct bind us all where we all have an interest in that standard’s authority being recognised by the agents around us. Now a standard of conduct is not recognised solely in compliance; it is recognised also in the breach as when the perpetrator feels remorseful and bystanders blame them. A standard of culpability tells us where blame is needed to give the relevant standard of conduct adequate recognition. Our bank teller will hand over the money reluctantly but without either feeling guilt or experiencing blame. It is otherwise in the case of our engineer. Knowing that what she does is gravely wrong, she will comply with the demands of the extortionists only with great compunction. Since her fear is functioning as an excuse rather than as a justification, guilt is expected and, in her case at least, may constitute adequate recognition of her wrongdoing.23 And once an obligation has been adequately recognised, any reason we had to further recognise it by blaming the violator is much diminished and may be altogether eliminated. Their guilt pre-empts our blame.24 It will now be asked what constitutes ‘adequate recognition’ of a standard of conduct. How much guilt or blame should greet a given delict? When should the delict be partially or completely excused? When should apt blame be publicly expressed and if so how and by whom? My answer is: whatever recognition is required to ensure that the standard of conduct in question registers the interest which grounds its bindingness. The required recognition may vary from person to person depending on their relation to the act and to the agent. Frequently it will involve the wrongdoer having an attitude to what they did which differs from that of the people around them. In this light, the presence of an excuse now appears as a consideration which complicates an agent’s blameworthiness in the same general way as do issues of standing, the passage of time, need for forgiveness and so forth. You are not absolved from guilt at wrongdoing just because others don’t have the standing to blame you or ought to forgive you, or ought to be focused on other things, etc., and nor are you off the hook simply because you have an excuse. Though there is much more to be said about ‘adequate recognition’, we have done enough to explain why the excusatory force of an excuse cannot motivate us. Compliance with a standard is a crucial form of recognition and so our standards of culpability should not incentivise violation. But they would incentivise it if those seeking exculpation could be guided by them. In saying this, I do not mean to imply that the sole point of recognizing a standard by blaming non-compliance is to raise the probability of future compliance. On the contrary, where the standard binds us, such recognition (whether public or private) is also to be valued for its own sake as an affirmation of the rights of the wronged, the claims 305

David Owens

of society, etc.25 And that includes the recognition accorded to the standard when a perpetrator feels guilty about what they have done despite being excused. I’ll conclude by arguing that at least some standards of culpability are conventional, namely those imposed by the agent’s social role. Compare our engineer with a typical soldier.26 Both have obligations imposed upon them by laws, organizational regulations, and the commands of their superiors, obligations setting standards of conduct appropriate to their rather different roles. Both are required to face down certain dangers (the engineer mustn’t run away when the plant begins to malfunction). Nevertheless, the standards of culpability associated with these roles diverge. A gun to the head would excuse (without necessarily justifying) much if not all of the wrongdoing someone commits in their capacity as an engineer. A soldier is not so fortunate. Failure to obey an order to attack will hardly be excused on the grounds that the soldier rightly expects to be shot by the enemy if he complies (Garner 2016: 451–460). The whole point of that role and its duties would be frustrated if disobeying such an order even with extreme compunction constituted adequate recognition of its binding force. Blame (and more) is apt.27 An objector may respond that our standards of culpability are sensitive to non-normative facts like the agent’s maturity for the young are given more latitude when it comes to impulsive behavior. Indeed so. The rules defining a given social role will bind their occupants only in so far as they serve some worthy function and how well certain rules serve that function will depend on the psychology typical of the occupants of the roles.28 Given facts about the general run of humanity, it may be apt to blame violators only once they have undergone training. The best way to design the soldier role is to require the same conduct of all soldiers (training them all up) whilst cutting younger soldiers who violate the rules some slack (Gardner 2007: 129). But note that such rules may still bind us even where that role could have been better designed. For any number of reasons, it might be hard to transition from a system where immaturity is no excuse to a system in which it is. Where society really needs soldiers and the rules defining the role are otherwise good enough to serve that purpose, then (a) young soldiers currently have no excuse for panicking, etc., but (b) the system ought to be reformed so that immaturity, lack of experience, etc., becomes an excuse. Sound standards of conduct and culpability will often impose the same expectations on all who occupy the relevant role, though meeting those expectations may be much harder for some than for others: the fact that I’m especially fragile, prone to panic and so forth is not even a partial excuse. One hopes that those who find themselves in the military possess (or develop in training) a sense of duty which trumps the fear of death. But soldiers who fail are criticised for their cowardice or immaturity, not excused by it. To settle whether this soldier is culpable for giving in to a mortal threat, we need not ask whether this individual has actually developed the motivational psychology that their training was meant to induce. We need ask only what they ought to be like after training and that is determined by the conventions governing the occupants of their role.29

Notes 1 For further details, see (Owens 2022: Chapter One). 2 On exemptions see (Sliwa 2019: 70–71) (Wallace 1996: Chapter 6) and (Simester 2021: 17.5). Since his understanding of excuses differs from mine, Wallace puts reflexes into the excuse box (op cit. 140–1).

306

Excuse, capacity, and convention 3 On justified wrongdoing see (Gardner 2007: 77–82) (Simester 2021: Chapter  17, Sec. 6) and (Owens 2012: 89–95). Though cases of self-defence do not present us with conflicting obligations, other situations might. I have made two promises and it turns out that I cannot fulfil them both. The right thing to do here is to fulfil the more urgent promise but it doesn’t follow that I do no wrong. Here, even though I am justified and should not be blamed, I’ll feel some compunction about the breach. That is the grain of truth in the idea that conflicting obligations excuse conformity (Sliwa 2019: 54). 4 Compare (Simester 2021: Chapter 16) with (Gardner 2007: 98–101). Rule utilitarians and others distinguish the principles on which you should act from those that should be used to evaluate your action as either right or wrong. 5 In Scanlon’s view the bank teller is normally justified in handing over the money because they lack ‘eligible alternatives’ (Scanlon 1998: 279–280, 291–292), though in ‘rare’ cases, the clerk is driven out of their mind by the threat and so acts with diminished responsibility (Scanlon 1998: 400– 401). For Rationalists like Scanlon (together with (Wallace 1996: 143–147) and (Bruno 2022)), ‘excuses’ like duress must function as either justifications or exemptions. Gardner formulates a compromise position according to which fear, anger, etc., may excuse even where they do not justify the action provided these emotions are themselves justified (Gardner 2007: 86, 110–111, 134). I shan’t add this as a further condition on excuses. Perhaps even an unreasonable fear of flying sometimes excuses and unreasonable anger can excuse your shouting at least when you are tired or stressed. (Simester 2021: Chapter 19, Sec. 1.1). As Sliwa notes, exhaustion is neither reasonable nor unreasonable (Sliwa 2019: 62–63). 6 For a related example, see (Simester 2021: 17.3.1). Can a rationalist accommodate such examples by distinguishing the blameworthiness of an act from its permissibility? For Scanlon an act’s permissibility is a function of the reasons that apply (i.e., justifying reasons) whilst its blameworthiness is a function of the reasons the agent takes themselves to have (i.e., their motivating reasons) (Scanlon 2008: 124–125)). On this view, your culpability is a function of your motivating reasons. The problem for the rationalist is that our engineer is excused even though they neither have nor take themselves to have reasons adequate to justify threatening to destroy the city. 7 Similarly for an insult and other forms of provocation (Austin 1979: 177). 8 Thus, whilst you can propose a justification or cite an exemption prior to the act that requires exculpation, excuses cannot be offered in advance. 9 I leave it open whether and how such a motivation might affect the ‘moral worth’ of the action. 10 (Dan-Cohen 1984: 633–634, 671) (Gardner 2007: 138–139) and (Simester 2021: Chapter 17, Sec. 4) make the point with respect to criminal liability. 11 It is also true that the rage will not excuse me if I somehow arranged the provocation in order to benefit from the excuse it provides. 12 Since the function of a rule is to guide us, I call the conventions around excuse ‘standards’ of culpability, as in ‘standard heart rate’ (which may not be the statistically normal one). 13 Elsewhere I’ve dubbed this ‘reflective control’ (Owens 2017: 2–3) and I used the notion of a capacity to formulate it. I would now argue that the capacity for reflective control is often imputed to us on the basis of conventional standards of culpability. 14 (Hart 2008: 32). For a statement of the capacity theory as applied to duress see (Frankfurt 1988: 36–46). Sliwa argues that what all excuses have in common is that they render us ‘incapable’ of acting on our ‘present directed intentions’ (Sliwa 2019: 54–59) yet our engineer seems to retain that capacity. 15 The notion of incapacity can also be used to analyze the compulsions noted earlier – obsessions, addictions, and phobias – but here I’ll concentrate on passions that afflict us all. 16 Both my rejection of the capacity theory and my adoption of a normative alternative were inspired by (Watson 2004: 41–58 and Chapter 11) and (Gardner 2007: 124–128). 17 I might be able to exercise indirect control by taking my medicine, etc., and am responsible for so doing. 18 One could simply define ‘X is capable of A-ing’ as ‘X would do A in a range of circumstances C’ where these circumstances have some special practical relevance. That may be how we should understand the claim that certain glass is ‘unbreakable’ or certain mountain range ‘impassable’ but I doubt it would give us the notion that the capacity theory has in mind.

307

David Owens 19 (Frankfurt 1988: 42) says that someone in the grip of irresistible desire for fame may be coerced into doing something by the prospect of celebrity provided only that he does not wish to give in to this desire, yet such an offer would not normally be regarded as a good excuse. 20 Kant thought that capacities are imputed to us on the basis of standards of conduct (Kant 1996: 163–164). 21 On the other hand, in contexts in which things like ambition and acquisitiveness are to be encouraged (e.g., a business) they might be recognised as excuses for carelessness etc. 22 For discussion of the nature of blame, see (Wallace 1996: Chapter 3), (Scanlon 1998: 267–277) (Scanlon 2008: 122–179) and (Owens 2012: 27–34). 23 (Owens 2012: 41–42) and (Sliwa 2019: 63–71). 24 Where they have the excuse and feel no guilt, we blame them for acting without compunction. 25 (Scanlon 2003: 222–223). Thus I accept Hart’s critique of Bentham’s deterrence-based doctrine of excuse (Hart 2008: 43–44). 26 (Gardner 2007: 128–131, 167–172) focuses on how social role affects one’s culpability. The point is even more obvious when it comes to the level of skill one is expected to have (Gardner 2007: 109). In the latter case blame may focus on one’s failure to seek the relevant training but if one receives the training and remains incompetent, blame will persist. 27 What about those who occupy different social roles with conflicting standards of conduct and culpability? Here, the right reaction is often emotional ambivalence, as when we recognise justified wrongdoing. Questions of culpability qua soldier and qua friend can’t always be settled by invoking some underlying pre-conventional principle, one which stands behind these social roles and adjudicates their claims on us. 28 Whilst some of these social roles are undertaken voluntarily, others are not (e.g., conscripted soldier or sibling). Our choice often helps to explain why we are subject to certain conventional standards of both conduct and culpability but not always. 29 Thanks to Alex Sarch, Crescente Molina, Daniele Bruno and Jens Timmerman for written comments and to Sandy Steel, Steve Schaus, David Enoch, Massimo Renzo, Hasan Dindjer, Maria Alvarez, Eliot Michaelson, David Papineau, Tom Pink, James Stazicker, Sacha Golob, Beau Mount, Alex Kaiserman, Daniel Miller, Sophia Moreau, Jessica Brown, Laura Valentini, Ulrike Heuer, Scott Hershovitz, Ezequiel Monti, Julia Markovits, Andrew Simester and Max Kiener for discussion as well as to audiences at Oxford University, the University of St Andrews, Pompeu Fabra University Barcelona, the Analytical Legal Philosophy Conference, University College London, and Kings College London.

Further reading For a defence of rationalism about responsibility, see Chapters 5 and 6 of Wallace, J. (1996). Responsibility and the Moral Sentiments. Cambridge: Harvard University Press. Watson, G. (2004). Agency and Answerability. Oxford: Oxford University Press. Chapters 2–3 and 11 contains stimulating discussion of exemptions, justifications, and excuses, as does  Gardner, J.  (2007). Offences and Defences. Oxford: Oxford University Press Chapters 5–8.

References Austin, J. (1979). Philosophical Papers. Oxford: Oxford University Press. Bruno, D. (2022). “Being Fully Excused for Wrongdoing,” Philosophical Quarterly. DOI: 10.1111/ papq.12425. Dan-Cohen, M. (1984). “Decision Rules and Conduct Rules: An Acoustic Separation in Criminal Law,” Harvard Law Review, 97(3), 625–677. Frankfurt, H. (1988). The Importance of What We Care About. Cambridge: Cambridge University Press. Gardner, J. (2007). Offences and Defences. Oxford: Oxford University Press. Garner, R.-P. (2016). “A Tale of Two Moralities,” Journal of Philosophy, 113(9), 446–462. Hart, H. (2008). Punishment and Responsibility. Oxford: Oxford University Press. Kant, I. (1996). Practical Philosophy. Cambridge: Cambridge University Press.

308

Excuse, capacity, and convention Owens, D. (2012). Shaping the Normative Landscape. Oxford: Oxford University Press. ———. (2017). Normativity and Control. Oxford: Oxford University Press. ———. (2022). Bound by Convention. Oxford: Oxford University Press. Scanlon, T. (1998). What We Owe to Each Other. Cambridge: Harvard University Press. ——— (2003). The Difficulty of Tolerance. Cambridge: Cambridge University Press. ———. (2008). Moral Dimensions. Cambridge: Cambridge University Press. Simester, A. (2021). Fundamentals of Criminal Law. Oxford: Oxford University Press. Sliwa, P. (2019). “The Power of Excuses,” Philosophy and Public Affairs, 47(1), 37–71. Wallace, J. (1996). Responsibility and the Moral Sentiments. Cambridge: Harvard University Press. Watson, G. (2004). Agency and Answerability. Oxford: Oxford University Press.

309

PART 3

The practice of responsibility

Section 7: Being and holding responsible

24 BLAMING Leonhard Menges

1 Introduction Since about the mid 2000s, blame has become a widely discussed philosophical topic. While the early debate mostly focused on the relationship between blame and responsibility, blame as a phenomenon is now examined on its own and approached from many different directions. This has generated lively discussions and many insights. However, it has become hard to say how the different accounts of blame relate to each other. The aim of this chapter is to help keep track. It starts (section 2) with a brief overview of three families of theories of blame (see Tognazzini and Coates 2018 for a more detailed overview). In section 3 I will propose that an important, but often not explicitly discussed, question is “what should we want from a theory of blame?” and I will present two answers. In section 4, I will critically discuss the proposal that a theory of blame should give a unifying explanation of our everyday blame practices. In section 5, I will propose an alternative. The idea is to start with the most interesting moral and philosophical problems that are related to blame (think of free will problems). Philosophical theories of blame should primarily help us solve these issues, but do not need to provide unifying explanations.

2  What is blame? A (very) brief overview According to a family of recent theories that I will call functionalist theories, the nature of blame must be accounted for in terms of its function (see Wang 2021 for different functions of blame). Consider the question of what the heart is. A plausible answer starts by looking at what the heart does: it pumps blood. Based on this observation, one can say that the heart is whatever fulfills this function in the bodies of certain living things. Functionalist accounts of blame approach the question of what blame is by looking at what blame does in our minds and social practices. They describe that function and contend that blame is whatever fulfills it. Some theorists propose that the function of blame is to signal the blamer’s commitment to norms and values (Shoemaker & Vargas 2021). When I think that people violated a norm I accept, then I can signal my commitment in various ways, such as confronting them, coldly de-friending them on social media, or experiencing resentment

315

DOI: 10.4324/9781003282242-35

Leonhard Menges

without changing my open conduct – in the latter case I would also be the addressee of the signal. All this is blame, according to the view under consideration, because it signals my commitment to the norm. Other functionalist accounts argue that blame’s function is to express one’s protest (Smith 2013) or to initiate or sustain conversations about what the blamee did (McKenna 2013; Mason 2019, chap. 5). Whatever plays these roles, the idea goes, is blame because to blame just is to respond in a way that fulfills one of these functions. The second, older family of theories holds that the nature of blame cannot be accounted for in terms of its function. For lack of a better word, I call them non-functionalist theories. They approach the question of what blame is by looking for attitudes that are present in all blame responses and that explain why these responses count as blame. A classic nonfunctionalist account inspired by Peter Strawson (1962) identifies certain emotions as the core of blame (e.g., Wallace 1994, chaps. 2, 3, 2011; Wolf 2011; Menges 2017; Portmore 2022). It says that to blame someone is to have a certain emotion, such as a kind of anger in case of blaming others and guilt in case of self-blame. Another non-functionalist account contends that the core of blame is a certain desire-belief pair (e.g., Sher 2006; Arpaly & Schroeder 2014). On this view, to blame is to believe that someone acted badly and to desire that the person had acted better. The emotions and conduct that we often associate with blame must, then, be explained by reference to the desire–belief pair. A third family of views combines functionalist with non-functionalist ideas, which is why I call them mixed theories. According to one such theory developed by Tim Scanlon (2008, 2013), blame partly consists in judging that someone impaired a relationship that we have with that person – the non-functionalist ingredient. When a friend tells a mean joke about me, I can judge that she thereby impaired our friendship. My response is only blame, according to this view, if it plays a specific social or mental role, namely, to revise our relationship according to my judgment. This can be done in various ways, such as resenting or coldly de-friending her – the functionalist ingredient. Miranda Fricker’s (2016) account of blame can be understood as another mixed theory. It says that blame necessarily involves a judgment that a person is at fault – the non-functionalist ingredient. In paradigmatic cases this judgment is communicated, which can be done in very different ways – the functionalist ingredient. To sum up, functionalist theories of blame answer the question of what blame is by referring to whatever plays certain mental or social roles. Non-functionalist theories say that specific attitudes constitute blame independent of their mental or social roles. And mixed theories say that to blame is to have a specific attitude together with whatever plays a certain social or mental role.

3  Theories of blame: what should we want from them? To find out which theory of blame to accept, we need to know how to evaluate them. We need to ask: what should we want from a theory of blame? Unfortunately, this question is rarely discussed. In what follows, I will present two mostly implicit motivations for developing theories of blame. One starting point for a theory of blame is the observation that blame is a common and potentially problematic phenomenon in our everyday lives: we blame each other all the time in very different ways even though most of us do not like being blamed and we see that blame often goes wrong. This is philosophically interesting. Thus, we can ask: what is the essence of 316

Blaming

these different things that we call blame? The goal is to provide a unifying explanation of our everyday thinking and talking about blame and of our blame responses. The best theory would identify the common core of all blame responses – be it the fulfillment of certain social or mental functions, a certain group of attitudes, or a mix of both. Thereby, it would explain why all the different responses are blame responses. Call an approach along these lines unifying. An alternative motivation for developing theories of blame is to solve problems that we are confronted with in moral or philosophical thinking. Consider, for example, debates about free will. Many authors characterize free will in terms of responsibility (for an overview see O’Connor & Franklin 2021): it is the control that is necessary for being responsible for something. Responsibility is often characterized in terms of praise- and blameworthiness: your being responsible for something is your being such that it is in principle appropriate to blame or praise you for that thing if it is bad or good. Thus, free will can be characterized in terms of the control that is necessary for blameworthiness. Therefore, one way to better understand what free will is, is to develop a theory of blame: once we know what blame is, we can better understand what free will is because free will is characterized in terms of being worthy of blame. Thus, philosophers who are interested in free will may be motivated to develop theories of blame. Below, I will present further questions that can motivate one to develop theories of blame. In general, I will call an approach of this kind problem-driven. These approaches are highly idealized. Actual theories of blame are often partly unifying and partly problem-driven, and often it is unclear to which of the two categories a specific theory belongs. But it is useful to work with the idealization because I will now show that a fully unifying approach is problematic.

4  Problems for unifying explanations Let us take a closer look at theories that aim at a unifying explanation of our blame practices. The standard but often implicit way to achieve this goal is to propose an inference to the best explanation. The approach is, roughly, this: we know from our everyday lives that our blame practices have certain features, and a philosophical theory of blame should explain these features. It should identify the common core of all blame responses that can explain the various features that our blame practices have. Here is a list of uncontroversial explananda that a theory of blame should explain (see, e.g., Smith 2013; Shoemaker 2013; Shoemaker & Vargas 2021): (1) blame can be private or public; (2) we can blame ourselves, people close to us, absent people, and even people we have never met, or who are long dead; (3) we can blame people without experiencing a feeling; (4) blame seems to be something people can be worthy of; (5) blaming people often has a sting in the sense that it is often bad for the target of blame. If we could show that one theory can explain these features and another cannot, then we have reason to prefer the first theory over the second. However, different accounts of blame can make sense of them in similarly good ways. Take for example Scanlon’s account (2008, chap. 4). It says that blaming people is, roughly, revising the relationship one has with them because one takes them to have impaired this relationship. Such a revision can happen in one’s mind by changing one’s attitudes towards the other person, but the response can also 317

Leonhard Menges

be expressed in open action by, for example, writing an email. One can change the relationship one has with oneself, by, for example, feeling guilty, but not every relationship revision needs to involve a feeling. On the assumption that we have a moral relation with every person, we can also revise it with long dead or distant people by, for example, not desiring that they had or have a happy life. Some people seem to be worthy of such a relationship revision, namely those who have impaired the relationship. As many people like having good relationships, it is stinging for them when their relationships are revised. There are questions about Scanlon’s account of blame (e.g., Chislenko 2019). But the general picture is well-suited to explain the features listed previously. Now consider the emotion account that says, roughly, that to blame is to have an emotion towards blame’s target. We can keep the emotion to ourselves or express it. We can be angry at the long dead or at our neighbor, and we are often angry at ourselves or feel guilty. Many people don’t like being the target of these emotions, which explains the stinging aspect of blame. Insofar emotions can be appropriate, it also seems as if some people are appropriate objects of angry blame, which is why it seems as if some people are blameworthy. A standard worry about the emotion account is that it cannot make sense of the fact that we sometimes blame people without experiencing a feeling (e.g., Sher 2006: 88). However, this rests on confusing emotions with feelings (see Menges 2017: 259). According to (almost) all theories of the emotions, having an emotion is closely connected with having feelings (see Scarantino  & de Sousa 2021). However, there is room to have an emotion without feeling something in a particular moment. For example, I  can be sad about my friend’s death ever since she died five years ago. This can be true even if I have not constantly felt sadness for five years. The same can be true for blame emotions. I can resent my neighbor ever since she put up an ugly fence five years ago even if I have not constantly experienced resentment. Again, there are questions about the emotion account. Generally, though, it can explain the core features identified previously, just like the relationship view. The same is true for most other accounts of blame. Thus, I believe that we cannot determine which account to accept by checking how they explain these features of our practice. Theories that aim at a unifying explanation may present additional features of our practice and demand that they should also be explained. Consider   (6) to blame is not simply to believe something (Sher 2006, chap. 1);   (7) we sometimes blame people for what they would do in certain circumstances even if, in fact, they are not in these circumstances and don’t do it (Shoemaker & Vargas 2021);   (8) we sometimes blame people even though we believe that they are not blameworthy (Pickard 2013);   (9) while it is mostly inappropriate for adults to blame children, it is often appropriate for children to blame other children (Reis-Dennis 2021); (10) blame is “the characteristic reaction of the morality system” (Williams 1985: 177); (11) there is non-moral blame such as epistemic or causal blame (Boult 2021; Chislenko 2021); (12) judging someone to be at fault for something and being against this something is a form of blaming (e.g., Arpaly & Schroeder 2014: 161); (13) if one has forgiven people then it is inappropriate to keep blaming them (e.g., Hughes & Warmke 2022, sec. 2.3). 318

Blaming

Some of these explananda come in tension: theories that explain one of them have problems explaining some of the others. For example, it is hard to see how one can make sense of, first, the idea that blame is the characteristic response of the morality system and, second, the idea that there is non-moral, epistemic or purely causal blame, such as blaming the brakes for the crash. An obvious reply is to distinguish between moral blame on the one hand and non-moral blame on the other and to say that one is only concerned with one of them. Note, however, that this reply already gives up on the goal to offer a unifying explanation of our blame practice. Our practice seems to involve moral and non-moral blame, which is why a unifying account of blame itself must make sense of both. Focusing on, for example, moral blame would only aim at an explanation of some kind of blame, which is not enough for a fully unifying explanation. Let me show in more detail how two explananda come in tension by focusing on (12) and (13). Fricker (2016: 170) considers a situation in which I think that it is the drivers’ fault that there is a strike and you think that it is the company’s fault. It seems correct to say that I blame the drivers and you blame the company. Every theory of blame, the idea goes, should explain this. Some theories have problems meeting this requirement. The emotion account, for example, says that my response is not blame unless I have an emotion. However, there are situations in which I find fault and do not have any emotion. Other theories can explain why my response is blame. For example, my finding fault signals my commitment to certain norms, expresses protest, or initiates a conversation with you or myself. Functionalist theories that say that every response that fulfills one of these functions is, therefore, a blame response, can explain why my response is a blame response. This seems to be an advantage of functionalist theories. Now, take explanandum (13). Imagine that, after an adequate apology, I have forgiven my sister for ruining my last birthday. However, I do not forgo blaming her. Intuitively, there would be something unfair, unfitting, incoherent, or odd about my blaming her for it. Every theory of blame should explain this, the idea goes. Some theories of blame, such as the emotion account, have no problem here. It says that to blame is to have an emotion, such as resentment. Obviously, there is something odd about forgiving my sister for ruining my birthday and, at the same time, resenting her for it. Thus, the emotion view can explain the oddity. Functionalist theories, by contrast, have problems here. They say that every response that plays a certain role is a blame response. However, there are some responses that play the relevant roles, and there is nothing inappropriate about forgiving my sister and, at the same time, responding in some of these ways. Imagine that I speak in a cool and quiet moment to close friends about how I was hurt by her conduct. This seems to be fine. There is nothing odd about forgiving S for having done x and, in a cool and quiet way, discussing S’s having done x. This signals my commitment to certain norms, is a form of cool protest, and invites my friends to have a conversation about the status of ruining peoples’ birthdays. Thus, the functionalist theories that explain how judging people to be at fault for something and how being against this something is a blame response thereby fail to make sense of the idea that forgiving people makes it odd to keep blaming them. Generally, there is reason to think that theories that can make sense of (12) have, therefore, problems making sense of (13) and vice versa. That is, accounts that explain why finding fault is a blame response have difficulties explaining why forgiving makes blame inappropriate. Reversely, accounts that explain why forgiving makes blame inappropriate have problems explaining how finding fault can be a blame response. 319

Leonhard Menges

It is unclear how to solve the tension. One attempt says that judging a person to be at fault and being against it – which fulfills certain functions – is not sufficient for blaming. Something must be added, such as an emotion. However, it is unclear how to justify this claim. For, in our everyday practice, it is completely fine to say of my friend, who does not have a blame emotion, that she blames the company for the strike. Alternatively, one could say that forgiving a person for doing something and blaming the person for this thing fit together. But again, it is hard to argue for this. Imagine that I tell my sister: “I still blame you for what you did on my last birthday.” It seems fully appropriate for her to reply: “I thought that you forgave me, I thought that we settled this.” Thus, there is a real tension between these two potential explananda. To sum up, first, accounts of blame that aim at developing a unifying explanation of our everyday practices typically adopt the method of proposing inferences to the best explanation. Second, there is an uncontroversial group of explananda and the standard accounts make sense of them. These explananda do not help settle which theory is best. Third, there is a group of more controversial potential explananda. They consist in features of our everyday practices that can be explained by some theories but not by others. Importantly, no theory seems to be able to explain all of them because some of them come in tension with each other.

5  The problem-driven approach How should we proceed? I  see three options. The first is to keep looking for a unifying account of all features of our blame practice. The second is to start a discussion about which aspects of our blame practices really are explananda and which aren’t. Thus, before developing our theories of blame, we should spend more energy making clear what the theory aims at explaining and, more importantly, why the theory should explain these and not some other features. The third option is to give up on the goal of developing a unifying explanation of our blame practices by identifying a common core. One attractive alternative (I’m not claiming that it is the only one) is to adopt the problem-driven approach. The idea is to start with an interesting social or philosophical question that has something to do with blame (of course, the problem should not be “what is the common core of all blame responses?”). Then, we can use theories of blame as tools to answer that question. Thus, we should want from theories of blame that they help us solve important social and philosophical problems. In what follows I will present two such problems. First, plausibly, blame often goes wrong, is destructive, painful, and an instrument of the powerful to keep others down. However, it also seems plausible that blame is sometimes important to make clear that we do not accept certain forms of conduct. Thus, we are confronted with a problem: what are the aspects of our blame practices that are worth preserving and what are the aspects we should try to get rid of? Let me start with the valuable aspects. Some argue that blame helps strengthen the norms and values that a well-functioning society needs (e.g., Wallace 1994: 69). Others propose that blame responses fit perfectly with the relational structure of moral norms (Darwall 2006, pt. 1; Wallace 2011) or that blame practices facilitate shared moral knowledge (Sliwa 2019). Others again suggest that blame practices help develop and sustain human capacities to respond to reasons (McGeer 2019). And some argue that it is of non-instrumental value when the guilty suffer appropriately in being (self-)blamed (e.g., Clarke 2016; McKenna 320

Blaming

2019). Some of these aspects seem worth preserving. A world that lacks these aspects of our blame practices is, therefore, in some respect, worse than our world. Now consider the problematic aspects of blame. Blame has a sting and is often harmful. Insofar harm is non-instrumentally bad, blame is often non-instrumentally bad. Moreover, blame and counter-blame can easily escalate and lead to destructive resistance rather than constructive conversation (e.g., Pereboom 2021, chap. 1). The anger involved in many blame responses can cloud our senses and make us ignore important features (e.g., Pettigrove 2012). Certain blame practices are problematic on a social level (for the following, see Zheng online first). Consider the idea that in the face of worldwide injustice, morality demands that we do what we can to make the world a just place. If we fail in some instance by not going to yet another demonstration or not giving yet another sum to charity, it appears inappropriate and de-motivating to blame us in standard ways. However, some form of criticism seems necessary to remind us of the ideal to make the world as just as possible. Thus, we need a better way to respond to some of our failings in an unjust world than standard blame responses. Recall that the question we are concerned with is “what are the aspects of our blame practices that are worth preserving and what are the aspects we should try to get rid of?” Theories of blame should help us answer it and they can do so without developing a unifying explanation of our blame practices. All we need is a plausible characterization – not a unifying explanation – of relevant parts of our blame practices. Such a characterization does not need to offer necessary conditions for blame, nor does it need to identify its core element. It should identify a group of responses that clearly are blame responses. Then, we can sensibly ask what is good and what is bad about them. For example, resenting others and feeling guilty are surely blame responses. Perhaps, there is blame without resentment or guilt, but it makes sense to ask what is good and what is bad about resentment- and guilt-blame. This is an interesting question. An answer to it can be the basis for proposing social reforms. Perhaps, we can re-shape our practices by preserving the good things about our current practices and getting rid of the bad ones. Perhaps, things are more difficult and the very features that make blame in some respect good make it in other respects bad (see McKenna 2019; Reis-Dennis 2019). The conclusion would then be that there is something tragic about our blame practices because we cannot have the good without getting the bad. More generally, theories of blame do not need to identify the core of blame for helping us discuss the question of what is good and what is bad about it. All we need are detailed characterizations of important kinds of blame. Second, take free will problems again. Intuitively, there seems to be something unjust about blaming people if what they did was determined by something beyond their control – this is what R. Jay Wallace calls “the seductive lure of incompatibilist pictures in ordinary moral consciousness” (Wallace 1994: 117). On the other hand, there seem to be blame responses that are clearly appropriate even if the targets of blame were fully determined. Obviously, we can appropriately blame the brakes for the crash in a deterministic universe and signaling commitment to the moral norm not to ruin people’s birthday in a cool and reasonable way seems unproblematic even if the world is deterministic. This confronts us with a problem: what, exactly, are the blame responses skeptics and incompatibilists should be concerned with, and what are the responses they do not need to worry about? We need an answer to this question to make sure that incompatibilists, compatibilists, skeptics, and so on, talk about the same thing when they ask if humans can be blameworthy (even if determinism is true). 321

Leonhard Menges

Some of those who doubt that humans are blameworthy in our world characterize the responses they are concerned with in terms of negative emotions or the (harmful) expressions of negative emotions (Pereboom 2014: 128, 2021, chap. 2; Rosen 2015). Others account for blame in terms of punishing conduct (G. Strawson 1994; Caruso & Morris 2017). A third view argues that the relevant kind of blame should be understood in terms of claim forfeiture (Menges 2021). The idea is that the relevant kind of blame involves some response that would normally infringe the right of its targets and the belief of the blamers that the targets have forfeited this right just because of what they did and how they were when they did it. Skeptics can argue that, because of determinism or luck, humans never forfeit their rights just because of what they do and how they are. Therefore, skeptics conclude, the relevant kind of blame always infringes the rights of their targets. Once it is clear what part of our blame practices skeptics are concerned with, they should admit that other parts of our blame practices are unproblematic (see Pereboom 2021). Similarly, those who try to defend the idea that humans in our world are sometimes blameworthy for what they do should make clear what they take blame to be. Then, it will probably turn out that some defenders of blameworthiness have one thing in mind when they defend human blameworthiness and some skeptics have something else in mind when they argue against human blameworthiness. Generally, it is reasonable to discuss what kind of blame should be at issue in debates about free will without first developing a unifying explanation of our blame practices. What we need is a characterization of certain aspects of our blame practices. Then we can ask if these aspects are undermined by skeptical or incompatibilist arguments. There are other problems that can motivate theories of blame that I cannot discuss in detail here. Think of the debate about the responsibility gap in machine ethics that is based on the question of who is blameworthy when an “intelligent” machine, such as a self-driving car, causes serious harm without any human fault (see, e.g., Matthias 2004; Nyholm 2018). With this question in mind, we could look at different kinds of blame and ask: is this kind of blame appropriately directed at the consumer, the designer, or the machine itself? Depending on how one construes the relevant kind of blame, the answers will probably vary. For example, it seems appropriate to judge that a machine caused something bad and to be against the machine’s causing it. It also seems appropriate to initiate a conversation about the machine’s harmful conduct with others. According to some theories, we would, thereby, blame the machine, and this seems to be perfectly fine. But it seems less obvious that we can appropriately revise our moral relationship with the machine. Thus, machines may be the appropriate target of blame in some respects, but not in others. More generally, a theory of blame may help clarify or even bridge the responsibility gap. Similarly, we can start with the question of when people lose the standing to blame others. A  hotly debated question is if and why having committed a certain wrongdoing without having adequately apologized for it undermines the standing to blame others for a wrongdoing of the same kind (e.g., Fritz & Miller 2018; Todd 2019; see King’s chap. 26 in this volume). Again, it is important to make clear what kind of blame is at issue, for it seems plausible that having done the same wrong undermines the standing to some forms of blame but not to others. We do not need a unifying account of the blame practices. Rather, we need accounts of different kinds of blame and to discuss whether some of these kinds of blame can be fully appropriate coming from someone who performed the same wrongdoing in the past and has not adequately apologized for it. 322

Blaming

6 Conclusion What should we expect from a theory of blame? One answer is that we should want a unifying explanation of our blame practices. I have suggested that current accounts fail on this because, first, it is unclear what explananda the theory should focus on and, second, because some potential explananda are in tension with each other. I have then suggested that this would not undermine the importance of developing philosophical accounts of blame. The key idea is to start with more specific problems such as: “What is good and what is bad about our blame practices?” “What is the kind of blame that should be at issue in debates about incompatibilism or skepticism about free will and responsibility?” “What kind of blame should be at issue in debates about the responsibility gap?” “What is the kind of blame that is undermined when we have performed the same wrongdoing we blame others for?” These and other questions can be fruitfully discussed without a unifying explanation of our blame practice.1

Note 1 Thanks to Hannah Altehenger, Gunnar Björnsson, Sebastian Köhler, Michael Schefczyk, Jan Willem Wieland, and the participants of the online workshop for this handbook on Responsibility Practices in October 2022 for helpful comments and suggestions. Special thanks to Max Kiener for very helpful comments and questions. Work on this chapter was supported by the Austrian Science Fund (P 34851-G) and is part of the project “The Sense of Responsibility Worthy Worrying About”.

Further reading The best overviews on blame are the papers by Coates, Justin D.,  & Tognazzini, Neal. (2018). “Blame,” in E. N. Zalta (ed.), The Stanford Encyclopedia of Philosophy, Fall. Metaphysics Research Lab, Stanford University. https://plato.stanford.edu/archives/fall2018/entries/blame/ and Coates, Justin D., & Tognazzini, Neal. (2012). “The Nature and Ethics of Blame,” Philosophy Compass, 7, 197–207). The book they edited (2013). Blame: Its Nature and Norms. New York: Oxford University Press still shapes the discussion. For a representative functionalist account of blame, see Shoemaker, D., & Vargas, M. (2021). “Moral Torch Fishing: A Signaling Theory of Blame,” Noûs, 55, 581–602. For powerful defense of a non-functionalist account see Portmore, D. W. (2022). “A Comprehensive Account of Blame: Self-Blame, Non-Moral Blame, and Blame for the Non-Voluntary,” in A. B. Carlsson (ed.), Self-Blame and Moral Responsibility (pp. 48–76). Cambridge: Cambridge University Press. For a classic mixed theory see Scanlon, T. (2008). Moral Dimensions: Permissibility, Meaning, Blame. Cambridge, MA: Harvard University Press, ch. 4.

References Arpaly, N., & Schroeder, T. (2014). In Praise of Desire. New York: Oxford University Press. Boult, C. (2021). “Epistemic Blame,” Philosophy Compass, 16(8). https://doi.org/10.1111/ phc3.12762. Caruso, G. D., & Morris, S. G. (2017). “Compatibilism and Retributivist Desert Moral Responsibility: On What Is of Central Philosophical and Practical Importance,” Erkenntnis, 82(4), 837–855. https://doi.org/10.1007/s10670-016-9846-2. Chislenko, E. (2019). “Scanlon’s Theories of Blame,” The Journal of Value Inquiry, July. https://doi. org/10.1007/s10790-019-09703-7. ———. (2021). “Causal Blame,” American Philosophical Quarterly, 58(4), 347–358. https://doi. org/10.2307/48619319. Clarke, R. (2016). “Moral Responsibility, Guilt, and Retributivism,” The Journal of Ethics, 20(1–3), 121–137. https://doi.org/10.1007/s10892-016-9228-7.

323

Leonhard Menges Darwall, S. (2006). The Second-Person Standpoint: Morality, Respect, and Accountability. Cambridge, MA: Harvard University Press. Fricker, M. (2016). “What’s the Point of Blame? A  Paradigm Based Explanation,” Noûs, 50(1), 165–183. Fritz, K. G., & Miller, D. (2018). “Hypocrisy and the Standing to Blame,” Pacific Philosophical Quarterly, 99(1), 118–139. https://doi.org/10.1111/papq.12104. King, M. (2024). “The standing to Blame,” in M. Kiener (ed.), Routledge Handbook of Responsibility (chap. 26). New York: Routledge. Hughes, P. M., & Warmke, B. (2022). “Forgiveness,” in E. N. Zalta (ed.), The Stanford Encyclopedia of Philosophy, Spring. Metaphysics Research Lab, Stanford University. https://plato.stanford.edu/ archives/spr2022/entries/forgiveness/. Mason, E. (2019). Ways to Be Blameworthy: Rightness, Wrongness, and Responsibility. New York: Oxford University Press. Matthias, A. (2004). “The Responsibility Gap: Ascribing Responsibility for the Actions of Learning Automata,” Ethics and Information Technology, 6(3), 175–183. https://doi.org/10.1007/ s10676-004-3422-1. McGeer, V. (2019). “Scaffolding Agency: A Proleptic Account of the Reactive Attitudes,” European Journal of Philosophy, 27(2), 301–323. https://doi.org/10.1111/ejop.12408. McKenna, M. (2013). “Directed Blame and Conversation,” in D. Justin Coates & Neal A. Tognazzini (ed.), Blame: Its Nature and Norms (pp. 119–140). New York: Oxford University Press. ———. (2019). “Basically Deserved Blame and Its Value,” Journal of Ethics and Social Philosophy, 15(3). https://doi.org/10.26556/jesp.v15i3.547. Menges, L. (2017). “The Emotion Account of Blame,” Philosophical Studies, 174(1), 257–273. ———. (2021). “The Kind of Blame Skeptics Should Be Skeptical About,” Canadian Journal of Philosophy, 51(6), 401–415. https://doi.org/10.1017/can.2021.38. Nyholm, S. (2018). “The Ethics of Crashes with Self-Driving Cars: A Roadmap, II,” Philosophy Compass, 13(7). https://doi.org/10.1111/phc3.12506. O’Connor, T., & Franklin, C. (2021). “Free Will,” in E. N. Zalta (ed.), The Stanford Encyclopedia of Philosophy, Spring. Metaphysics Research Lab, Stanford University. https://plato.stanford.edu/ archives/spr2021/entries/freewill/. Pereboom, D. (2014). Free Will, Agency, and Meaning in Life. New York: Oxford University Press. ———. (2021). Wrongdoing and the Moral Emotions. New York: Oxford University Press. http:// dx.doi.org/10.1093/oso/9780192846006.001.0001. Pettigrove, G. (2012). “Meekness and ‘Moral’ Anger,” Ethics, 122(2), 341–370. Pickard, H. (2013). “Irrational Blame,” Analysis, 73(4), 613–626. Portmore, D. W. (2022). “A Comprehensive Account of Blame: Self-Blame, Non-Moral Blame, and Blame for the Non-Voluntary,” in A. B. Carlsson (ed.), Self-Blame and Moral Responsibility (pp. 48–76). Cambridge: Cambridge University Press. Reis-Dennis, S. (2019). “Anger: Scary Good,” Australasian Journal of Philosophy, 97(3), 451–464. https://doi.org/10.1080/00048402.2018.1520268. ———. (2021). “Rank Offence: The Ecological Theory of Resentment,” Mind, 130(520), 1233–1251. https://doi.org/10.1093/mind/fzab006. Rosen, G. (2015). “The Alethic Conception of Moral Responsibility,” in R. Clarke, M. McKenna, & A. Smith (ed.), The Nature of Moral Responsibility: New Essays (pp. 65–88). New York: Oxford University Press. Scanlon, T. M. (2008). Moral Dimensions: Permissibility, Meaning, Blame. Cambridge, MA: Harvard University Press. ———. (2013). “Interpreting Blame,” in D. J. Coates & N. A. Tognazzini (ed.), Blame: Its Nature and Norms (pp. 84–99). New York: Oxford University Press. Scarantino, A., & de Sousa, R. (2021). “Emotion,” in E-N. Zalta (ed.), The Stanford Encyclopedia of Philosophy, Summer. Metaphysics Research Lab, Stanford University. https://plato.stanford.edu/ archives/sum2021/entries/emotion/. Sher, G. (2006). In Praise of Blame. New York: Oxford University Press. Shoemaker, D. (2013). “Blame and Punishment,” in D. J. Coates & N. A. Tognazzini (ed.), Blame: Its Nature and Norms (pp. 100–118). New York: Oxford University Press.

324

Blaming Shoemaker, D., & Vargas, M. (2021). “Moral Torch Fishing: A Signaling Theory of Blame,” Noûs, 55(3), 581–602. https://doi.org/10.1111/nous.12316. Sliwa, P. (2019). “Reverse-Engineering Blame,” Philosophical Perspectives, 33(1), 200–219. https:// doi.org/10.1111/phpe.12131. Smith, A. M. (2013). “Moral Blame and Moral Protest,” in D. J. Coates & N. A. Tognazzini (ed.), Blame: Its Nature and Norms (pp. 27–48). New York: Oxford University Press. Strawson, G. (1994). “The Impossibility of Moral Responsibility,” Philosophical Studies, 75(1–2), 5–24. Strawson, P. F. (1962). “Freedom and Resentment,” in G. Watson (ed.), Free Will (pp. 72–93). New York: Oxford University Press, 2003. Todd, P. (2019). “A Unified Account of the Moral Standing to Blame,” Noûs, 53, 347–374. https:// doi.org/10.1111/nous.12215. Tognazzini, N., & Coates, D. J. (2018). “Blame,” in E. N. Zalta (ed.), The Stanford Encyclopedia of Philosophy, Fall. Metaphysics Research Lab, Stanford University. https://plato.stanford.edu/ archives/fall2018/entries/blame/. Wallace, R. J. (1994). Responsibility and the Moral Sentiments. Cambridge, MA: Harvard University Press. ———. (2011). “Dispassionate Opprobrium: On Blame and the Reactive Sentiments,” in R. J. Wallace, R. Kumar, & S. Freeman (ed.), Reasons and Recognition: Essays on the Philosophy of T. M. Scanlon (pp. 348–372). New York: Oxford University Press. Wang, S. T. (2021). “The Communication Argument and the Pluralist Challenge,” Canadian Journal of Philosophy, 51(5), 384–399. https://doi.org/10.1017/can.2021.30. Williams, B. (1985). Ethics and the Limits of Philosophy. Cambridge, MA: Harvard University Press. Wolf, Susan. (2011). “Blame, Italian Style,” in R. J. Wallace, R. Kumar, & S. Freeman (ed.), Reasons and Recognition: Essays on the Philosophy of T. M. Scanlon (pp. 332–347). New York: Oxford University Press. Zheng, R. (online first). “Moral Criticism and Structural Injustice,” Mind, no. fzaa098. https://doi. org/10.1093/mind/fzaa098.

325

25 COMMUNICATING PRAISE Daniel Telech

Introduction Praising is a communicative phenomenon. Or, at least, there is a kind of praise that is both integral to our responsibility practices and, in some important sense, communicative in nature. The claim that praise (or even one kind of praise) is communicative, however, may seem mysterious, or just flatly mistaken. For, in the case of blameworthy action, the wrongdoer can plausibly be understood to be called to account for his wrongdoing in being blamed. That is, the claim that blame is communicative – or that it involves “moral address” – finds intuitive articulation in the idea that to blame another is to ‘hold them accountable’ (or, to ‘hold them answerable’), often understood in terms of the blamer’s demanding that the wrongdoer properly acknowledge and ‘account for’ (or, ‘answer for’) their wrongdoing, by expressing guilt or remorse, apologizing, offering redress, etc. (Watson 1987; Darwall 2006; McKenna 2012; Shoemaker 2015; Helm 2017). The accountabilitytaking responses sought by blame are in this way remedial responses, presupposing a previously disregarded weighty normative consideration, one that perhaps persists in modified form after the wrongdoing (Gardner 2007: 33; Nelkin 2015: 363). But if blame’s communicativeness is intelligible as a way of calling the wrongdoer to account for their wrongdoing, a similar explanation will be lacking for praise.1 For, there is nothing for which the praiseworthy agent owes the praiser an account – nothing to account, or answer, for. After all, the ‘rightdoer’ already properly regarded the relevant moral reasons in having acted praiseworthily; indeed, it is in virtue of their having properly regarded (and acted upon) the relevant reasons that they are praiseworthy.2 And yet, a number of theorists endorse the idea that praise is communicative in the sense of being addressed to and seeking a response from the rightdoer, i.e., the praiseworthy agent. Sometimes praise is construed as communicative in a far weaker sense than this response- or uptake-seeking sense of communicativeness. Consider the claim that “[p]raise communicates information about others’ evaluations – of specific performances, qualities of the self, or the entire person. It can be an important source of information about the self” (Crocker 2021: xvii), or that “praise conveys the message that one has the ability to succeed” (Delin & Baumeister 1994: 225; cf. Holroyd 2007: 268). Here, praise DOI: 10.4324/9781003282242-36 326

Communicating praise

communicates in the sense of conveying or implying certain information, without being essentially addressed to someone from whom a response is sought. Without denying that praise conveys information, e.g., that the praiser positively evaluates something ascribable to the agency of the praisee, I  reserve talk of praise’s communicativeness to the uptakeseeking sense of communication. The aim of this chapter is to place on firmer ground the claim that our responsibility practices feature communicative praise, i.e., praise that is communicative in the uptakeseeking sense. I do so primarily by identifying and responding to three challenges to the idea of uptake-seeking praise: the descriptive adequacy objection; the normative objection; the redundancy objection. But first, in order to sufficiently distinguish between praise and nearby phenomena, like flattery and (mere) judgments of praiseworthiness, I  provide an outline of praise.

1  An opinionated outline of praise The sense of “praise” of interest to me is that associated with ‘giving credit’ to another for something, e.g., an action, and as such, of ascribing it to their agency.3 More substantively, I take it that to praise someone for some action is, at a minimum, to positively evaluate and take non-instrumental satisfaction in their performance of that action – where non-instrumental satisfaction consists in the subjective satisfaction of a non-instrumental desire4  – and, on that basis, to (be motivated to) communicate one’s attitude to the target.5 We need not worry ourselves here in spelling out the precise content of praise’s positive evaluation; at a minimum, it represents the action as a) particularly good so as to merit praise and b) ascribable to the agency of the praise, e.g., an expression of the agent’s good quality of will that exceeds what others’ can demand of her. That the praisee is represented as meriting a positive response on the basis of some agential contribution distinguishes praise from its curry-favoring cousin, flattery.6 It is insufficient for praise that the agent merely positively evaluates an agent’s action in the previously characterized way; a judgment (or belief) of praiseworthiness might do this. But, intuitively, one might judge that an agent acted praiseworthily (e.g., that they manifested a laudable degree of good quality of will) without praising them in the relevant sense. Perhaps the praiseworthy agent is a distant historical figure or one’s enemy, in which case one’s taking the agent to be praiseworthy might lack moral psychological import for one, and so, fail to translate into praise. Or imagine that the judgment of praiseworthiness is made by Satan or the amoralist, who, in thinking that some agent acted praiseworthily, is left indifferent or has the accompanying thought, ‘what a sucker!’7 To give credit, as one does in the sense of praise of interest to me, in addition to positively evaluating the agent for their action, one must take non-instrumental satisfaction in their performance of that action (and be motivated to communicate this evaluation and satisfaction, in a sense to be further specified shortly). The satisfaction taken in the positively evaluated action must be of a non-instrumental kind. For, if the ground of one’s satisfaction in another’s performance of some generous deed, say, were solely that one placed a bet forecasting this kind of action, one’s satisfaction (even if accompanied by the right kind of evaluation of the action) will not be of the right sort to constitute praise. Praise involves non-instrumental satisfaction in the positively evaluated action. Some evidence for thinking that we’re on the right track comes from reflection on the fact that expressions of admiration (and similar attitudes) are naturally understood as 327

Daniel Telech

expressions of praise, at least when agent-directed.8 When I  admire, say, my peer’s (A’s) going out of her way to make a visitor feel welcome, I am, inter alia, positively evaluating and taking non-instrumental satisfaction in A’s (kind, fair-minded, considerate, supererogatory, generous) action. Consider what else I am doing in expressing my admiration to A, in, e.g., saying ‘it was very considerate of you to host a dinner to welcome B to the department’. I am framing A’s praiseworthy action in a particular way salient to me, i.e., as considerate. I am, further, inviting A to respond to my way of admiring A’s action. As my admiration involves crediting A for her treatment of B, in inviting A to respond to my admiration of her, I am inviting A to accept credit from me. (Perhaps A will judge me poorly positioned to evaluate her action, in which case she might reject my praise without denying that she acted praiseworthily, that she is the fitting target of praise.) Of course, B himself might respond to A’s praiseworthy action, e.g., in expressing gratitude to A for her action, perhaps saying ‘thank you for your kindness and friendliness in welcoming me to the department’. Gratitude too, at least gratitude directed to agents for their benevolent actions, involves positively evaluating some action in which one takes non-instrumental satisfaction. Here too, perhaps more vividly, B’s grateful response frames A’s action in terms salient to B and invites A to take credit for the action so framed. While it might be possible to praise an agent in the above sense without therein feeling a positive emotion toward that agent for that action, I take it to be no accident that, following Strawson (1962), interpersonal emotions (or “reactive attitudes”) like admiration and gratitude are often identified as vehicles of praise. While gratitude involves positively evaluating and taking satisfaction in another’s acting well toward oneself (or toward another with whom one identifies, e.g., one’s child), admiration involves positively evaluating and taking satisfaction in another’s acting well toward another (with whom one does not identify). Without treating responses like admiration and gratitude as necessary for praise, I proceed under the assumption that praise is paradigmatically instantiated in attitudes like these. These are attitudes in which one values an agent’s action non-instrumentally. That is, assuming that thus valuing is a way of taking non-instrumental satisfaction in an agent’s positively evaluated action, which, what’s more, seeks uptake from its target, we can see why admiration and gratitude emerge as paradigmatic vehicles of praise in the communicative sense. The Strawsonian approach coheres with my treating praise as a communicative phenomenon, for, following Gary Watson (1996), a large cluster of Strawsonians understand the reactive attitudes as communicative phenomena, or forms of ‘moral address’.9 In saying that praise invites a response, I am claiming that praise is communicative in more than the information-implying sense. After all, invitations – in addition to informing their recipients of various matters (e.g., that some event will (or may) take place at some time) – seek a response of a particular sort. They call for uptake. What kind of communicative response makes an invitation successful? It is acceptance of the invitation. The uptake sought by praise, accordingly, is acceptance of the invitation issued by praise. Consider my admiration and B’s gratitude of A  for A’s action. Despite their differences, both aim for A to value her own action, as framed by the praiser, with the praiser. That is, they are invitations to co-value the praiseworthy agent’s action. As the valuing of A’s action that is involved in my admiration and B’s gratitude is valuing that construes B’s action as worthy of, or meriting, a positive response, the valuing sought in turn will be valuing wherein one accepts credit from the praiser. If discursive expression of admiration and gratitude is the paradigmatic way in which we recognize and value others’ praiseworthy actions, it is 328

Communicating praise

natural that it is via discursive expression of positive self-reactive attitudes, like self-approbation and pride, that praisees paradigmatically recognize and value their own actions in response to the praise of others.10 On the invitational view, then, praise issues an invitation to accept credit in discursively co-valuing the praiseworthy agent’s action in terms salient to the praiser.

The objection from descriptive adequacy One might wonder whether the invitational view of praise captures what we’re doing when engaging in praise. That it is descriptively inadequate is not an automatic disqualifier for a view of praise, for one could be engaged in a revisionary, prescriptive, project, one that aims to characterize how we should praise (where this prescription gets its normative force from, e.g., considerations of justice, fairness, or utility). But while the invitational view of praise might offer local prescriptions about how to praise in order to secure praise’s soughtafter response, it is meant to capture what we are doing, at least in paradigmatic cases, in praising others. That is, continuous with the Strawsonian methodology of aiming to elucidate the nature of moral responsibility by attending to the subtleties of our responsibility practices, as we find them, the invitational view (at least that under consideration here) purports to be descriptively adequate. To see why one might doubt the descriptive adequacy of the invitational view, consider the oddness of the following exchange: • D: ‘It was really considerate of you to help F with that problem of hers.’ • E: ‘I recognize your valuing of my action, and I value my action (with you) in response’. Not only do rightdoers characteristically not respond to praise as E does, rightdoers seem to regularly deflect or reject praise, by saying things like ‘it was nothing’, ‘don’t mention it’, etc. These are intuitively responses of the opposite sort than those predicted by the invitational view. They do not look like self-valuing responses. In reply, norms of modesty presumably temper the way in which praise is received and, by extension, the ways in which pride felt in response to praise is expressed (Leech 1983).11 Given the association of certain facets of pride with vices like arrogance (Tracy & Robins 2007, 2014; Tracy et al. 2009), it should not be surprising that we have internalized strategies to express pride in ways that do not signal conceit. In this context, responses like ‘don’t mention it’ or ‘it was nothing’ are conventional ways of discursively giving praise uptake. Elinor Mason makes this kind of point in outlining a view on which praise’s acceptance similarly generates a kind of joint valuing: Imagine that my friend decides to donate 10 per cent of her income to charity. . . . When I praise her, I am entering into a communication: I am telling her that I recognize the moral value of her act, and the effort she has put into it. I want her to accept my praise. Verbally she may brush it off, but that is compatible with her accepting it. We tend to say things like, ‘oh, you know, it’s not much’. But that can be a way of accepting praise. In accepting my praise, my friend does her part of the praise conversation, accepts my assessment, quietly takes pleasure in it, and we jointly affirm our shared values. (Mason 2019: 108, italics added) 329

Daniel Telech

There is a big difference between the seemingly tepid response, ‘it was nothing’, and one that conventionally rejects praise, e.g., ‘I did not do that’ or ‘I regret doing that; it conflicts with my values.’ Furthermore, on the assumption that the rightdoer responded, ex ante, to the value of the intended action for which they are praiseworthy – without which they would not be praiseworthy  – it is difficult to see how, ex post, the rightdoer would not be disposed to value what they have done. What’s more, in being praised, the rightdoer is receiving feedback that confirms their earlier assessment of the value of their action. While it is surely possible for the rightdoer to have valued their action ex ante but not ex post – perhaps they underwent a shift in values in the interim – normally (or at least, in the absence of misjudgment or other error), their positive attitudes toward the action as to be done will translate into positive attitudes toward the action as done. Further, insofar as the valuable action is viewed as done by them – i.e., as one’s own doing – it will presumably be valued as such. So, while norms of modesty may mediate our responses to praise such that we downplay our praiseworthy actions, downplaying one’s praiseworthy action is compatible with valuing it in the mode of pride. Additionally, although the pride in question is a species of agential-pride – pride that targets an expression of one’s agency – the pride sought by praise is not simply pride in ‘having done the right thing’, but pride in, as it were, having righted another. That is, it is patient-focused; it is a way of valuing one’s own action as significant for another (Telech 2021: 164). After all, assuming the praiseworthy agent is morally praiseworthy (for some other-regarding action), she will not simply have done the right thing, but have done the right thing from other-regarding motives, i.e., in a patient-focused way. It is thus no surprise that we often respond to praise in ways that make reference to our original reasons for action, e.g., ‘I’m am happy to have been able to help’.

The normative objection Next, one might object to the uptake-seeking conception of communicative praise on the following grounds: it seems unjust, unfair, or some such, that, having acted praiseworthily, the rightdoer should now owe a response to the praiser upon being praised. On what basis can the praiser legitimately expect the kind of uptake that, according to the invitational view, praise calls for? While blamers (perhaps especially, victims of wrongdoing) are intuitively licensed to put normative pressure on wrongdoers in blaming them, whence comes the praiser’s entitlement to the rightdoer’s response? Surely, the praiseworthy agent does not owe the praiser a response, much less the kind of reparative response associated with the entitlement to seek uptake in blame. In reply, it should be granted that praise does not communicate a normative expectation of (or demand for) uptake. One might worry, however, that once it is denied that praise comes with an expectation of response – or that there is normative pressure on the praisee to respond in the previously characterized way – the idea that praise is communicative in more than the information-implying sense is rendered opaque. We can put the normative objection in terms of the following dilemma: either a) praise communicates that the praisee should (or is expected to, or owes it to the praiser that they) give praise uptake or b) it does not. If a), communicative praise is inappropriate, for the praiseworthy agent does not owe the praiser an account of their rightdoing (and to treat them as if they do is to treat them unjustly). If b), communicative praise is normatively impotent; it is not clear how praise

330

Communicating praise

is uptake-seeking rather than merely information-providing. Robert Wallace (2022: 468) presses a version of the second horn of this dilemma: The recipient of blame should feel (or know that they should feel) guilt; if otherwise, they must correct the blamer’s assessment of what happened in an accounting. The recipient of gratitude need not accept it as praise, or justify their actions, or offer an excuse. They can even downplay what happened. It is hard to see how these responses indicate that a person is being held to account as a feature of our moral practices, even if these responses indicate something communicative and interpersonal has happened, namely, an internal sense of uptake. How is this sense of recognition as internal uptake anything more than recognizing in yourself, merely attributing to yourself, a beneficent will? The normative objection can be defused. First, it is true that the invitational view understands praise to put normative pressure on the praisee. So, the invitational view denies b), specifically the implication that praise is normatively impotent. Nevertheless, praise does not communicate an expectation of uptake. The praisee is under no requirement to give praise uptake, as the blamee arguably is with respect to blame. That is, the invitational view also denies a). To see how both a) and b) can be rejected, let us turn to the distinctive features of invitation. Valid invitations provide their addressees with reason to accept the invitation because invited. That is, like demands and requests, invitations are directives. The directive here is discretionary in that the reason provided, like that of request and unlike that of command, is non-obligatory, i.e., not a conclusive reason. Invitations seek acceptance, rather than compliance. Whether to do as invited is up to the invitee’s discretion. And yet, invitations put normative pressure on the addressee to accept, for this is what invitations seek. As Martin (2021: 75) notes, invitations ‘carry with them a certain legitimate pressure to accept, where the invitee needs a good reason to refuse (beyond say “I don’t feel like it”)’. In this respect, invitations contrast with offers. Both seek some reply, but speech act theorists understand offers as neutral between acceptance and rejection, while “invitations prefer acceptances and disprefer rejections” (Walker 2013: 456). Evidence for this is found in the normative remainder generated by declining a valid invitation, namely the appropriateness of an expression of regret, e.g., ‘I’m sorry I  can’t make it’.12 Thus, even if we deny that praise possesses the normative force of demand – imperatival force – it may nonetheless be, as the invitational view maintains, that praise puts normative pressure on the praisee to give praise uptake.

The objection from redundancy Finally, one might worry that any view on which the praisee is called to recognise the value of their action must be redundant. For, the praiseworthy agent typically already knows they acted praiseworthily (assuming she acted from recognition of the relevant moral reasons), and so, will have apprehended the value of their action. After all, it’s in light of this value that the praiseworthy agent acted in the first place. So, the idea that praise seeks uptake in the praisee’s seeing herself as having done something valuable in pride is implausible; praise that sought this would be redundant.

331

Daniel Telech

Cheshire Calhoun (2021: 24–25) raises a version of this objection: Why does sending the gratitude message matter? Is it really to get people to selfapprovingly see themselves as the grateful see them? Those who do us favors, live up to normative expectations when most others wouldn’t, contribute to collective projects, engage in heroic rescues, and so on do so intentionally for just the right reasons – or at least, if they deserve to be appreciated or thanked, they do. Any message whose gist is “You did a good thing” simply affirms what the appreciative or grateful person must assume the target already knows. Thus, the target already has grounds for feeling self-approbation (along with noticing her good qualities, interpreting her actions in the best possible light, patting herself on the back, and doing something nice for herself). Thus, sending appreciation and gratitude messages seems pointless. The redundancy objection is useful in illustrating the shortcomings of the information-conveying view of praise’s communicativeness. If the information conveyed by praised is information about the act’s rightness (/right-making features), on the assumption that an agent’s being praiseworthy for acting rightly presupposes her being aware of and guided by the act’s rightness (/right-making features) (D’Arcy 1963: 160; Arpaly 2003: 79, 84; Zimmerman 1988: 50; Haji 1998: 175), praise will indeed by redundant. As Calhoun (2021: 25, n.28) elaborates, “the appreciative or grateful person must assume that the target intended to do something of positive moral import because of its positive moral import in order to be justified in thinking that appreciation or gratitude was the appropriate response.” The information ‘conveyed’ ex post will be information the praisee already possessed – indeed, was guided by – ex ante. And what’s more, the praiser will normally know this, assuming their praise is justified. In reply, first, praise does not simply tell the praisee that what they did was praiseworthy (/had positive moral import). It does this in terms salient to the praiser. Praiseworthy actions are replete with value that can be described in myriad ways reflective of the praiser’s particular concerns, their relative weightings of the normative considerations, their history, etc. One and the same action may, for example, be generous, compassionate, courageous, and so on. Different praisers might fittingly praise the action in distinct but non-competing ways, one for its generosity, one for its compassionateness, and so on. Even if the praiseworthy agent must have been guided by the act’s rightness, she need not have conceived of her action under the description salient to third parties, much less by the particular third party doing the praising. Secondly, and more importantly, according to the invitational view of praise, the uptake sought by praise is not for the praisee simply to recognize the value of their action. Its point is not exhausted by its content (including its content as informed by the aspects of the action salient to the praiser). This content comes in cordial form – that of invitation – as praise is itself an expression of good will (Wallace 2022: 476). This is reflected in the fact that other things being equal, we appreciate, feel gratified by, etc., others’ praise. Without collapsing the distinction between praise and reward (where the latter but not the former includes an intention to benefit), praise is reward-like in the following sense: it tends to promote the interests of, or benefit, its recipient. This should not be surprising given that praise is a way of valuing some agent for her praiseworthy action. In addition to verbally expressing gratitude and admiration  – which can be meaningful to the praisee over and

332

Communicating praise

above the meaning of its content – we ‘show’ our thanks and admiration in various ways meant to express the value, to us, of the praiseworthy agent’s action. The invitational view of praise captures the intuition that praise paradigmatically benefits the praisee. For, invitations represent it to be in the invitee’s interest to accept the invitation. This is one important dimension along which requests are distinguished from invitations. As Drew and Couper-Kuhlen (2014: 6) write, “[w]hen A requests B to do something, A is the beneficiary; but when A invites B to (do) something, then B becomes at least a co-beneficiary.” In the case of praise, the addressee’s being a co-beneficiary of the praising interaction will be a matter of her accepting praise in co-valuing her action in terms salient to, and presented by, the praiser. Thus, even when the praiseworthy agent knows that they are praiseworthy and also that the would-be praiser judges them praiseworthy in some particular way, communicative praise has a point. In praising another, we are inviting the praiseworthy agent to accept credit in valuing their praiseworthy action as valued by us. If accepted, the praiser and the praiseworthy agent will co-value the latter’s action in a manner hosted by (because initiated by and framed in terms salient to) the praiser. While the praiseworthy agent will have valued the action ex ante, she could not have jointly valued the significance of that action for the praiser. Thus, praise that seeks uptake in the praiseworthy agent’s valuing of her action need not be redundant.

Conclusion This chapter presented and defended a view of praise as a communicative phenomenon, distinguishing communicativeness in the (mere) information-conveying sense and communicativeness in the uptake-seeking sense. On the view outlined, praise is communicative in that it invites the praiseworthy agent to accept praise by co-valuing their action in the evaluative terms supplied by the praiser. Praise is invitational in that it i) provides the praisee with a (discretionary) directive reason to accept praise; ii) seeks to host the addressee in co-valuing her action; and iii) characteristically benefits the praisee. Thus, although it is not a response that calls the praisee ‘to account’ in a remedial sense, this is no challenge to the proposal that praise is a communicative phenomenon.13

Notes 1 For elaboration, see King (2014); Macnamara (2013a, 2013b); Eshleman (2014); Telech (2020). 2 What’s more, even if it is possible to be praiseworthy for some unwitting rightdoing (such that the praiseworthy agent is not aware she did the right thing, under the relevant description), praise of the unwitting rightdoer would not intuitively amount to a form of holding to account. 3 I speak of praise as targeting (agents for their) actions for simplicity’s sake. Other candidates for proper objects of praise include virtues of character, and non-volitional attitudes like desires, beliefs, and emotions. One’s response to the question, ‘what are the proper objects of praise?’ will be determined by one’s account of praiseworthiness; a traditional assumption is that moral responsibility generally (and by extension, praiseworthiness) presupposes voluntary control, thus ruling out responsibility for the non-voluntary. But see Adams (1985), Scanlon (1998), and Smith (2005, 2008). 4 To desire X non-instrumentally is to desire X at least in part for its own sake. On non-instrumental (or, ‘intrinsic’) desires, see Arpaly and Schroeder (2014). 5 An account of praise’s communicative aim is well served by focusing on praise in its paradigmatic form, i.e., praise that is in fact communicated (cf. Fricker 2016; McKenna 2012). While

333

Daniel Telech the motivation to address the praisee might fail to find communicative expression (much less, its sought-after uptake), I limit my scope to communicated praise. 6 On flattery, see Eylon and Heyd (2008). 7 See Arpaly and Schroeder (2014: 160–161) on this point. 8 For skepticism about there being a tight connection between positive emotion and praise, see Stout (2020: 217). Stout, however, treats congratulations (including, for non-moral achievements, e.g., a student’s acquiring 17,000 Instagram followers) as expressions of praise, but many cases of congratulations won’t involve taking non-instrumental satisfaction in some exercise of agency, either because the congratulator does not have a non-instrumental desire that is satisfied by the act (true of me in relation to someone’s gaining many social media followers, assuming the latter qualifies as an action), or because one is congratulated for something non-agential (e.g., winning the lottery). On other objections (and responses) to the view that positive emotion is essential to praise, see Telech (2022: 5–6). 9 Darwall (2006); McKenna (2012); Macnamara (2013a); Shoemaker (2015); Telech (2021). 10 See Telech (2021). See also Lippert-Rasmussen (2022), though he denies that praise is essentially invitational; sometimes it issues demands. See also Mason (2019: 108–109). See also Macnamara (2013a: 908), who, though not taking praise to be invitational, writes, ‘moral agents receptively recognize themselves as having done something morally good with self-approbation’. Of gratitude, which he takes to be the positive analogue of resentment, Shoemaker (2013: 117) writes, ‘the aim is fundamentally communicative, to get the heighter [i.e., the praiseworthy agent] to robustly acknowledge what he has done and how he has made the heighted feel from the heighted’s perspective’. 11 Leech’s (1983: 136) Politeness Principle contains among its maxims, ‘The modesty maxim: Minimize the expression of praise of self; maximize the expression of dispraise of self’. Within the modern Chinese context, see Yuegou’s (1990: 246) Self-denigration maxim: ‘(a) denigrate self and (b) elevate other’. 12 Not all invitation attempts will generate valid (i.e., directive-providing) invitations, just as not all command or request attempts will generate valid commands and requests. 13 The ideas in this chapter benefited from presentations at the University of Oxford and Tilburg University. For helpful comments and questions, I  thank Rachel Achs, Mark Alfano, Cheshire Calhoun, Jules Holroyd, Max Kiener, Matt King, Elinor Mason, Leo Menges, Thaddeus Metz, and Andrea Westlund.

Further reading Macnamara, C. (2013). “ ‘Screw You!’  & ‘Thank You’,” Philosophical Studies, 165(3), 893–914, develops an influential communicative account of expressions of the positive and negative reactive attitudes as “recognitives,” speech acts expressing emotional recognition that seek acknowledgement in expression of self-reactive attitudes from their targets. Martin, A. (2014). How We Hope: A Moral Psychology. Princeton: Princeton University Press, chapter 5, develops an account of “normative hope,” an agent-directed stance wherein one upholds norms as aspirational and from which positive reactive attitudes (vehicles of praise, for the Strawsonian) are proposed to be intelligible. Telech, D. (2022) “Praise,” Philosophy Compass, 17(10), 1–19, provides a survey of possible views of praise’s nature (e.g., the judgment view, the action view, the emotion view, the conative-alteration view) and outlines several norms (a norm of merit/fittingness/desert, an epistemic norm, a norm of standing, a fairness norm) that may govern praise. Telech, D. (2021). “Praise as Moral Address,” Oxford Studies in Agency and Responsibility, 7, 154–181, introduces the invitational view of praise on which this chapter builds.

References Adams, R. M. (1985). “Involuntary Sins,” Philosophical Review, 94, 3–31. Arpaly, N. (2003). Unprincipled Virtue: An Inquiry into Moral Agency. Princeton: Princeton University Press. Arpaly, N., & Schroeder, T. (2014). In Praise of Desire. Oxford: Oxford University Press.

334

Communicating praise Calhoun, C. (2021). “Appreciating Responsible Persons,” Oxford Studies in Normative Ethics, 11, 9–28. Crocker, J. (2021). “Foreword: Praise in the Ecosystem and the Ecosystem,” in E. Brummelman (ed.), Psychological Perspectives on Praise. Abingdon-on-Thames: Routledge. D’Arcy, E. (1963). Human Acts. Oxford: Clarendon Press of Oxford University Press. Darwall, S. (2006). The Second-Person Standpoint: Morality, Respect, and Accountability. Cambridge, MA: Harvard University Press. Delin, C., & Baumeister, R. (1994). “Praise: More Than Just Social Reinforcement,” Journal for the Theory of Social Behaviour, 24(3), 219–241. Drew, P., & Couper-Kuhlen, E. (2014). “Requesting – from Speech Act to Recruitment,” in P. Drew & E. Couper-Kuhlen (eds.), Requesting in Social Interaction (pp. 1–34). Amsterdam: John Benjamins Publishing Company. Eshleman, A. (2014). “Worthy of Praise: Responsibility and Better-than-Minimally-Decent Agency,” Oxford Studies in Agency and Responsibility, 2, 216–242. Eylon, Yuval, & Heyd, David. (2008). “Flattery,” Philosophy and Phenomenological Research, 3, 685–704. Fricker, Miranda. (2016). “What’s the Point of Blame? A Paradigm Based Explanation,” Noûs, 50(1), 165–183. Gardner, J. (2007). Offences and Defences: Selected Essays in the Philosophy of Criminal Law. Oxford: Oxford University Press. Haji, I. (1998). Moral Appraisability. Oxford: Oxford University Press. Helm, B. (2017). Communities of Respect: Grounding Responsibility, Authority, and Dignity. Oxford: Oxford University Press. Holroyd, J. (2007). “A Communicative Conception of Moral Appraisal,” Ethical Theory and Moral Practice, 10(3), 267–278. King, M. (2014). “Two Faces of Desert,” Philosophical Studies, 169(3), 401–424. Leech, Geoffrey N. (1983). Principles of Pragmatics. London and New York: Longman. Lippert-Rasmussen, K. (2022). “Praising Without Standing,” The Journal of Ethics, 26(2), 229–246. Macnamara, C. (2013a). “ ‘Screw You!’ & ‘Thank You’,” Philosophical Studies, 165(3), 893–914. Macnamara, C. (2013b). “Taking Demands Out of Blame,” in D. J. Coates & N. Tognazzini (eds.), Blame: Its Nature and Norms (pp. 141–161). Oxford: Oxford University Press. Martin, Adrienne. (2021). “Personal Bonds: Directed Obligations Without Rights,” Philosophy and Phenomenological Research, 102, 65–86. Mason, E. (2019). Ways to be Blameworthy: Rightness, Wrongness, and Responsibility. Oxford: Oxford University Press. McKenna, M. (2012). Conversation and Responsibility. Oxford: Oxford University Press. Nelkin, D. (2015). “Psychopaths, Incorrigible Racists, and the Faces of Responsibility,” Ethics, 125(2), 357–390. Scanlon, T. M. (1998). What We Owe to Each Other. Cambridge, MA: Harvard University Press. Shoemaker, D. (2015). Responsibility from the Margins. Oxford: Oxford University Press. ———. (2013). “Qualities of Will,” Social Philosophy & Policy, 30(1–2), 95–120. Smith, A. (2005). “Responsibility for Attitudes: Activity and Passivity in Mental Life,” Ethics, 115, 236–271. ———. (2008). “Control, Responsibility, and Moral Assessment,” Philosophical Studies, 138, 367–392. Stout, N. (2020). “On the Significance of Praise,” American Philosophical Quarterly, 57(3), 215–225. Strawson, P. F. (1962). “Freedom and Resentment,” Proceedings of the British Academy, 48, 1–25. Telech, D. (2020). “Demanding More of Strawsonian Accountability Theory,” European Journal of Philosophy, 28(4), 926–941. ———. (2021). “Praise as Moral Address,” Oxford Studies in Agency and Responsibility, 7, 154–181. ———. (2022). “Praise,” Philosophy Compass, 17(10), 1–19. Tracy, J. L., Cheng, J. T., Robins, R. W., & Trzesniewski, K. H. (2009). “Authentic and Hubristic Pride: The Affective Core of Self-Esteem and Narcissism,” Self and Identity, 8(2), 196–213. Tracy, J. L., & Robins, R. W. (2007). “The Psychological Structure of Pride: A Tale of Two Facets,” Journal of Personality and Social Psychology, 92(3), 506–525. ———. (2014). “Conceptual and Empirical Strengths of the Authentic/Hubristic Model of Pride,” Emotion, 14(1), 33–37.

335

Daniel Telech Wallace, R. H. (2022). “A Puzzle Concerning Gratitude and Accountability,” The Journal of Ethics, 26, 455–480. Walker, T. (2013). “Requests,” in M. Sbisa & K. Turner (eds.), Pragmatics of Speech Actions (pp. 445– 466). Berlin: De Gruyter Mouton. Watson, G. (1987/2004). “Responsibility and the Limits of Evil,” in G. Watson (ed.), Agency and Answerability: Selected Essays. Oxford: Clarendon Press of Oxford University Press. ———. (1996/2004). “Two Faces of Responsibility,” in G. Watson (ed.), Agency and Answerability: Selected Essays. Oxford: Clarendon Press of Oxford University Press. Yuegou, G. (1990). “Politeness Phenomena in Modern Chinese,” Journal of Pragmatics, 14, 237–257. Zimmerman, M. J. (1988). An Essay on Moral Responsibility. Totowa, NJ: Rowman & Littlefield.

336

26 THE STANDING TO BLAME Matt King

1 Introduction There are (at least) three different ways our blame can go wrong.1 First, it can go wrong because of facts about the one being blamed. Perhaps they aren’t really blameworthy, and my blame would be unfitting or mistaken. Second, blame can go wrong because of facts about the instance of blaming itself. Perhaps my blame is too strong or the timing or circumstances are ill-chosen. Even if there’s no question of a blamee’s blameworthiness, my blame can be objectionable because of how or when I’m blaming. Finally, even if my blame is fitting and proportional, well chosen and placed, it can go wrong because of facts about the blamer themselves. If I’m unabashedly blaming another for the very thing I do regularly, then I am a hypocrite. If I blame another for something that’s none of my business, then I’m a meddler. In both cases, there is nothing wrong with my blame as such or its relation to the one blamed. My blame may be perfectly proportioned and entirely fitting. The problem is instead something about me, something that makes my blame objectionable. It is this last category that is the focus of discussions regarding the standing to blame. As Tognazzini and Coates (2021) put it, “The general idea here is that there may be facts about the person who is blaming that undermine her standing to blame. It’s not her place, she isn’t well positioned, she doesn’t have the authority, and so on.” Similarly, (though importantly, not equivalently), Marilyn Friedman (2013: 276) writes, This question [“Who are you to blame me?”] directly challenges . . . the blamer’s status to blame anyone for the particular sort of wrongdoing in question. That status is often referred to as either the “standing” or “authority” to blame. On this characterization, the standing to blame is required in order to legitimately blame. While this is compatible with such blame still being justified by extraneous moral considerations, such as the fact that the blame might have really good consequences, those who lack the standing to blame are nonetheless in some sense disqualified from blaming. Elsewhere, I have called this the “orthodox view” of the standing to blame.2

337

DOI: 10.4324/9781003282242-37

Matt King

As commentators have recently noted, however, there is no consensus regarding just what exactly the standing to blame picks out. Metaphors are routinely employed: there’s talk of being positioned or placed. A  common refrain points to the charge, “Just who are you to blame me?” Something about the blamer and their relative relation to the one blamed is grounds for complaint. For some, it is additionally grounds for dismissing or otherwise ignoring their blame. But despite these exhortations, there is no common view about what disputes regarding the standing to blame and its conditions amount to. There is some cautious consensus, however, that whatever the standing to blame is, it plays a role in hypocritical blame. Hypocritical blame has received outsized attention in the literature, and it strikes many as the clearest case of standingless blame. Similarly, though perhaps less paradigmatically, cases of meddling blame, where the transgression appears to be none of one’s business, are often cited as further instances3 where the standing to blame is in some way compromised.4 This chapter explores the standing to blame. Section 2 characterizes two different methodologies of explanation, providing a contrast in the phenomena to be explained. Section 3 examines the predominant strategies for explaining those phenomena, with a particular focus on hypocritical blame. Section  4 outlines some grounds for skepticism regarding the standing to blame. Finally, Section 5 briefly examines whether there are related issues involving the standing to praise. (Section 6 gives a brief conclusion.)

2  Structure of the phenomena To begin, we can distinguish two approaches to the standing to blame. These approaches differ in terms of the role given to the notion of standing and the phenomena they take to be primary. The first option seeks to explain what goes wrong with certain cases of blame, foremost among them, it seems, hypocritical blame. This approach to the standing to blame begins with instances of blame in which, while the grounds for blame are clearly impeccable, there is nonetheless something objectionable about the blamer blaming the blameworthy. Taking hypocrisy as a paradigm, such an approach might note that it is the blamer’s hypocrisy that renders their blame inappropriate under the circumstances. The second step would be to explain why their hypocrisy renders such blame inappropriate in the way that it does. So, on this approach, the standing to blame is a mechanism for explaining what goes wrong with certain instances of blame. Still, it is open to this first approach to opt for a different explanatory notion; the standing to blame is merely one option for what’s wrong with hypocritical blame. Since this approach leaves open whether the notion of standing is explanatory (even if it is a plausible or intuitive option), I call this the unsettled approach. In contrast, a different approach takes the central phenomenon to be the kind of status that hypocrites lack. Thus, on this alternative approach, we begin not with the wrongness of certain instances of blame, but with a distinctive kind of missing status or complaint. Sticking with hypocrisy as the paradigm, this approach takes it as plain that hypocrites are missing a status required for legitimate blame, and so seeks to explain that status they are missing. Since this approach presumes that the standing to blame either explains the wrongness of the blame or otherwise underlies the phenomenon to be explained, I call this the settled approach. (To put the contrast briefly, an unsettled approach primarily looks to explain the problem with certain instances of blame, and is consequently permissive about how to best explain what goes wrong, whereas a settled approach presumes the problem lies in a kind 338

The standing to blame

of missing status, and so primarily seeks to explain the lack of standing as such. The former is unsettled about the theoretical role the status of the blamer ought to play, while the latter takes that status to either be explanatory or the thing to be explained.) It can be difficult to disentangle the approaches. At times it is ambiguous whether a given account is settled or unsettled about the role of the standing to blame. Furthermore, the contrast here illustrates that the two approaches need not be competitors, as they aim to explain different things. Indeed, it can be tempting to think that only settled approaches are giving an account of the standing to blame. Unsettled approaches, since they might well appeal to alternative explanatory mechanisms, may or may not speak at all to the standing to blame, if that term is restricted to a special kind of status.5 Nevertheless, it is worth distinguishing these two approaches because doing so helps reveal certain constraints on our theorizing. In particular, if the kinds of cases in which blame of the blameworthy is objectionable due to facts about the blamer diversify considerably, this may put some pressure on settled approaches. The broader the class of blamers that might not be able to legitimately blame, the more difficult the prospects become for distilling out a single status they all share that’s univocally explained. Unsettled approaches, which only seek to explain what has gone wrong in various cases of blame, have more theoretical flexibility in how to unify those cases or explanations. Approaches that are presumptively committed to a single status – lacking the standing to blame – might face special challenges as the kinds of cases diversify. As a brief illustration, consider meddlesome blamers. As with hypocrites, many have taken meddlesome blamers to also lack the standing to blame, but reflection suggests they are quite different from one another. Notice, first, that the role that status might play is different for each. In the case of hypocrites, most people can blame without problem; it is only hypocrites that are in the wrong class. But for meddlers, most people would be meddling under those circumstances; it’s only a select few who have the status (i.e., those for whom the matter is their business or who stand in the proper relationship to the one blamed). This means that if the relevant status here is the standing to blame, it would have to be both possessed by most potential blamers (since they are non-hypocritical) but also missing from most potential blamers (since they are would-be meddlers). While this doesn’t show that there could not be such a status, its geometry looks initially challenging.6 Of course, one might draw a different lesson from these differences and conclude that hypocritical blame is importantly different from meddlesome blame. Perhaps, then, they should be treated differently. Indeed, Patrick Todd notes that hypocritical blame involves a distinctive kind of complaint, “Who are you to blame me?” Meddlers, by contrast, are best met with the reply, “Mind your business.” As he puts it, “these are importantly different replies. One says: you lack the standing with me to say what you are saying. Another says: you lack the standing with morality to say what you are saying” (2019: 349). It thus remains an open question how unified the relevant phenomena to be explained are.

3  Dominant strategies For now, we can set aside questions regarding the best explanatory approach and focus instead on the supposed paradigm of hypocritical blame in order to consider options for how to best characterize the standing to blame. One thing is clear: the standing to blame should not be modeled on standing as it functions in other contexts. It might be tempting to treat the standing to blame like the standing to bring legal suit7 or the standing to command 339

Matt King

(i.e., authority).8 But these analogues are decidedly inapt, for each of the latter are enabling conditions on the very activity in question, such that when the condition goes unmet, the activity is impossible. One simply cannot bring suit unless one has the standing to do so. There is, strictly speaking, no such thing as a standingless suit. If one is found to lack standing then whatever motions one went through (e.g., filing paperwork and addressing the court) do not achieve their legal goal. Similarly, while you can tell someone what to do in the shape of a command, voicing it as an imperative, without the requisite authority you simply fail to issue a command.9 Whatever else is true of standingless blame, however, the blamers do not fail to blame. Indeed, it is their very blaming that is at issue. At most, the idea that the standing to blame is required to blame must be interpreted to mean that one cannot legitimately blame without the standing to do so. But this is a different sort of complaint than is captured by lacking legal standing or authority. And this is a significant result in and of itself. If the very notion of the standing to blame gathers intuitive force from its resemblance to standing elsewhere, then to the extent that it departs from that notion, the intuitive case for the standing to blame may be considerably weakened. (Such a result could motivate unsettled explanatory approaches to look elsewhere for explaining what’s wrong with the relevant cases of blame.) If we dispense with modeling the standing to blame on parallel notions of standing, the next obvious candidate is as a right or entitlement to blame.10 Unfortunately, few theorists have been explicit about the nature of this right. Two notable exceptions are Kyle Fritz and Daniel Miller (2022) and Ori Herstein (2017).11 Fritz and Miller argue that the standing to blame should be understood as a defeasible privilege (in the Hohfeldian sense) to blame the blameworthy.12 It is objectionable to act in ways to which one doesn’t have the right. In order to permissibly walk across your field, I must have the privilege to do so. Similarly, blame would be objectionable if I lack a privilege to do it, for then I would have a duty to refrain. Fritz and Miller argue that the standing to blame is the relevant permission involved. Blamers who lack the standing to blame lack permission to blame, and their blame is objectionable as a result.13 Their explanation for why hypocrites lack the privilege to blame is that that right is grounded on the moral equality of persons. It is in virtue of our moral equality as persons that we all have a (defeasible) right to blame each other. The problem for the hypocrite is that they fail to respect the moral equality of persons by making an exception of themselves. Hypocrites therefore lack the right to blame, and so their blame is objectionable. In contrast, Herstein understands the standing to blame as a kind of power.14 He argues that blamers normally give special reasons to the ones they blame, what he calls “directivereasons.” These reasons direct the blameworthy to (among other things) change their ways and feel remorse merely because the directive has been given. (In this way, the standing to blame looks like a kind of authority.) When hypocrites blame, since they lack the relevant power, the blameworthy are permitted to dismiss their blame without even considering it. In so doing, they do not culpably neglect any directive-reasons because the obligation to comply with those reasons requires standing. I don’t have the space here for further elaboration or even conscientious consideration of either view. Instead, I’ll (briefly) gesture toward some of the limitations of both accounts, which highlight some general challenges any view about the standing to blame faces. First, consider blame’s notorious heteronomy. There is an enormous range of behaviors, attitudes, and the like, all of which are plausible candidates for blame (or, for involving 340

The standing to blame

blame). An account of the right to blame should also plausibly respect this heteronomy (or, at least, explain it away). For instance, Herstein’s account is limited to the normative force of directives, a kind of speech-act of which blaming is but one member. Many take privately held hypocritical blame to be no less objectionable than expressed, but such blame does not exhibit the features of any directive. For different reasons, we might similarly worry about Fritz and Miller’s defeasible privilege, precisely because it is unclear why we need permission to blame in general. While blame can have harmful effects in certain circumstances (when expressed, say), it can also be exceptionally benign. Second, whatever the prospects for either account with respect to hypocritical blame, extending them to meddlesome blame appears fraught. (Here we reencounter one of the constraints on theorizing from the previous section.) What’s curious is that in cases of hypocritical blame most would have the standing to blame under the circumstances, only the hypocrite’s standing is compromised, whereas in cases of meddlesome blame most fail to have the standing to blame, only those whose “business” it is possess the relevant right. A challenge here is to specify a general right – the right to blame – that can make sense of these very different structures. Notably, this challenge persists even if one takes hypocritical and meddlesome blame to represent distinct problems. For the right to blame will still be present in both kinds of cases, since both involve blamers blaming. To the extent, therefore, that the right to blame is a right to blame generally, one’s account of that right will have to reckon with the dynamics of blame in all its circumstances, including meddlesome cases. Neither limitation is decisive, of course. Rather, both reinforce how more work on clarifying the nature of the standing to blame remains to be done.

4  Varieties of skepticism There are principally three ways in which the dominant strategies have been resisted. In each instance, we can see again the importance of distinguishing the two basic approaches to the phenomena. Some defend skepticism about the relevant data. For example, Macalester Bell (2013) contends that, among other things, hypocritical and meddling blamers do not necessarily lack the standing to blame. According to Bell, blame has many different functions and values, only some of which are thwarted by particular instances of improper blame. If that’s right, then if we’re after an explanation for what makes such instances of blame improper, it won’t be that such blamers lack standing. Alternatively, if we’re after an explanation for why certain blamers lack a certain status, then hypocritical and meddling cases won’t be relevant. A different kind of skepticism concerns our prospects for explanation. In arguing for a unified account of the standing to blame, Patrick Todd concludes that there may be no explanation for why certain classes of blamers lack the standing to blame: they just do (2019: 371–372). In this way, the standing to blame might be basic, having little more to say in its explication than that it is the status lacked by certain kinds of improper blamers or reflected by certain kinds of complaint. If that’s right, then our explanatory ambitions may be necessarily frustrated. If our objective is to appeal to the standing to blame to explain certain cases of problematic blaming, we may succeed in securing an explanation, even if it will be ultimately truncated, running out in an appeal to a basic kind of fact.15 Alternatively, if our objective is an explanation for why certain blamers lack a certain status, then it looks like such a goal is ruled out. 341

Matt King

One might instead defend skepticism about the standing to blame itself (King 2019, 2020). Elsewhere, I argue that there is no good case for a right to blame that could do the requisite explanatory work. Since I take the primary task to be explaining what goes wrong with, in particular, hypocritical and meddlesome blame, my position is that a right to blame doesn’t fit the phenomena and isn’t required for successful explanations of them. However, if what we’re after is explaining why certain blamers lack a certain status, then my position appears to be a non-starter, since it denies that there is such a status. Obviously, I’m partial to an unsettled approach. We shouldn’t presume that there is anything like the standing to blame. Nonetheless, there is certainly an intuitive pull about the charge, “Who are you to blame me?” And those who favor a settled approach to the phenomena are likely to emphasize retorts of this form to hallmark cases of improper blame. I am additionally skeptical that such replies in fact constitute a legitimate charge rather than serving as a useful distraction. It’s worth noting, for instance, how being able to silence would-be critics would be self-interestedly advantageous. It would in many ways be to our benefit if we could simply dismiss the criticisms of others, rather than reckoning with their substance or the further messy dynamics of engagement over our wrongdoing. To the extent that we wish to avoid those charges, we will be motivated to find justifications to de-legitimize their source. Establishing these claims, however, exceeds the bounds of the present discussion.

5  The standing to praise? Just as hypocrisy has dominated discussions of the standing to blame, so, too, does blaming dominate discussions of the ethics of responding to the responsible. (Indeed, the relevant subliterature here typically goes by the name “the ethics of blame.”) No doubt a large reason for the asymmetry is that there are more obvious moral concerns regarding blame than praise. If one takes blame to be punitive, or at least unpleasant, then one will likely be imposing costs or potential harms on the one blamed. Praise, by contrast, is generally pleasant, so at worst you might be directing an undeserved benefit someone’s way. But there are good reasons to consider whether similar issues might arise for praise. (And, if not, why not.) We can see with the standing to praise a parallel bifurcation regarding the relevant explanatory objective that we found in the case of blame. On the one hand, we might start with cases of objectionable praise and look for an explanation for why they’re objectionable. Such a project might point us toward a missing standing to praise (or, of course, it might just point us to a host of disparate considerations). On the other hand, we might frame our explanatory ambition in terms of understanding why certain praisers lack a kind of status (i.e., why they are in no position or aren’t well placed to praise). Initially, we might have our doubts. The predominant case of standingless blame  – hypocritical blame – finds no obvious analogue on the positive side of things. One who is praiseworthy for the same things is just as well placed as anyone, perhaps more so, to praise others for that thing.16 Moreover, we might find a right to praise curious on its face. Whatever the attractions of the idea that we need to clear some bar in order to properly render blame do not obviously attend to directing acclaim others’ way. The charge, “Just who are you to praise me?” is not so common an expression. Nonetheless, there do seem to be cases of problematic praise that parallel those of blame. Consider meddling blame. In the typical case of meddlesome blame, the one blamed is apt to respond with something like, “Mind your own business!” But if we switch from blame 342

The standing to blame

to praise in the case, it isn’t clear we substantially alter the relevant dynamic. The good doings of others may be just as likely to lie outside one’s business as their bad doings; certain strangers may be poorly positioned to praise me for my accomplishments. If anything, the wrongdoing of others would seemingly give us greater license to intrude into their “business.” Whether the standing to praise is as robust a topic as its negative counterpart is far from clear, but it certainly warrants further attention.17

6 Conclusion In many ways, examination of the standing to blame and its related phenomena remains in its infancy. The literature has yet to coalesce around clearly demarcated questions or concerns. As such, it remains a bit conceptually messy. Different theorists seemingly have separate explanatory concerns, so we are not yet settled on the relevant targets. Additionally, there is significant skepticism, both about the motivating cases and the candidate right involved. As a modest recommendation for future work, we might do better by abandoning the “standing to blame” altogether, and instead clearly articulating our explanatory targets. Are we trying to explain what goes wrong with particular cases of blame? And in what ways does it go wrong exactly? Are we giving an account of a presumed right to blame? And what is the nature of that right? Are we explaining the grounds for a specific kind of complaint? And in what ways does that complaint generalize? As it stands, there appears to be no one thing unequivocally captured by “the standing to blame.” Unless we can sharpen our collective focus, we risk talking past one another in our efforts to better understand the ways in which blame of the blameworthy can still go wrong.18

Notes 1 These are not intended to be exhaustive. 2 See King 2019. Fritz and Miller (2022) criticize my characterization, arguing that many treat the standing to blame as a right to blame that is perfectly compatible with such blame being allthings-considered appropriate or justified even when that right is absent. But this misunderstands the orthodox view. What we all have to agree on, it seems to me, is that the standing to blame is a status (because some can have it and others can lack it) and that its absence renders one’s blame in some sense inappropriate. If one thinks that whether or not one has the standing to blame makes no difference to any kind of appropriateness of one’s blame, it’s unclear why we should focus distinctively on a status-notion at all, rather than just focusing on whether blame is justified or not (all things considered) under the circumstances. 3 These two categories do not exhaust the potential places we might find standingless blame. Part of what is at issue in the literature is precisely which kinds of objectionable blame are relevant. 4 I will alternate between talk of “standingless blame” and “compromised standing.” I mean the two to be roughly interchangeable, remaining neutral so far as it goes regarding whether standing is a digital or analog notion. 5 See also Tognazzini and Coates 2021 for an alternative taxonomy. 6 A separate source of constraint may be that the myopic focus on hypocrisy unduly limits the scope of inquiry. See Seim 2019 for a discussion focused on meddling. 7 See, e.g., Sabini & Silver 1982. 8 See, e.g., Darwall 2010. 9 Compare the legal standing to marry. If one is already married, then one lacks the standing to marry, and so cannot marry. One can fill out the forms, make various promises, hold various ceremonies, but none of that can change one’s legal status.

343

Matt King 0 See, e.g., Cohen 2006; Radzik 2012. 1 11 I can’t do justice to their full arguments here. In any case, my aim is to outline their strategies, not to analyze them. 12 In Hohfeld’s classic analysis, there are four types of rights: claims, privileges, powers, and immunities. Privileges are rights that imply the rightholder has no duties to refrain from the substance of the right (i.e., they are at liberty to do the thing in question). 13 It’s a bit unclear exactly what the nature of this objection is for Fritz and Miller. They allow that one could be “all things considered” justified in blaming hypocritically, despite thus lacking the right to blame, under circumstances where, say, some large value could be obtained by blaming (e.g., saving many lives). Presumably, the idea is that when one lacks a privilege to blame, one has a duty not to blame, but that duty can be overridden in particular circumstances. But it isn’t the lack of the privilege that explains the duty, but rather the duty that explains the lack of privilege. I may have the privilege to whistle showtunes. It would nonetheless (normally) be wrong for me to whistle showtunes during a memorial service. The explanation for why it would be wrong, however, is that it would be rude (or disruptive or disrespectful or whatever). Because of those features, I have a duty to refrain, and that of course implies that I lack the privilege to do it. But the objection to such whistling would be that I’m being rude, not that I lack the right to do it. 14 In the classic Hohfeldian analysis, powers are rights that allow the rightholder to change the rights and duties of another. One with power of attorney can perform various legal actions on behalf of another, thereby changing their legal status and correlative rights. 15 Though, Todd divides the phenomena to be explained, treating meddlesome blamers and hypocritical ones separately. 16 Though see Lippert-Rasmussen 2022; Telech 2022 for examinations of potential parallels. 17 See also Fritz & Miller 2022 for an examination of potential parallels between the standing to blame and the standing to forgive. 18 My thanks to the workshop audience and to Maximilian Kiener and Daniel Telech for written comments on an earlier draft.

Further reading Cohen, G. A. (2006). “Casting the First Stone: Who Can, and Who Can’t, Condemn the Terrorists?” Royal Institute of Philosophy Supplement, 58, 113–136 is widely regarded as (one of) the first significant publication on the topic. Fritz, K.,  & Miller, D. (2018). “Hypocrisy and the Standing to Blame,” Pacific Philosophical Quarterly, 99, 118–139 outlines one of the central views on at least hypocritical blame. Fritz, K., & Miller, D. (2022). “A Standing Asymmetry Between Blame and Forgiveness,” Ethics, 132(4), 759–786 is one of the few texts to try to give a positive account of the nature of the right to blame and contains a good discussion of various theoretical approaches. King, M. (2019). “Skepticism About the Standing to Blame,” Oxford Studies in Agency and Responsibility, 6, 265–288 makes the case for skepticism about the standing to blame, understood as the right to blame. Another good piece defending skepticism is Bell, M. (2013). “The Standing to Blame: A Critique,” in D. J. Coates & N. Tognazzini (eds.), Blame: Its Nature and Norms. Oxford: Oxford University Press. For an alternative framework in considering the standing to blame, see Tognazzini, N., & Coates, D. J. (2021). “Blame,” in E. Zalta (ed.), The Stanford Encyclopedia of Philosophy (Summer 2021 Edition), Edward N. Zalta  (ed.), .

References Bell, M. (2013). “The Standing to Blame: A Critique,” in D. J. Coates & N. Tognazzini (eds.), Blame: Its Nature and Norms. Oxford: Oxford University Press. Cohen, G. A. (2006). “Casting the First Stone: Who Can, and Who Can’t, Condemn the Terrorists?” Royal Institute of Philosophy Supplement, 58, 113–136. Darwall, S. (2010). “Authority and Reasons: Exclusionary and Second-Personal,” Ethics, 120, 257–278. Friedman, M. (2013). “How to Blame People Responsibly,” Journal of Value Inquiry, 47, 271–284.

344

The standing to blame Fritz, K., & Miller, D. (2018). “Hypocrisy and the Standing to Blame,” Pacific Philosophical Quarterly, 99, 118–139. ———. (2022). “A Standing Asymmetry Between Blame and Forgiveness,” Ethics, 132(4), 759–786. Herstein, O. (2017). “Understanding Standing: Permission to Deflect Reasons,” Philosophical Studies, 174, 3109–3132. King, M. (2019). “Skepticism About the Standing to Blame,” Oxford Studies in Agency and Responsibility, 6, 265–288. ———. (2020). “Attending to Blame,” Philosophical Studies, 177(5), 1423–1439. Lippert-Rasmussen, K. (2022). “Praising Without Standing,” The Journal of Ethics, 26(2), 229–246. Radzik, L. (2012). “On the Virtue of Minding Our Own Business,” Journal of Value Inquiry, 46, 173–182. Sabini, J., & Silver, M. (1982). Moralities of Everyday Life. Oxford: Oxford University Press. Seim, M. (2019). “The Standing to Blame and Meddling,” Teorema, 38, 7–26. Telech, D. (2022). “Praise,” Philosophy Compass, 17(10), e12876. Todd, P. (2019). “A Unified Account of the Moral Standing to Blame,” Noûs, 53, 347–374. Tognazzini, N., & Coates, D. J. (2021). “Blame,” in E. Zalta (ed.), The Stanford Encyclopedia of Philosophy (Summer 2021 Edition), Edward N. Zalta (ed.), .

345

27 APOLOGY AND FORGIVENESS Transaction, conversation, or joint narrative?1 Andrea Westlund

The pivotal scene of Ingmar Bergman’s movie Fanny and Alexander has 10-year-old Alexander locked in an intense psychological struggle with his step-father, the stern disciplinarian and local bishop who married Alexander’s mother after her first husband’s untimely death. Alexander has made up a semi-fantastical tale about the bishop and his former family, effectively accusing the bishop of being responsible for their deaths. First, the outraged bishop goads Alexander into lying about what he did. Then, he uses the threat of corporal punishment to pressure Alexander into confessing both his slander and his perjury. Confession in hand, he proceeds to administer punishment anyhow. And just when you might think it’s over, he demands that Alexander personally ask his forgiveness. Here, Alexander draws the line: “I will never apologize!” Only, as it turns out, he will – under threat of further punishment. The whole ritual unfolds as a horrific display of power, rooted in misplaced resentment and injured pride, and ending in submission rather than anything remotely like reconciliation (Bergman 1982). Apology has been given a prominent place in standard accounts of forgiveness, often being cast a paradigmatically good reason to foreswear hard feelings toward a wrongdoer. A wrongdoer’s sincere apology allegedly allows one to give up one’s moral anger without compromising one’s self-respect. (Let us call this the standard view of apology, so as to correlate it with the standard view of forgiveness.) But apology rituals like that recounted previously appear to have a starkly different structure and function: though the bishop acts under the guise of moral education, it is clear that Alexander is meant to be put in his place and compelled to show due regard for the bishop’s authority. Here, the demand for an apology looks like nothing more than an effort to “pay back” the humiliation of having been slighted, or to “save face” and restore one’s injured pride. According to Martha Nussbaum (2016), such examples are not an outlier. Nussbaum argues that our apology practices are deeply rooted in a morally repugnant, “transactional” ritual, in which wrongdoers are required to abase themselves before their victims as a way of compensating for the injury inflicted by an offense. Nussbaum argues not only that this ritual is ultimately futile (payback cannot right a wrong or restore moral balance), but that to demand an apology is to intrude on another’s “inner world” (Nussbaum 73) and even to exercise a form of violence against the other’s self. DOI: 10.4324/9781003282242-38 346

Apology and forgiveness

The standard view and the transactional model (as interpreted by Nussbaum) thus give starkly different assessments of apology’s significance in forgiveness. On the standard view, apologies have reason-giving power: they can make it rational for a victim to forgive a wrongdoer, or to resume relations with a wrongdoer, or to change their stance toward the wrongdoer in some other relevant way.2 On the transactional model, it is the victim who exercises power, extracting “payment” from the wrongdoer and granting forgiveness in return. According to the standard view, apology is morally powerful. Understood transactionally, it appears instead to be morally disempowering. Can the standard view be redeemed? I take the standard view to reflect a widely shared view of apology’s power and significance, which would need to be explained by any viable model of apology. As Nussbaum construes it, the transactional model would seem to debunk rather than explain or support the standard view. In this chapter, I explore the limitations of the transactional model and contrast it with two other models of apology: a conversational model and a narrative model. Each of these alternative models illuminates important aspects of our moral practices of apology and forgiveness, in ways that help to redeem the standard view of apology’s role and value (that is, the view that apologies have morally valuable, reason-giving power). I do not regard the conversational and narrative models as competing with one another, but as playing largely complementary roles. My aim is not to defend one over the other, but rather, to explain how each avoids the pitfalls of the transactional model.3 I will, however, highlight a feature of the narrative model that has not received as much attention it deserves. Specifically, I argue that a form of joint storytelling plays a larger role in apologies than has hitherto been widely recognized, and that this model helps us to see how apologies can be mutually empowering, rather than amounting to an exercise of power by one party over another. Let’s begin by considering our target practice in more detail. An apology is typically understood to have the following components: one must acknowledge a wrongdoing and take responsibility for it, one must express some regret or remorse over the wrongdoing, and one must express a commitment to change, at least in the sense that one will not repeat the wrong.4 Apologies that lack one or more of these features are at very least flawed, and may not even count as apologies at all. Locutions such as “I’m sorry if you were offended” or “I’m sorry that happened to you”, for example, are clear examples of bad or failed apologies. In such cases, the speaker might express sadness or perhaps even regret over the impact of an action, but fails fully to acknowledge the wrongdoing, take responsibility for it, or commit to change. Often, such attempted apologies come across as entirely insincere, such that even the expression of sorrow or regret rings hollow. Apologies that make reference to mitigating factors are rather different and more complicated, though very common. One might, for example, say something like “I’m so sorry, I didn’t realize . . .” or “I am sorry, I didn’t mean to . . .” Here, the wrongness of the act may be acknowledged, and the apologizer may express sincere regret or remorse and a commitment to change. (Now that she knows better, she is determined not to do the same thing again.) But whether or not the speaker takes full responsibility for the act is less clear. On the one hand, a complex apology will fail if the apologizer appears to be reaching for excuses instead of taking responsibility where she should. On the other hand, however, if the mitigating factors to which she appeals are such as to release her from blame, or at least to reduce her blameworthiness, one might think that such locutions are not, properly speaking, apologies at all – perhaps such “apologizers” are instead asking to be excused.5 347

Andrea Westlund

Sometimes we even use such phrases as “Please excuse me . . .” to flag that there are mitigating factors in play. Yet apologizers who lacked full knowledge of or control over what they were doing often resist the thought that they bear no or even reduced responsibility for the act or the harms it incurred, and they often feel and express genuine remorse – especially when the harms were severe. Sometimes this is because the apologizer rightly takes responsibility for their shortcomings – they could have learned more, or exercised greater care. But even when it would seem cruel to blame a person for what they could not have foreseen or for what they genuinely misunderstood or could not control, it often still matters to the would-be apologizer that they were in fact the one who did the thing or caused the harm in question, and they may express sincere remorse and a desire to atone or make amends, even while simultaneously maintaining that it was not their intent to harm. (Consider, for example, life-changing cases in which a person causes an accidental death.) Although the stakes are lower in the case of more minor, “everyday” slights and injuries, there, too, apologizers sometimes offer apparently exculpatory explanations of their actions even while expressing sincere remorse and explicitly taking responsibility for the harms they have incurred. I will return to these cases shortly, since I think the narrative model can help us to better understand their complexities. Let us return, for now, to the standard view and consider the relationship between apologies and forgiveness as depicted therein. On the standard view, to forgive is to forswear resentment (and perhaps other negative emotions or attitudes) for good moral reason. Forgiveness can seem highly puzzling on this view: if your resentment is justified, and if forgiveness entails giving up your resentment, what could justify forgiving? If you give up your resentment without good reason, you seem to be condoning a wrong act and, since you were the one who was wronged, you seem also to be manifesting a lack of appropriate selfrespect. So, in order for forgiveness to be “uncompromising”, as Pamela Hieronymi puts it, we must identify the right sort of reason to give up one’s justified resentment. Jeffrie Murphy (1988) influentially argues that in order to merit forgiveness a wrongdoer must in some way distance or separate themselves from the wrong act. Murphy identifies repentance as the clearest way in which wrongdoers can separate themselves from their acts: when a wrongdoer repents, according to Murphy, “he withdraws his endorsement from his own past behavior” (Murphy 26) and joins the victim in condemning it. Murphy argues that several other commonly cited reasons for forgiving (that the wrongdoer had good motives, or has suffered enough, or has been humbled or humiliated, or “for old times’ sake”) can all be taken to involve some form of separation between the wrongdoer and the act, even if somewhat less obviously or directly than repentance.6 Hieronymi’s account of uncompromising forgiveness is similar: according to Hieronymi, justified resentment protests a threatening claim made by a wrong act – a claim that says, in effect, that one may be treated in a way that is incompatible with one’s worth. The wrongdoer’s sincere repentance or apology withdraws their endorsement from the claim made by a past act, thus rendering it no longer a present threat. Ongoing resentment, in such a case, would have no point. More recently, Christopher Bennett (2022) has argued that an apology is an expressive act of self-dissociation, through which the wrongdoer joins the moral community in condemning the wrong act and makes a commitment to stay within the moral community (Bennett 17). This account, too, locates the normative force of apology in an act of self-distancing undertaken by the apologizer. In all three formulations, the wrongdoer not only distances themselves from their own past self, they do so via an expression of solidarity with the victim or the moral community. 348

Apology and forgiveness

Understood transactionally, it is hard to see how an apology could serve to distance the wrongdoer from her previous act in the right sort of way. On the transactional model, the initial moral injury to the victim is, in effect, answered by a new injury to the wrongdoer – the humiliation of abasing oneself before one’s victim. Murphy, as noted, suggests that having been humiliated can, like repentance, serve as a reason to forgive. But it is not entirely clear how it is supposed to do so. Self-abasement does not, in itself, entail sincere repudiation of one’s former self or past action. It might be done reluctantly, resentfully, or under duress, as in Alexander’s case. It might be a mere act, undertaken to manipulate or curry favor with one’s addressee. If the self-abasing agent does genuinely repudiate her earlier acts, this would seem to be a fortunate coincidence rather than an inevitable result of her performance. Of course, genuine feelings of humiliation might accompany acts of self-abasement, particularly if those acts are compelled. Nussbaum, as we’ve seen, imagines the self-abasing apologizer as experiencing a kind of violation of or intrusion into the self, which could lead to feelings of alienation or dissociation. When acting under duress, for example, one might not recognize oneself in one’s words and deeds. One might even feel that one is betraying oneself. But this is not, of course, the sort of distancing or dissociation that can function as a reason to forgive: the self-repudiation imagined by defenders of the standard view purports to express the wrongdoer’s authentic self, rather than estranging her from it. In short, the self-abasing words, actions, and postures required by the transactional account are, at best, only contingently related to the morally significant form of self-repudiation required by the standard view, and thus at best contingently related to the reasongiving power of a (good) apology.7 Put bluntly, the transactional model empowers victims by putting them in a position to humiliate wrongdoers – but the apologies they extract are likely to leave them with the wrong kind of reason to forgive.8 Of course, if we could simply opt not to pursue apologies in a transactional manner, this line of critique won’t seem very deep. Nussbaum suggests, in passing, that receiving an apology is quite a different matter from extracting one (Nussbaum 141). So why not just rest with the thought that apologies are only meaningful if freely and sincerely offered, and that it is counterproductive to demand or extract one? At the heart of Nussbaum’s critique, however, is the thought that to treat apology as a prerequisite for forgiveness is to treat it as something that can be legitimately demanded of a wrongdoer, and that this is already implicitly to treat apology in an incipiently coercive way – as a form of payment, which, if not offered freely, may be extracted from wrongdoers by victims. The idea that apologies may be owed and even in some sense demanded, and that they are a prerequisite for forgiveness, does seem integral to the standard view. So, it seems, if we are to preserve a central role for apology in practices of reconciliation, we must be able to make sense of how “paying one’s dues”, in the form of an apology, could possibly have the right sort of reason-giving power. For this, we will need a different model, or metaphor, for guiding our understanding of apologies. Fortunately, we find at least two other models of apology in the literature, to which I will refer as the conversational model and the narrative model. I will briefly outline each of these models, and show how each accommodates the idea that apology can be a good reason to forgive. Neither, I argue, collapses into the bare transactional model, though each makes room for relevant demands and expectations. On a conversational model, an apology is best understood not as a part of a (covertly coercive) transaction, but as something analogous to a move in a conversation. Generally, 349

Andrea Westlund

of course, an apology literally is a move in a conversation, not just analogous to one, which can be confusing. To properly understand the conversational model, it is thus important to focus on apology as a performative – a bit of speech that not only says something but also does something with a particular moral significance. (It is, in Austinian terms, an illocutionary act.) There are various ways of construing what an apology does, and I won’t commit myself to a specific, detailed account here. As a reasonable approximation, however, one might describe an apology as a speech act that signals the apologizer’s acceptance of a new and different attitude toward what they have done and its implications. When successful, it constitutes a reconciliation not between victim and wrongdoer (though it might have such reconciliation as a further aim and help to make it possible) but between previously divergent interpretations of or attitudes toward the wrongdoer’s action and the victim’s moral standing. A conversational model of apology fits into a broader, conversational model of moral responsibility, such as that developed by Michael McKenna, on which our practices of blame and forgiveness traffic in the negotiation of moral understanding and the (re)alignment of reactive attitudes, rather than in the extraction and payment of moral debt. For McKenna, an agent’s action carries moral meaning insofar as it is indicative of her quality of will: her action may indicate good will toward another, or ill will, or indifference, or something else in between. McKenna refers to this sort of meaning as “agent’s meaning”, drawing an analogy with Gricean “speaker’s meaning” (what the speaker intends to communicate, in uttering a particular sentence). A wrong act is like an opening move in a conversation, to which the “audience” may react by holding responsible and blaming the wrongdoer. The wrongdoer may then react in turn, offering a plea or excuse or some sort, or a justification, or an apology. To follow through on the analogy, each of these moves must itself be regarded as an act that expresses or signals an agent’s quality of will – not just as an utterance with a particular semantic content.9 The conversational account, thus understood, easily accommodates the standard account both of apology itself and of apology’s relationship to forgiveness. A sincere apology acknowledges that the victim is right to hold the wrongdoer responsible; indeed, the apologizer joins with the victim in taking up that stance. The apologizer effectively holds herself responsible, endorsing the victim’s resentment and condemnation of the act. She creates the needed distance between herself and the previous act by repudiating the quality of will it expressed, and expressing a different, morally appropriate quality of will toward the victim. This apologizer does not exchange her own suffering and humiliation for forgiveness (even if she does feel anguished or humbled by the experience); rather, her act of apologizing sends a new message that makes it possible for the victim to forgive her without compromising her own self-respect. Nor does the conversational model appear to collapse into the transactional model in the way envisioned by Nussbaum. As a model for our responsibility practices more broadly speaking, it clearly involves a place for normative demands. To blame another is to hold her to account, subjecting her to a demand for a response. As noted, that response might be a justification, an excuse, or an apology. If the wrongdoer defiantly refuses to respond, she violates a moral requirement, and moral relations with her will continue to be disturbed. The victim may continue to subject her to negative reactive attitudes, withdraw trust, and exclude her from certain interactions and relationships. Others in the moral community may do likewise. But there is no implication that the victim is entitled to coercively extract an apology from the wrongdoer as a way of evening the score. To be a bit tongue-in-cheek 350

Apology and forgiveness

about it, the bishop’s coercive response to Alexander isn’t good conversation. In fact, it isn’t really conversation at all: it is an exercise of brute power, a type of move that doesn’t properly figure in the relevant moral practice, and which subjects wrongdoers to the wrong sort of requirement.10 The conversational model is powerful and appealing, not least because it helps us to make sense of the standard view of the relationship between apology and forgiveness (or between apology and reconciliation more generally). Nonetheless, insofar as it tends to focus our attention on canonical expressions of responsibility, remorse, and commitment, it abstracts somewhat from what I’ll call the narrative framing of an apology. A narrative model of apology is closely related to the conversational model, insofar as it treats the interaction between wrongdoer and victim as having the structure of a dialogue (not a transaction). But it is focused more broadly on the narrative structure and content of each party’s contribution to that dialogue, rather than (more narrowly) on the moral meaning of the expressive acts that constitute the main dialogical moves. Charles Griswold (2007), for example, argues that each party to such a dialogue offers a narrative that frames their actions and experiences in a particular way. A narrative, as he construes it, is not a mere chronology of events, but rather a “meaning-making unification” (Griswold 184) thereof, which brings together multiple perspectives on a chain of events and the parties involved in it. These perspectives include the parties’ own past and present perspectives on themselves as well as their past and present perspectives on one another. A “dialogical” narrative of forgiveness is developed by the wrongdoer and victim together, through the “giving and receiving of accounts, and the demand for as well as offering of explanations that make intelligible one’s past deeds and warrant trust in promises about the future” (Griswold 184). There is, to be sure, a canonical “script” for apology narratives, and for the larger narratives of blame and forgiveness in which they figure. A victim expressing blame will typically recount the ways in which she was wronged, assign responsibility to the wrongdoer, and perhaps press a claim for some sort of redress. The wrongdoer, in the act of apologizing, can be expected to recount events in a way that incorporates each of the core features discussed previously: acknowledging a wrong, taking responsibility for it, expressing remorse, and making a commitment to change. This narrative will juxtapose the wrongdoer’s current perspective, from which they acknowledge and appreciate the perspective of the victim, with their past perspective, which they now see as flawed and incomplete. An apologetic narrative opens up distance between the wrongdoer’s past and present self by juxtaposing their past and present perspectives on their action and endorsing the present perspective as their own. Beyond the canonical elements, however, Griswold seems to suggest that apology narratives will also be personalized and humanizing. The apologizer will explain how they came to commit the wrong act – what they were thinking and how they saw the situation at the time. This sort of contextualization, combined with the wrongdoer’s present recognition of their past mistakes, helps the victim to see the wrongdoer in a new and different light – not just, it seems, because the wrongdoer has joined the victim in repudiating the wrong act, but because in telling their story, they make themselves intelligible – even sympathetic – characters, and reveal themselves as not “monstrously bad” (Griswold 103). To do that, they must tell a story that is a story of change, but also, in a sense, a story of continuity. As Griswold puts it: “[The offender] will have to show that she is not just a wrong-doer; that wrong-doing is not all of her, and indeed that she is becoming her better self” (Griswold 351

Andrea Westlund

104). Drawing on Joseph Butler’s seminal sermons on resentment and forgiveness, Griswold treats “recognition of the humanity common to offender and victim” both as a starting point for this process of dialogical narration and as “a necessary step on the way to forgiveness” (Griswold 79). The narrative account of apology, thus construed, offers a thicker characterization of the interaction between offender and victim than does the conversational account. Not only does it commit itself to a somewhat richer, fuller “script” for apologies (couching the key apologetic moves in a broader self-description or explanation), it depicts the dialogical narrative constructed by the parties as implicitly aimed at a form of mutual emotional understanding:11 an apology doesn’t just express solidarity with the perspective of the victim, but also invites the victim to engage sympathetically with the perspective of the (now repentant) wrongdoer. Each party, in the blame–forgiveness dialogue, must enter into the perspective of the other, and when successful, each comes to see the other in a new light and feel differently about them than they did before. Admittedly, some apologies seem less narrative in form than others. At one extreme, the apologizer’s “narrative” might be stripped down to the basic components of apology itself: acknowledging and taking responsibility for a wrong, expressing remorse, and making a commitment to change. Though arguably narrative in some minimal sense, these components might in some cases be expressed in a highly conventional form and presented without much by way of further explanation of the wrongdoer’s self-conception or frame of mind at earlier and later times. This is not entirely mysterious: Bennett (2022) has argued, drawing on work by Robinson, that certain acts can, by convention, come to have expressive power by becoming symbols of a particular emotion (Bennett 11). In some cases, performing an expressive act that effectively symbolizes the emotional distancing of oneself from one’s previous self or actions might be enough to provide a reason to forgive, without placing that act within the sense-making context of a narrative. In some cases, one might even argue, narrative elaboration serves as an unwanted distraction – centering the wrongdoer and her thoughts rather than the victim. Perhaps it asks too much of a victim, at least in some cases, to expect her to enter sympathetically into the perspective of the wrongdoer and hear their story. In such cases, a more stripped down conversational model might better serve our purposes. Even so, the narrative model of apologies has at least one distinctive strength, to which I alluded previously: it can make an important contribution to our understanding of “complex” apologies. Complex apologies, as I am using the term, are apologies that include not just the canonical elements but also some explanation of what the wrongdoer was thinking at the time – a characterization of her own earlier point of view, which might include reference to various respects in which her understanding was flawed or in which her intentions failed to align with her actions. Finer-grained accounts of what the agent thought she was doing or how she conceptualized the situation do not always function as excuses. An account like Griswold’s can help to clarify this point: entering into the offender’s perspective and appreciating its intelligibility can change one’s view of the wrongdoer without undermining a judgment of culpability. Seeing someone’s missteps as intelligible – in the sense of being understandable and even relatable – is not the same as seeing them as excusable. But, it can help to humanize the wrongdoer – to see them as merely fallible rather than as monstrous – and to open one’s heart to the possibility of forgiveness. If an offender can’t or won’t make themselves intelligible, by contrast, there is a sense in which they and their motives will remain confoundingly alien to their victims. Consider the 352

Apology and forgiveness

(often) rhetorical question “Who do you think you are?” with which one might be inclined to angrily confront an offender. This exclamation expresses a certain outraged bafflement at how anyone could see their way to doing what the offender has done. It suggests that the offender must be placing themselves above the victim and disregarding the victim’s moral standing. In some cases, an apology that includes all the canonical elements might still leave the victim feeling perplexed and unsettled. “But how could you have done it?” one wants to know. “What were you thinking?”12 A complex apology owns up to the wrong while at the same time placing it in a broader interpretive context that makes it seem less alien and more understandable, thus helping to build a basis for trust and facilitate the restoration of a default level of goodwill.13 I hope it is already apparent that the conversational and narrative models of apology are not incompatible. Indeed, the narrative model can be seen as complimenting the conversational model, elaborating one more specific form that the dialogue between wrongdoer and victim might take. A narrative that juxtaposes the wrongdoer’s past and present perspectives may serve as a vehicle for the expressive acts through which the wrongdoer joins the victim in repudiating the wrong.14 Apologies that flesh out and juxtapose offenders’ past and present perspectives, thereby articulating their narrative self-understanding, can clarify in a more fine-grained way the “quality of will” expressed by an action and by a subsequent apology. After all, actions aren’t limited to expressing “good” will or “ill” will simpliciter, but may express good or ill will in varying degrees or in different respects. A relatively minor wrongdoing, for example, might not express positively ill will, or thorough-going disregard, but might instead manifest something more like insufficiently good will, or an insufficient degree or attentiveness or regard to the victim’s perspective or experience.15 Narrative apologies offer an interpretive framework for understanding the nature and significance of the wrong, without (necessarily) disowning responsibility for it. Some narratives might simply confirm what the victim already took to be the case and thus consolidate that story between the parties. But in other cases, a narrative apology might play a more transformative role, inviting to the victim to revise their conception of the wrong and of the wrongdoer in some way. Narratively reframed, a wrong may become easier to forgive – either because it now appears to be a lesser or different sort of wrong, or because, even if rather significant, the offender’s mistake no longer seems baffling or monstrous and there is some basis for the restoration of trust. It might take some back and forth to get to a point at which offender and victim share an interpretation of the sequence of events in question. On this picture, an apology is a collaborative affair, in which the parties jointly work out a shared understanding of what has transpired and how to feel about it. Like the conversational model, the narrative model has resources to clearly differentiate itself from the transactional model. One way of construing what it is to be held to account is to be subject to a normative demand for an account that makes sense of what one has done. Like conversational demands, the demands of joint narration have a normative force internal to the practice in which they arise, which does not depend on any entitlement, on the part of the victim, to coercively extract what is owed. Indeed, if the motive to self-narrate is to satisfy a powerful other who threatens one with further sanctions, the self-narration will be an act of submission, rather than of self-accounting, and it will not play the role carved out for it on the narrative model. We might borrow a term from H. L. A. Hart (1961), here, and say that Alexander (in Bergman’s film) is “obliged” 353

Andrea Westlund

to confess by the bishop, who is powerful enough to back his demands with credible threats of harm. But to suggest that the demand for an apology can only ever amount to an attempt to oblige, in this sense, is to suggest that there are no genuinely authoritative normative requirements internal to responsibility practices themselves. This suggestion seems unwarranted.16 Surprisingly, Nussbaum herself treats Griswold’s account of apology and forgiveness as paradigmatically transactional in nature. The requirement that the apologizer engage in selfexamination and offer a narrative account of their wrongdoing and commitment to change strikes Nussbaum as distinctly intrusive – drawing on Foucault, she treats the demand for self-narration as a particularly insidious exercise of power by the listener over the speaker. Certainly, if an apologetic narrative is nothing more than a one-sided confession, the terms of which are entirely set by the confessor and imposed on the confessee, the exercise is likely to feel disempowering to the confessee. Alexander feels disempowered by the bishop for precisely this reason. He knows what he is expected to say, it is clear that the required story would not be his story, and that deviations from the script will not be heard or accepted. But Griswold’s actual account is far more dialogical than this representation makes it seem. The narrative element of an apology is precisely where the apologizer has an opportunity to make herself heard or known – a discursive space in which she can participate in self-interpretation and self-representation, rather than simply taking on the fixed identity of “wrongdoer” (or, worse yet, “monster”). Paradoxically, perhaps, a narrative apology can be empowering to the apologizer as well as to the victim – it is a way of being seen or heard for who she takes herself to be, even as it is an opportunity to show that she sees and hears the victim for who they are. So much of what is wrong with Alexander’s exchange with the bishop is apparent on the surface: the bishop is a powerful adult, and Alexander is a relatively powerless child. Alexander’s transgression was arguably excusable in the first place – or at least of a much lesser magnitude than the bishop makes it out to be. The bishop’s disproportionate response, if anything, speaks to his own insecurity about his standing in the family. The most salient problem, of course, is that the bishop coerces Alexander’s apology, which is a distortion of the sense in which one who has been wronged is entitled to demand such a thing. In cases like Alexander’s, the very prospect of apologising feels like an affront to one’s dignity because one is being denied a role in the construction of a narrative in which one figures. Alexander is cast as a character in a set piece, and his choices within that piece are highly limited. Under more favorable conditions, an apology is an opportunity to tell one’s own story, not just an act of submission to another’s narrative framing of one’s action. There are, of course, bad and self-serving ways of telling one’s story. Such efforts at apology typically fail. But it is often important to our practices of moral reconciliation – and, ultimately, to practices of forgiveness – that a story be told. Without it, the moral meaning of an offender’s action will often not be clear, nor will it be clear when or why trust may be restored. Paradoxically, perhaps, successful apologies may be morally empowering to the wrongdoer as well as to the victim. While apology rituals that simply extract concessions are exercises in humiliation rather than in moral reconciliation, the narrative model gives us a way of construing apology as genuinely interactive and creative. It explains how apology can be more than an exercise in self-effacement, and serve as an invitation to see the wrongdoer in a different light – which is, after all, a prerequisite to forgiveness on most views.

354

Apology and forgiveness

Notes 1 I am grateful to Maximilian Kiener for his excellent editorial support and feedback, and to Christopher Bennett for detailed comments on the penultimate draft of this chapter. At an earlier stage I also benefited immensely from discussion at an online workshop for contributors to this volume. Thanks to all those who participated for their helpful comments and suggestions. 2 See Adrienne Martin (2010: 534) for a useful gloss on apology’s reason-giving power. 3 Henceforth, when I refer to the transactional account, I will mean the transactional account as understood by Nussbaum. As one reviewer has pointed out, there may be other, and more charitable, ways of fleshing out the transactional model. I suspect alternative interpretations will still lack important elements captured by the conversational and narrative accounts, but I will not take on this more general argument here. 4 Different accounts offer slight variations on these elements, and also may differ over which elements are required to count as an apology at all and which are merely typical. See, for example, Alice MacLachlan (2015) and Christopher Bennett (2022). My tripartite list is meant simply as a rough summary of features that are commonly included as either required or typical. 5 Thanks to Leo Menges for suggesting this interpretation in a workshop discussion. 6 Interestingly, Murphy categorizes apologies as part of a ritual of humiliation that compensates for a status-injury by “equalizing” the parties, thus appearing to embrace a view of apology more like Nussbaum’s transactional account. As I go on to argue shortly, however, it is not obvious how rituals of humiliation are meant to effect the right kind of distance between the wrongdoer and the act, and so I remain unconvinced that having been humiliated could serve as the right kind of reason to forgive. Murphy does note, however, that in the “best cases” (Murphy 28) an apology will serve as a vehicle for repentance, and thus would admit of the more straightforward sort of explanation. 7 See Bennett (2022) for an interesting and related discussion of what he calls the “performing deference” view of apology, which he attributes to Jeffrie Murphy and which bears a considerable resemblance to Nussbaum’s transactional view. 8 Nussbaum herself argues in favor of unconditional love as an alternative approach to wrongdoers, which accords a lesser (if any) place to the practice of apologizing. 9 Bennett (2022) offers an account of apology as an expressive action, which explains how certain canonical postures, words, and gestures could come to carry the right sort of significance and perform the normative work of apologies. I see his view as complementary to the conversational model, and indeed, as unpacking in greater detail how that model might work. 10 These two sentences loosely but intentionally echo a line from the movie The Philadelphia Story, when C. K. Dexter Haven exclaims to Tracy Lord “That’s no good. That’s not even conversation!” (Cukor 1940). Tracy’s lapse is of a different sort from the bishop’s but, to my ear, the line fits. 11 The idea that narrative understanding is a form of emotional understanding, and that it brings about a form of emotional closure, is a common thread in various influential views of narrative in the philosophical literature. See, for example, Peter Goldie (2012), Jenefer Robinson (2005), and J. David Velleman (2003). 12 John Doris, too, notes that apologies that include explanations (or “rationalizations”) will often be more satisfying than ones that do not (Doris 2015: 149). Rationalizations can help to show that “the problematic behavior is compatible with “good will, affection, or esteem” (Doris 2015: 150), and thus play a positive role in the restoration of relationships. 13 Alice MacLachlan (2015) examines the role of apology in restoring trust in detail. Although narrativity is not a central theme of the paper, MacLachlan does suggest that, in the case of public apologies, a confirmed, “explanatory narrative” that explains what went wrong and how can be reassuring and even empowering to its addressees, since it takes away the spectre of an unknown, amorphous “monster under the bed” (MacLachlan 450) and replaces it with a more mundane, comprehensible set of failures and mistakes. While the language of monstrosity may be entirely coincidental, it nicely echoes the language used by Butler and Griswold to characterize our perception of wrongdoers whom we have not (yet) come to understand. 14 Thanks to Elinor Mason for pointing out, in discussion, that we can usefully distinguish between the work done by an expressive act and the mechanism by which it is carried out.

355

Andrea Westlund 5 Thanks to Leo Menges for making this suggestion in discussion. 1 16 I refer, here, to a distinction famously drawn by H. L. A. Hart in The Concept of Law (Hart 1961), between being (psychologically) obliged and being (normatively) obligated.

Further reading Bennett, Christopher. (2022). “What Goes on When We Apologize?” Journal of Ethics and Social Philosophy, 23(1), 1–21 offers an account of apology as an expressive action, which explains how certain canonical postures, words, and gestures could come to carry the right sort of significance and perform the normative work of apologies. (1970). Butler’s Fifteen Sermons Preached at the Rolls Chapel, ed. T. A. Roberts. London: S. P. C. K, includes influential sermons on resentment and forgiveness by Joseph Butler, which serve as a starting point for much contemporary work on anger, forgiveness, and apology. Griswold, Charles. (2007). Forgiveness: A Philosophical Exploration.  New York: Cambridge University Press, offers a narrative account of rituals of apology and forgiveness. Hieronymi, Pamela. (2001). “Articulating an Uncompromising Forgiveness,” Philosophy and Phenomenological Research, 62(3), 529–555 develops an influential account of “uncompromising” forgiveness, in which a wrongdoer’s sincere repentance or apology can serve as a reason to forgive insofar as it withdraws their endorsement from the threatening claim made by their past act. McKenna, Michael. (2012). Conversation and Responsibility. New York: Oxford University Press defends a conversational account of responsibility and related practices such as blaming, apologizing, and forgiving. Nussbaum, Martha. (2016). Anger and Forgiveness: Resentment, Generosity, Justice.  New York: Oxford University Press gives a critique of apology and forgiveness understood as a transactional ritual.

References Bennett, Christopher. (2022). “What Goes on When We Apologize?” Journal of Ethics and Social Philosophy, 23(1), 1–21. Bergman, Ingmar (dir.). (1982). Fanny and Alexander. Cinematograph. 188 minutes. Butler, Joseph. (1970). Butler’s Fifteen Sermons Preached at the Rolls Chapel, ed. T. A. Roberts. London: S. P. C. K. Cukor, George. (1940). The Philadelphia Story. Metro-Goldwin-Meyer. 112 minutes. Doris, John. (2015). Talking to Our Selves: Reflection, Ignorance, Agency. New York: Oxford University Press. Goldie, Peter. (2012). The Mess Inside. Oxford: Oxford University Press. Griswold, Charles. (2007). Forgiveness: A Philosophical Exploration. New York: Cambridge University Press. Hart, H. L. A. (1961). The Concept of Law. Oxford: Oxford University Press. Hieronymi, Pamela. (2001). “Articulating and Uncompromising Forgiveness,” Philosophy and Phenomenological Research, 62(3), 529–555. MacLachlan, Alice. (2015). “ ‘Trust Me, I’m Sorry’: The Paradox of Public Apology,” The Monist, 98, 441–456. Martin, Adrienne. (2010). “Owning Up and Lowering Down: The Power of Forgiveness,” The Journal of Philosophy, 107(10), 534–553. McKenna, Michael. (2012). Conversation and Responsibility. New York: Oxford University Press. Murphy, Jeffrie G. (1988). “Forgiveness and Resentment,” in Jeffrie G. Murphy & Jean Hampton (eds.), Forgiveness and Mercy. Cambridge: Cambridge University Press. Nussbaum, Martha. (2016). Anger and Forgiveness: Resentment, Generosity, Justice. New York: Oxford University Press. Robinson, Jenefer. (2005). Deeper Than Reason: Emotion and its Role in Literature, Music, and Art. Oxford: Oxford University Press. Velleman, J. David. (2003). “Narrative Explanation,” Philosophical Review, 112, 1–25.

356

28 TAKING RESPONSIBILITY Elinor Mason

‘Taking responsibility’ has several everyday senses. One thing that could be meant is that someone is voluntarily taking on an obligation, usually for a practical matter. For example in discussing plans for a picnic, someone might say, “I’ll take responsibility for the food”. In that case, they mean that they will accept an obligation to organize the food, and will accept being held responsible (blamed and held liable) if they do not organize the food. This is fairly straightforward, and the reference to responsibility here is not the sense that agency theorists are interested in, but rather a matter primarily of interest in normative ethics: of how obligations work, and how they can be acquired. Another usage of the phrase, ‘taking responsibility’ applies to something that has already happened rather than something that the agent is undertaking. Someone might say, ‘I take responsibility for the damage to your car’. In that case, they could be saying that they are responsible (and always were), and now they are owning up to that fact. But the phrase does not always imply that the person accepts that the damage was their fault, or that they feel they are responsible in a basic sense. People often talk about ‘taking responsibility’ when they accept liability, that is, when they accept that it is their duty to repair or recompense for a harm to another person. Liability does not always require basic responsibility: there are other links to the agent that will justify liability. It may be that the agent’s children or pets damaged the car, and so she is not responsible in a basic sense. Recently, the notion of ‘taking responsibility’ has been deployed in agency theory in a rich sense, so that by taking responsibility one becomes responsible. And the sense of ‘responsibility’ here is the ambitious sense, according to which responsibility implies something along the lines of that one is correctly credited with the act, that it stems from one’s own agency, that it is the agent’s own act. Let’s call that ‘basic responsibility’ to distinguish it from other uses of the word ‘responsibility’. On the face of it, the idea that one could become responsible by taking responsibility is puzzling: how could someone retroactively change the facts about who is responsible for what? And it is puzzling whatever sort of account of basic responsibility we are drawn to. If we think that basic responsibility is tied to free will then it is obviously odd to claim that we could retroactively make something a free act. But even if we accept a compatibilist account, and think that basic responsibility depends on an act being chosen by the agent 357

DOI: 10.4324/9781003282242-39

Elinor Mason

without inappropriate causal influences, basic responsibility seems to be based on something that happens before and during the act, not what happens after that act has been completed. So it is apparently equally odd for a compatibilist to claim that an agent can make themselves responsible by taking responsibility. In this chapter I examine various accounts of taking responsibility and defend the apparently outlandish possibility of retroactively changing the responsibility facts. Let us start with Susan Wolf’s argument. Wolf begins with Bernard William’s account of moral luck – the phenomenon whereby our moral record is affected in some way by outcomes that are not under our control.1 Wolf compares a case where someone has acted in a faulty way to a small degree, and their action has a very bad outcome, to a case where an agent acts badly to the same degree, but gets away with no bad outcome. Adapting one of Williams’ examples, she imagines a lorry driver who has failed to have his brakes checked. A child unexpectedly runs out in front of the vehicle and is killed. In such a case the driver will presumably feel awful about what has happened, and will blame themself. By contrast, a driver who is negligent to the same degree in leaving the brakes in need of maintenance, but who (though mere luck) does not harm anyone, will not feel awful. Furthermore, these reactions seem fitting – we think it is natural and right for the driver who killed someone to feel awful and to blame themselves, and it makes sense that the driver who does not, would not feel awful. The puzzle is that the motivations and voluntary behavior of the two drivers is the same. So, they should blame themselves to the same degree. It seems to be a fundamental principle of any account of responsibility that we are responsible (and therefore praise- or blameworthy) only for what we do, where doing something is acting voluntarily, or at least, through some sort of motivated action. Outcomes that are a matter of luck are not doings. If we want to stick to that principle we should insist that luck can make no difference to the lorry drivers. However, as both Williams and Wolf point out, that view does not reflect our actual moral responses. We think that the driver who kills a person should feel something extra, that there is something severely lacking if they do not. And therein lies the puzzle of moral luck: on the one hand it seems as though luck can make no difference to our moral ledger, and on the other hand it seems that it does. There are various way that we might try to solve the puzzle without claiming that the agent is responsible and properly blameworthy for outcomes that are a matter of luck. One of William’s own ideas, the idea of ‘agent regret’ can be deployed to avoid the conclusion that responsibility is infused with luck: there are emotional reactions other than remorse that acknowledge a harm done. Agent regret is a rational reaction to being the causally responsible factor behind a harm, and we should not confuse that with remorse, which involves a sense of moral responsibility. Generally, we can point to a range fitting reactions and responses to inadvertent harm that do not commit the agent to thinking that they are responsible. There are various emotional reactions, and also more practical reactions, such as apologizing (some forms of which need not involve admitting guilt in the moral sense), or accepting some form of liability.2 Wolf’s account of things could be read as a version of this strategy, but her story is nuanced, and some of what she says points to something that is close to the idea that the agent in this sort of case can take responsibility and thereby become responsible. She observes that there is something seriously lacking in the driver who kills someone and then manages to detach the fault from the outcome in their own mind, telling themself, “My fault here was pretty minor, I should not feel bad”. Wolf considers the idea that the driver is 358

Taking responsibility

trying to minimize their responsibility, and she hones in on what is disturbing about that: it is the driver believes the child’s death has nothing to do with them. As Wolf puts it, The problem is not that he refuses to accept what responsibility he objectively has for the child’s death; it is that he fails to take responsibility for it, in a way that goes beyond that. He reveals a sense of himself – his real self, one might say – as one who is, at least in principle, distinct from his effects on the world, whose real quality and value, for better and for worse, is at best impurely indicated but not at all constituted by the goods and the harms, the successes and the failures that comprise his life in the physical world. It is as if he draws a circle around himself, coincident with the sphere of his will. (Wolf 2001: 12–13) Wolf goes on to describe a “nameless virtue”, which would involve a tendency to feel something more than agent regret (but maybe not quite guilt), and a corresponding willingness to accept accountability for what one does, where take an expansive view of what is done: that is to say, we allow that not everything we do in the relevant sense is fully voluntary. I’ll come back to the sort of thing that may count as a doing in the expansive sense. First, let’s focus on what it is to accept accountability, for that is what is crucial to how radical Wolf’s proposal is. If we read ‘accountability’ as liability, Wolf’s proposal is not radical at all. Of course it is a virtue to accept liability for some of the unlucky bad outcomes of our being in the world. As Wolf says, if I break a vase, I should offer to pay. At some points Wolf’s remarks point to an interpretation in terms of liability: a virtuous agent accepts more liability than her level of fault demand. Indeed, it is not part of our concept of liability that it is precisely connected to fault, as the category of strict liability illustrates: we can be liable for what others do, what our pets do, and there are other areas of strict liability. However, at times, Wolf seems to be saying something more ambitious. We expect the vase-breaker, like the truck driver who hit the child, to acknowledge that the consequences of his behavior have something specifically to do with him. We expect the vase-breaker to offer to pay, then, not only because we want him to be generous, but because we expect him to accept contingency in the determination and assessment of who he is. (Wolf 2001: 14) Wolf’s radical claim here is that the outcomes of our actions may justifiably affects assessments of who we are, even when those outcomes are partly down to luck. Our agency – who we are as agents – is not just down to what we do voluntarily, but down to how things turn out. We can rephrase this in terms of responsibility – if we are responsible for actions that spring from our agency, and agency is affected by outcome luck, we can be responsible for outcomes that are a matter of luck. Let’s pause and clarify the proposal that is coming into view. The radical idea is that we can become responsible by taking responsibility. But given the everyday ambiguities of the phrase ‘taking responsibility’, we need to be very clear about what is meant by the claim that one can take responsibility. If all we mean is that one can accept responsibility for something that one is actually responsible for, we have not said anything new.3 The idea must be that the person who takes responsibility was not previously responsible. 359

Elinor Mason

Furthermore, we must be clear that the talk of taking responsibility does not refer to taking on liability, or duties of reparation, but refers to becoming responsible, in a sense that agency and responsibility theorists are interested in. Wolf’s view is a little slippery on these issues. It is not completely clear what sort of responsibility someone would end up if they exercise the nameless virtue. Wolf thinks that when we inadvertently cause harm, it is appropriate for us to feel (something like) remorse, but not for others to blame us. Thus the sort of responsibility that is taken is not the same as the responsibility we have when we act knowingly or recklessly. Wolf thinks that it would be appropriate to take on duties of reparation. But this is compatible with an account of liability as opposed to responsibility. Furthermore, the role of the ‘taking’ is not completely clear. On one interpretation of Wolf’s account, there is contingency in who we are. And that is why we should expect to see that contingency reflected in our own (and others’) assessments of who we are. On that interpretation, Wolf is offering us an account of responsibility for luck, rather an account of taking responsibility. At the same time, other remarks that Wolf makes suggest an interesting account of taking responsibility. Most importantly, Wolf gives an account of the reasons for taking responsibility that do not reduce to a claim that the agent is responsible and so should take responsibility. She says that we must take responsibility for contingencies of fortune in order to engage properly with others: To form one’s attitudes and judgments of oneself and others solely on the basis of their wills and intentions, to draw sharp lines between what one is responsible for and what is up to the rest of the world, to try in this way, to extricate oneself and others from the messiness, and the irrational contingencies of the world, would be to remove oneself from the only ground on which it is possible for beings like ourselves to meet. If we define ourselves in ways that aim to minimize the significance of contingency and luck, we do so at the cost of living less fully in the world, or at least at the cost of engaging less fully with the others who share that world. (Wolf 2001: 14–15) The point of this discussion of Wolf is to illuminate what more work needs to be done to render an account of taking responsibility clear and convincing. We need to show that taking responsibility really changes the situation, that it is not just an acknowledgement of responsibility that was already there. Crucially, it is essential to explain the nature of the responsibility that is taken. If it is not mere liability, what is it? Wolf is not completely clear on that. And (not unrelatedly, of course) we need an account of which things one can take responsibility for. For Wolf, the answer is that we can take responsibility for things that we do, given an expansive sense of our doings. Additionally, we need an account of why someone would take responsibility. Wolf offers an interesting answer here: it is connected to our interpersonal involvements. Let’s spell these desiderata out a bit more neatly. 1. Does the role of ‘taking responsibility’ genuinely shift us from a situation where there was no responsibility to one where there is responsibility? 2. What sort of responsibility is taken (it should not be responsibility for some future state of affairs and it should not be liability)? 3. What range of things can we take responsibility for? 4. Why would someone take responsibility? 360

Taking responsibility

Finally, in answering these questions an account of taking responsibility should have an answer to the skeptical question that I started with: how on earth could someone retroactively change the responsibility facts? David Enoch (2012) gives some different answers to Wolf. Like Wolf, he refers to Williams’ example of the lorry driver and, like Wolf, he thinks that there is something morally deficient in the driver who shrugs off the bad luck of having killed someone. So Enoch has a clear answer to the last question, why we should take responsibility. His view is that, taking responsibility for things we are not previously responsible for is sometimes a moral duty. He also has an answer to question 3: he argues that we that we can take responsibility for anything in our ‘penumbral agency’. For Enoch, this is a matter of being related to the act in some relevant way, and is much wider than than any standard conception of what falls within our agency. Indeed, penumbral agency may include things that other people do (our children, for example). This gives us a clear answer to question 1 – if we can take responsibility for things that are not even indirectly caused by our own agency, then obviously the role of responsibility taking is robust and moves us from a situation where there was no responsibility to one where there is. Enoch’s problem is that the clear answers to those questions have a high cost: they leave him hard pressed to justify his claim that what is taken on really is responsibility rather than liability. On Enoch’s picture, taking responsibility is a normative power. We can do it at will, when we see that we have a moral duty to do so. But this picture does not look like a picture of responsibility. Rather, the obvious thing we have a moral duty to take on, and the obvious thing we take on by an act of will, and can apply even to the actions of others, is liability, not responsibility.4 Enoch admits that his account does not get at responsibility in the sense that it is related to ‘core agency’, as he calls it, the sense in which one’s acts are indelibly marked on one’s moral ledger. (2012: 115.). However, he argues that one can take responsibility in “an answerability or accountability sense”.5 According to Enoch, this is a matter of it being appropriate to offer apologies, excuses, or justifications. But that doesn’t get us very far. We need to know what the conditions of aptness are for apologies, and excuses or justifications. If we see those as part of liability, and not requiring backward looking responsibility (as we well might), then the sense in which responsibility has been taken is just liability. But if we think that apologies, excuses, and justifications are only apt when one really is responsible in the backward looking sense, then Enoch has not answered the question, but rather, pushed it up a level. Does answerability imply responsibility in the basic sense or not? If answerability is a sort of basic responsibility then the idea that one becomes answerable when one takes responsibility is radical and ambitious, but obviously needs support.6 The real problem here may be that Enoch’s view is too ambitious in the range of things that he thinks we can take responsibility for. In allowing that we can take responsibility for things that we were nowhere near responsible for in the first place – things that are not at all our own actions – he is making the task too hard, and inevitably all he will be able to show is that we can take on liability for those things. On Wolf’s view, recall, the things we can take responsibility for are our doings (expansively conceived), and it is not actually clear that on her view we were not responsible for those things in the first place. There is really only a very small role for taking responsibility. But perhaps that small role is what we need here. Perhaps we need to accept that on the best account of taking responsibility, the answers to question 1 (concerning the distinctness of the states before and after responsibility is taken) and question 4 (what range of things we can take responsibility for) may not 361

Elinor Mason

be very exciting. It may be that the potential for shifting the responsibility facts by taking responsibility applies only to cases where we were already close to being responsible. That would make sense, after all. We should be looking for a narrower account than Enoch offers. We find such an account in Joseph Raz. Raz is concerned with the category of ‘negligent’ acts, where negligence is not recklessness but rather the state of acting without awareness of the relevant properties. So, someone who forgets about their dog in the car on a hot day is not necessarily reckless – they may have taken all precautions it was rational to take – but still, sometimes, people forget important things. We cannot trace it back to a bad act or bad motivation, it is just a sort of glitch in their mental functioning. This is negligence in the sense Raz is concerned with. The category of negligent acts does not extend very far from the set of acts that are uncontroversially in our responsibility zone: certainly not as far as acts done by others or by entities that have some connection to us. As I say, this may be a strength of Raz’s account. Raz, like Wolf, argues that we can take responsibility for these acts. Like Wolf’s, Raz’s argument does not focus too much on whether we are already responsible for these acts or whether we move to a completely new situation. Raz is more concerned with explaining how it is that our sense of our own agency can push us to seeing ourselves as responsible for these acts. Whether that makes us responsible or is a matter of seeing that we are responsible is not, on Raz’s picture, the main issue. The main issue is why we should see ourselves as responsible. Raz’s account leans heavily on the idea that we have a “domain of secure competence”. Contrast things that happen that we are clearly not responsible for – the doings of others, the reflexes of our body, and so on. These things are not connected to our agency. There are other things that are very directly and obviously connected to our agency – our deliberate conscious doings. These are within our domain of secure competence. Sometimes, things within our domain of secure competence don’t work out, and not through something we control, rather, through unexpected failure of control. Take negligent actions. These actions are not controlled, but they are, as Raz puts it, within our domain of secure competence. They are acts that we normally would be able to control. When I forget something, like the dog in the car, I  do something that I  could normally be expected not to do. This is where Raz sees the important difference between these and other happenings that we do not control. Elinor Mason calls these cases of ‘ambiguous agency’, (Mason 2019a, 2019b) indicating that these are acts are not ones that are clearly agential, but they are not clearly non-agential either. They are ambiguous, not just because the observer cannot see whether the act is agential (in real life we are often unable to tell whether an act was deliberate or inadvertent), but in themselves. Ambiguous agency is a genuine grey area. Raz begins with an argument that makes it look as though he is arguing that we are already responsible for our negligent acts. Raz says that we are responsible in cases where our powers of rational agency malfunction – we have all the opportunities and capacities, and yet, we fail to notice something important, or fail to follow through on an intended action.7 Raz points out that his account explains how our capacities can render us responsible for our conduct, even when our conduct was not intentional (on Raz’s view, if we have a capacity and the opportunity to exercise it, we are responsible for both our failures and our successes). However, Raz recognizes that more is needed: why would we accept that this is the right way to think of the relationship between capacities and conduct? In particular, when conduct is inadvertent, why should we think that we are responsible? Raz says: “The 362

Taking responsibility

account has to, and can, be bolstered by an explanation of the significance of the actions for which we are responsible in our life, a significance which depends on responsibility having this sense [the sense including negligent actions]” Raz (2010: 16–17). This is where taking responsibility comes in. Raz argues that our sense of who we are depends on us affirming “a domain of secure competence” (2010: 17). We have a sense of ourselves as agents, and we need that to function in the world: The way we feel about ourselves, our self-esteem, our self-respect, the degree to which we are content to be what we are, or what we perceive ourselves to be, our pride in ourselves, our shame in how we are, or in how we conduct ourselves – all these and various other self-directed attitudes and emotions depend in part on competence in using our faculties of rational agency. (2010: 17)8 So, we are heavily invested in this domain of secure competence, and when things go wrong, we must react by buttressing our agency. We should, Raz thinks, hold ourselves responsible for malfunctions within our domain of secure competence. For Raz, taking on responsibility for negligent acts is a way of affirming our own agency.9 Raz is thus doing two things at once in this argument. He is telling us why we should take responsibility (because our sense of self is threatened if we do not), and in doing that, he is also telling us what sort of thing our concept of responsibility is. It is not a metaphysical thing, whose truth conditions are independent of us. Rather, it is a concept that we construct according to our needs. This is where Raz and Wolf converge. They both argue that an expansive view of agency and responsibility is better for our human-scale purposes. What is essential in both Wolf and Raz’s views is that the actions we can take responsibility for are ones that we might anyway decide fall into the category of responsible acts. Taking responsibility is not alchemy, we cannot produce gold out of straw. Rather, when an act is in the grey area of agency – the sort of act that was not produced by deliberate and successful rational processes, and yet was produced by the agent to the extent that it strikes as one that the agent should feel something responsibility related about (maybe agent regret, maybe more) – perhaps we can take responsibility. The act already straddles the category of the agential and the non-agential. So, according to this line of thought, we can choose to classify the act in the agential – we can decide to own it. A plausible story about taking responsibility is emerging. Different accounts may favor different reasons for taking responsibility, but the basic idea, that it sometimes makes sense to take an expansive view of our own agency, is promising. Of course, the sceptic will not be satisfied. They will insist that that if that act does not spring from our agency in the right way then nothing will make it agential. Imagine someone who thinks that ‘springing from agency in the right way’ means ‘being done with metaphysical free will’. For a free will libertarian like this, there is no way we could make ourselves responsible by taking responsibility. But for a compatibilist, there may be more wiggle room. For a compatibilist, responsibility is already a somewhat awkward compromise between the compelling idea that an agent must act on her own without external determination, and the undeniable fact that agents never ‘act on their own’. So the compatibilist has to decide which external influences matter, and which do not. As Harry Frankfurt (1971) puts it, the concept of a person is crucial here. Which influences disrupt personhood, and which constitute it? 363

Elinor Mason

Different compatibilists give different accounts of how to delineate the aspects of our being that we are responsible for from those that we are not. For some, our emotions and characters are in the realm of the responsible. For others, only what we do absolutely deliberately is relevant to responsibility. So why not include taking responsibility as a proposal for a route to responsibility? Again, the objection here will be that we cannot retrospectively change responsibility, and that all the parties to disagreements within compatibilism about which things are in the responsibility realm will agree on this: responsibility is backward looking! I will address this head on. But first, remember that the category of acts at issue here is the category of acts that are already plausibly agential in at least one respect: they look like cases of moral luck, in that we do think that an agent ought to feel something like responsibility for the act. We are engaging in a sort of reflective equilibrium here. Our intuitive reaction to these cases is that there is some sort of responsibility. Our neatest theories tell us there is not. So what should we say? Well, before we jettison the intuitive reaction, we should look at where it comes from. It is plausible to think that our tendency to think of ourselves and others as responsible in moral luck type cases reflects something we need from our responsibility system. And of course, this is just what Wolf and Raz argue: we need to see ourselves and others as solid. Now the question is, what weight does this kind of consideration  – our human and social needs – have, when the consideration conflicts with what was looking like a nice neat theory? To answer that, we have to think about what sort of truth we were hoping for from our theory. If I need to think that my friend has not betrayed me but all the evidence points to the fact that they have, I should go with the evidence. But that is not the structure of our problem here. Compatibilists are building a theory of responsibility. Indeed, the reason for building one in the first place is very much related to our needs. We can’t get along without a responsibility practice, as Strawson eloquently argues. So why not take our need to include responsibility for some moral luck cases seriously? And we can do that by allowing the phenomenon of taking responsibility. This simple argument will leave the hard-headed theorist unsatisfied. Their line is that responsibility cannot be taken after the fact, because responsibility is determined by the history of the act, and that that is essential to the concept of responsibility. The proponent of taking responsibility points out that the concept of responsibility is there to serve our needs, and that sometimes the best way to do that is to allow that responsibility can be taken. The objector will respond that this is the wrong place to allow needs into the picture. Compare a possible justification for the practice of promise keeping: we have such a practice because it serves our needs. But on a particular occasion we must stick to the rules.10 This is a fair point, but there is a response. The proposal is not that people could take responsibility whenever it suits them. For example, a mother would not be able to protect her child be taking responsibility for their crimes. The proposal is rather that the rules of the practice could include conditions under which taking responsibility was possible. There is still something that rankles the objector. It seems that whatever you can take on must look more like liability than responsibility. Basic responsibility is not the sort of thing that we can conjure. One way to give shape to this thought is in terms of remorse. The objector may say: ‘You can regret an inadvertent act, but you can’t take on responsibility, because you are not responsible! You can make reparations, of course, but you yourself admit that is not responsibility! You can’t feel remorse for something you did

364

Taking responsibility

inadvertently!’ The idea of remorse is interesting. It does seem that remorse, unlike agent regret, implies responsibility. Could one feel remorse for inadvertent bad consequences? Let’s grant that it can be rational to feel something rather complex in this sort of case. Agent regret, for example, can be explained as rational – it is a reaction to something that has come about through my actions, even if not something I am responsible for. My causal involvement licenses a particular sort of emotional reaction. That does not give us responsibility. But it is the thin end of the wedge. If there is an apt emotion that is related to my causal involvement, there may be an apt emotion that is related to my ambiguously agential involvement, and we can creep up the scale, so that what we have is strikingly close to the sort of emotion that is appropriate when I deliberately and self-consciously do something bad.11 And yet again, the sceptic will insist that whatever emotion is apt is not a responsibility emotion, but a ‘bystander emotion’, like regret. At this point, the nature of the deadlock may seem intractable. Proponents of taking responsibility do not have a knock down argument for the hard headed theorist. But they have a lot going for them: coherence with our practices, and relevant and persuasive considerations about what sort of theory our theory of responsibility is.

Conclusion In this chapter I  have tried to suggest that there are very promising accounts of taking responsibility, that give us a genuine sense in which taking responsibility changes the responsibility facts. A defender of this sort of view must show that we have good reasons to take responsibility in some cases. This is the easy part: both Wolf and Raz give plausible accounts, related to our relationships with others and our self-respect respectively. The hard part is showing that our concept of responsibility can be bent so as to accept a new sort of route to responsibility, so that one can become responsible for things that one was not responsible for solely in virtue of the casual history of the act. I have suggested that that may not be as big of a stretch as it is initially sounds.

Notes 1 Williams (1981). See also Adams (1985). Adams focusses on taking responsibility for our emotional states and tendencies. 2 Jordan MacKenzie (2017) argues that agent regret can be rational in the context of a social practice of taking on reparative duties in cases of moral luck. This may be true but agent regret does not correlate with basic responsibility, so MacKenzie does not show that we can take responsibility in the basic sense. 3 Fischer and Ravizza (1998) stress the importance of accepting responsibility in this sense. 4 For other broadly obligation centered accounts of taking responsibility see: Trystan Goetze (2021), who explicitly disavows the idea that we might take responsibility in the basic sense and focusses on providing an account of how vicarious responsibility works, and Garrath Williams (2020), who emphasizes that we have a collective duty to take collective responsibility. 5 Others who appeal to answerability in this context include Maximilian Kiener (2022), who argues that we have an interest in taking on ‘answerability’ where there is an answerability gap. Like Enoch, Kiener thinks that taking responsibility is a normative power. 6 Though there are various accounts of what they are. See, e.g., David Shoemaker 2011; Angela Smith 2012. 7 This is also how George Sher deals with this sort of case (Sher 2009: 183–184). More recently, Gary Watson has made a similar suggestion (Watson 2021).

365

Elinor Mason 8 The argument appears in his 2011 book, From Normativity to Responsibility, in slightly different words but essentially unchanged. 9 Elinor Mason (2019a, 2019b) takes a more Wolfian line, arguing that it is not so much for our own self-respect that we should take responsibility, but in order to show respect for and to maintain the respect of others. (Mason differs from Wolf in arguing that others can justifiably blame someone who has taken responsibility). Mason is also concerned with the gendered dimension of taking responsibility (Mason 2018). 10 This is Rawls’ point in ‘Two Concepts of Rules’ (1955). 11 For a more detailed version of this argument see Mason 2019b.

Further reading Enoch, David. (2012). “Being Responsible, Taking Responsibility, and Penumbral Agency,” in U. Heuer & G. Lang (ed.), Luck, Value, and Commitment: Themes from the Ethics of Bernard Williams. Oxford: Oxford University Press is where Enoch lays out his account of the virtue of taking responsibility. Raz, Joseph. (2011). From Normativity to Responsibility. Oxford: Oxford University Press is a collection of Raz’s essays brought together as a big picture of Raz’s views on normative ethics, responsibility, and the relationship between them. Wolf, Susan. (2001). “The Moral of Moral Luck,” Philosophic Exchange, 31(1), 1–19 is her foundational contribution to the theory of taking responsibility.

References Adams, Robert. (1985). “Involuntary Sins,” Philosophical Reviews, 94, 3–31. Enoch, David. (2012). “Being Responsible, Taking Responsibility, and Penumbral Agency,” in U. Heuer & G. Lang (ed.), Luck, Value, and Commitment: Themes from the Ethics of Bernard Williams. Oxford: Oxford University Press. Fischer, John Martin,  & Ravizza, Mark. (1998). Responsibility and Control: A  Theory of Moral Responsibility. Cambridge: Cambridge University Press. Frankfurt, Harry. (1971). “Freedom of the Will and the Concept of a Person,” Journal of Philosophy, 68(1), 5–20. Goetze, Trystan S. (2021). “Moral Entanglement: Taking Responsibility and Vicarious Responsibility,” The Monist, 104(2), 210–223. Kiener, Maximilian. (2022). “Can We Bridge AI’s Responsibility Gap at Will?” Ethical Theory and Moral Practice, 25(4), 575–593. MacKenzie, Jordan. (2017). “Agent-Regret and the Social Practice of Moral Luck,” Res Philosophica, 94(1), 95–117. Mason, Elinor. (2018). “Respecting Each Other and Taking Responsibility for Our Biases,” in Marina Oshana, Katrina Hutchison, & Catriona Mackenzie (eds.), Social Dimensions of Moral Responsibility. Oxford: Oxford University Press. ———. (2019a). “Between Strict Liability and Blameworthy Quality of Will: Taking Responsibility,” in David Shoemaker (ed.), Oxford Studies in Agency and Responsibility, vol. 6 (pp.  241–264). Oxford: Oxford University Press. ———. (2019b). Ways to Be Blameworthy: Rightness, Wrongness, and Responsibility. Oxford: Oxford University Press. Rawls, John. (1955). “Two Concept of Rules,” Philosophical Review, 64(1), 3–32. Raz, Joseph. (2010). “Responsibility and the Negligence Standard,” Oxford Journal of Legal Studies, 30(1), 1–18. ———. (2011). From Normativity to Responsibility. Oxford: Oxford University Press. Sher, George. (2009). Who Knew?: Responsiblity Without Awareness. Oxford: Oxford University Press USA. Shoemaker, David. (2011). “Attributability, Answerability, and Accountability: Toward a Wider Theory of Moral Responsibility,” Ethics, 121(3), 602–632. Smith, A. M. (2012). “Attributability, Answerability, and Accountability: In Defense of a Unified Account,” Ethics, 122(3), 575–589.

366

Taking responsibility Watson, Gary. (2021). “The Possibility of Pure Negligence,” in Pavlakos, George  & RodriguezBlanco, Veronica (eds.), Agency, Negligence and Responsibility. New York: Cambridge University Press. Williams, Bernard. (1981). “Moral Luck,” in Moral Luck: Philosophical Papers 1973–1980. New York: Cambridge University Press. Williams, Garrath. (2020). “Taking Responsibility for Negligence and Non-negligence,” Criminal Law and Philosophy, 14(1), 113–134. Wolf, Susan. (2001). “The Moral of Moral Luck,” Philosophic Exchange, 31(1), 1–19.

367

29 RESPONSIBILITY WITHOUT BLAME Bruce Waller

“Responsibility” has traditionally been used in several distinct senses (Zimmerman 2015: 45), and recently philosophers have added many more. The type of responsibility that has been the subject of the most attention and the greatest controversy is commonly called “moral” responsibility: the responsibility that justifies praise and blame, reward and punishment. This chapter will not attempt to survey the full range of responsibility accounts, nor even cover the currently most popular. It examines a small sample, focusing on those that raise intriguing questions about the deeply entrenched system of moral responsibility. Galen Strawson, Derk Pereboom, and Gregg Caruso have elucidated the traditional notion of moral responsibility and established a good starting point for examining the variety of views that claim the title of “responsibility.” Pereboom (2014) and Caruso (2018, 2021) call this traditional moral responsibility account the “basic desert” model. Strawson refers to it as “ultimate” or “true” moral responsibility (1994: 7, 9) and notes that it is “central to the Western tradition” (1994: 8). His dramatic account of divine punishment in the afterlife (1994: 9) focuses on the key feature of “ultimate” moral responsibility: it is detached from any practical considerations such as reform or deterrence and is based exclusively on justly deserved punishment or reward.

Role-responsibility and taking-responsibility Another type of responsibility – quite different from “basic desert” moral responsibility – is what H.L.A. Hart called role-responsibility. Role-responsibility is valuable and common: A sea captain is responsible for the safety of his ship, and this is his responsibility, or one of his responsibilities. . . . A sentry for alerting the guard at the enemy’s approach; a clerk for keeping the accounts of his firm. These examples of a person’s responsibilities suggest the generalization that, whenever a person occupies a distinctive place or office in a social organization, to which specific duties are attached to provide for the welfare of others or to advance in some specific way the aims or purposes of the organization, he is properly said to be responsible for the performance of these duties.  .  .  . If two friends, out on a mountaineering expedition, agree that the one DOI: 10.4324/9781003282242-40 368

Responsibility without blame

shall look after the food and the other the maps, then the one is correctly said to be responsible for the food, and the other for the maps, and I would classify this as a case of role-responsibility. (Hart 1968: 212) Our role-responsibilities are often very important to us, even to our sense of identity. I have role-responsibility for the classes I teach and will resent any effort to take over my classes and dictate the way I must teach them. That would imply that I am not capable of carrying out my teaching role effectively, and my role as a teacher is an important part of who I am. Role-responsibility does not imply moral responsibility. Imagine a soldier assigned the role of sentry by the commanding officer, but who – after an exhausting 16-hour march – is physically incapable of staying awake, his best efforts notwithstanding. Even believers in moral responsibility will have a hard time maintaining that the exhausted soldier is morally responsible for failing his role-responsibility, but his role-responsibility is undiminished. Or from the other direction, I have role-responsibility for planning the departmental picnic, and the results are fabulous. I claim that I justly deserve special credit for the successful picnic because it was my role-responsibility. Someone might respond that I  deserve no credit: it was indeed my role-responsibility, but I was lazy and disorganized, and the picnic preparations were a disaster until the diligent department secretary stepped in and brilliantly arranged everything. In that case my role-responsibility remains clear but my moral responsibility for the success is very doubtful, and that is sufficient to draw a clear line between role and moral responsibility. I can be assigned role-responsibility, and in many cases I  can take role-responsibility by volunteering for the task. But moral responsibility cannot be assigned, and it cannot be taken by claiming it. If the department chair assigns me the role of teaching a graduate seminar in medieval philosophy, then I have role-responsibility for teaching the course; but I am grossly unqualified for such an assignment, and it is no surprise that the course is terrible. At that point the chair might attempt to assign me the blame – the moral responsibility – for the bungled course, but that attempt will be in vain. The question of who if anyone is morally responsible for the disastrous course cannot be settled by assigning moral responsibility nor by my claim of moral responsibility. If I wish to prevent the chair from suffering blame for an ill-considered course assignment, I might attempt to claim moral responsibility for the disaster; but that claim will prove nothing about who justly deserves blame for the disastrous course.

Taking responsibility Appeals to taking responsibility are sometimes offered as extra support for dubious claims of moral responsibility. Harry Frankfurt’s “deep self” account of moral responsibility is based on reflective approval – at a higher “second order” level – of one’s motives and values; but just in case, Frankfurt adds “taking responsibility” to his account: “To the extent that a person identifies himself with the springs of his actions, he takes responsibility for those actions and acquires moral responsibility for them” (1975: 122). Frankfurt’s reflective “identification” is more likely to be socially self-serving than accurate: the greedy man consumed by the desire for wealth convinces himself that his deepest value is caring for his family (he desires wealth only to provide comfort and security for his beloved family). As the prophet Jeremiah taught us: “The heart is deceitful above all things, and desperately 369

Bruce Waller

wicked: who can know it?” (Jeremiah 17:9 KJV) What Jeremiah knew contemporary psychologists have confirmed and elaborated (Wilson 2002; Kahneman 2011; Panksepp  & Biven 2012; Dehaene 2014; Davies 2019). But even if one accepts Frankfurt’s dubious model of moral responsibility, that responsibility must be established by the addict’s decisive identification with and approval of his own values: “taking” responsibility contributes nothing. Robert Kane makes a brilliant effort to establish moral responsibility, scrupulously avoiding appeals to miraculous powers of self-creation and instead relying on quantum indeterminacy amplified by chaos: chaos that results from a deliberative struggle over attractive alternatives. The ultimate choice between the alternatives contains an important element of indeterminacy, and Kane appeals to “taking responsibility” to make that indeterminate choice the actor’s own choice: In making such a choice we say, in effect, “let’s try this. It is not required by my past, but it is consistent with my past and is one branching pathway in the garden of forking paths my life can now meaningfully take. Whether it is the right choice, only time will tell. Meanwhile, I am willing to take responsibility for it one way or the other.” . . . Suppose we were to say to such persons: “But look, you didn’t have sufficient or conclusive prior reasons for choosing as you did since you also had viable reasons for choosing the other way.” They might reply: “True enough. But I did have good reasons for choosing as I did, which I’m willing to stand by and take responsibility for.” (2007: 41–42) But if the choice is not already my own moral responsibility, then “taking responsibility” will not make it so. Suppose there is uncertainty over who is morally responsible for a tragic accident that was the result of negligence, and two candidates for moral blame – Giselle and Irene – are under consideration. Giselle volunteers to settle the issue by taking responsibility for the negligently caused accident. That may be noble of Giselle but it does nothing to settle the question of Giselle’s moral responsibility. Unlike role-responsibility, moral responsibility is not a responsibility one can “take.” Elinor Mason uses “taking responsibility” to expand moral responsibility into realms where many advocates of moral responsibility have doubts about its application: I argue that the zone of responsibility can be extended to include acts that we are not fully in control of, and acts whose moral status we are nonculpably ignorant about at the time of acting. This extension of responsibility happens through a voluntary taking of responsibility. (2018: 164) Mason insists that: “When an agent takes on responsibility, she is responsible. . . . My argument is that one can change one’s standing with regard to an inadvertent action, and if one takes on responsibility for it, one really is responsible” (2018: 179). Mason focuses on an important concern that is too often overlooked: deep but unrecognized harmful biases. In the moral responsibility system taking such concerns seriously requires treating the offenses as wrongs committed by an individual who is morally responsible for the harms and justly deserves blame. But Mason makes it clear that in many 370

Responsibility without blame

cases of implicit bias the biased person was shaped by a cultural environment that blocked awareness of the bias; and when the wrongs are the result of “nonculpable ignorance” it is difficult (even for those who believe in moral responsibility) to hold the biased person morally responsible for this destructive moral flaw. Mason finds that intolerable, insisting “we need something closer to full-blown blame, despite the admitted lack of bad will” (2018: 164). “Taking responsibility” is her proposed solution. Mason insists that when we unknowingly harm someone due to implicit bias, then: “The important issue is that we take seriously that we are wronging someone” (2018: 175), and “Taking it seriously involves taking responsibility for it, feeling bad about it” (2018: 175). Indeed we should feel deep remorse for our character flaws and the harm they cause, but taking moral responsibility is not the best way of taking the problem seriously, feeling genuine remorse, and fixing the problem. Suffering blame and shame for something over which one had no control does not encourage honest examination of the profound flaw in one’s character; and if an individual takes moral responsibility for a destructive implicit bias (of which she was unaware) she treats the cultural causes as if they are irrelevant. The deep problem of bias requires precisely the insightful causal analysis Mason gives (2018: 167) of the powerful effects of culture. Preserving belief in basic desert moral responsibility requires blocking inquiry into deeper causal (including cultural) factors: inquiry that reveals the decisive influence of causes we did not choose or control and of which we were often not aware. What you can do – when you recognize the terrible flaw in your character and the harm that your own flawed character has caused – is take seriously the moral wrong of bias and work diligently to expurgate it from your character and your culture.

Attributivism Mason favors an attributivist approach to moral responsibility because she is concerned about terrible wrongs stemming from deep flaws such as implicit bias and she demands that we take them seriously. Others – such as Pamela Hieronymi (2001: 530, 537, 546) – favor the attributive view for similar worthwhile reasons: they want to legitimize protest and anger over deep moral wrongs such as sexism and racism, no matter what caused the racists and sexists to become morally flawed. Anger at those moral wrongs and their deep social and individual destructiveness is an essential motivation for the hard work of bringing about change (Damasio 1994). For all those excellent reasons, the attributivists (Arpaly 2003; Hieronymi 2001, 2004; Smith 2008; Scanlon 1998, 2008; Sher 2006, 2009; Mason 2018, 2019; Talbert 2012, 2019) favor a narrowly focused account of moral responsibility: the key question is whether an action or a character trait is attributable to an agent for making assessments of that person’s moral character. If one performs morally bad acts from a morally flawed character that suffices for moral responsibility, and we should not be distracted by deeper questions concerning why such morally flawed characters were shaped. Attributivists embrace Ronald Reagan’s neoliberal perspective on moral responsibility: “Our forebears were never concerned about why a person misbehaves. We are straying from the principle of holding the individual responsible for his actions” (quoted in Beckett 1997: 66). It is important that we recognize and oppose and protest bad behavior; but it is most important to change that behavior and its deep cultural and psychological causes (Raine 2013). Legitimate anger and forceful protest require recognizing that racism and sexism are egregious moral wrongs that cause enormous suffering; they do not require that we blame those who are racists and sexists. To change the culture as well as flawed individuals we 371

Bruce Waller

must understand the deeper causes that are promoting racist and sexist attitudes. Blame and shame block deeper cultural inquiry and discourage individuals from looking hard at their own flawed characters: in their efforts to avoid blame they are more likely to hide their flaws – especially from themselves – or embrace their racism and sexism as morally justified.

Take-charge responsibility We cannot take moral responsibility, but there is an expanded version of role-responsibility – what I have called “take-charge” responsibility (Waller 2011: 103–114) – that we can and should take. Take-charge responsibility is the special responsibility we claim and take for our own lives and projects and choices. But like role-responsibility, take-charge responsibility is not the moral responsibility of just deserts. If I am minimally competent, it is of great importance – and psychological benefit – to exercise take-charge responsibility for my own life and career and choices. I want to be in control of those choices, and not under the control of someone else. Perhaps you could manage my life better than I can; that does not alter the fact that I want to have control – exercise take-charge responsibility – for my own life. But having take-charge responsibility for managing my own life does not imply that I am morally responsible for managing it well or ill. I am less competent than you, but I have as much right to my own take-charge responsibility as you do, and as much psychological need for exercising the valuable control afforded by take-charge responsibility. So long as we are competent the difference in our abilities has no bearing on our right and our need to take responsibility for our own life choices; but if the question is moral responsibility, then ignoring those differences results in profound injustice: the injustice of punishing and rewarding on the basis of differing abilities we neither chose nor controlled (unless we are blessed with miraculous powers to totally create ourselves).

The “benefits” of moral responsibility Advocates of the moral responsibility system often claim that it is justified by its moral benefits (Dennett 1984: 164; Smilansky 2000: 145–191; Vargas 2007: 155–156). One of the most impressive accounts of those moral benefits is Victoria McGeer’s “scaffolding” model of responsibility. McGeer treats developing moral competence as an intelligent practice involving the development of special skills: skills that require constant work to maintain. She emphasizes that “environmental feedback” – from one’s social group – is vital to developing and sustaining such moral skills. As McGeer states: “The capacity to recognize and respond to moral reasons is an essentially social skill, requiring social feedback to develop and maintain” (2019: 313). McGeer’s account of moral development is a great improvement over traditional accounts that treat moral development as a highly intellectualized individual process with little attention to the need for social support. But problems emerge when McGeer recommends moral responsibility practices – including “guilt, shame, remorse, and contrition” (2019: 313) – as the optimum method for promoting moral skills. The “blame and shame” of moral responsibility are not effective tools for accomplishing the vitally important goal of “developing and sustaining our capacity to recognize and respond to moral reasons” (2019: 313); to the contrary, they are more likely to limit and impede moral development. To understand the destructiveness of the moral responsibility system, consider it writ large in an industrial setting. Toyota’s early auto manufacturing process was disastrous. Its severely top-down control model demanded that workers follow 372

Responsibility without blame

orders without thinking or questioning. Mistakes on the assembly line were blamed on individual workers and severely punished: the problem was “solved” when the individual worker was fired. Workers hid mistakes when possible and tried to shift the blame to others when mistakes could not be concealed. Small problems were covered up until they became big problems, cars rolled off the assembly line with multiple defects, and Toyota became notorious for poor workmanship. Eventually the blame and shame control model was replaced by a no-blame systems approach (Bodek 2011) based on three radically different basic principles. First, rather than blaming workers for mistakes and problems, the detection of problems was regarded as a valuable part of improving the manufacturing process: small problems that would have been hidden by worried workers were instead exposed, examined, and fixed before they evolved into disasters; and workers who reported problems were treated as valuable contributors to improving the process. Second, problems were not treated as the fault of individual workers to be solved by firing an individual at the immediate problem source. Problems are systemic and solving them requires careful examination of the deeper causes. And finally, workers were treated as valuable contributors to accomplishing shared goals, and their expertise and insights were welcomed and treated with respect. The results were remarkable: Toyota was transformed from a manufacturer of inferior quality automobiles to a company with an earned reputation for high-quality workmanship with remarkably few manufacturing flaws. The success of the systems model in manufacturing led to its adoption in air traffic control, where near misses and catastrophic mistakes were proliferating. Instead of treating errors as evidence of individual negligence, the inevitable errors were now viewed as vitally important indications of deep systemic problems that required cooperative shared efforts to resolve (Sabatini 2008; Harris & Muir 2005). Small errors were reported and fixed before they became disasters, and cooperative efforts resulted in effective ways of radically reducing errors and preventing the inevitable errors from becoming disasters. Airline safety improved dramatically. The systems model rejects blame and shame and guilt and produces safer procedures and improved products. When we apply the same principles to our treatment of individuals – show respect for their abilities and persons, reject blame and shame and just deserts, seek to find and understand and fix the deep systemic social problems rather than narrowly focus on the individual  – the results are better behavior and a genuine appreciation of moral considerations. There is no doubt that blame and guilt have a deep influence; but what they primarily shape are improvements in our already polished ability to invent accounts of our behavior that make it socially acceptable (Haidt 2012: 46) and convince both ourselves and our conversation partners of our moral uprightness. Rather than enlarging the understanding of moral wrongs and deepening the appreciation of their destructiveness, they block examination of systemic causal processes and stifle deeper appreciation of the real value of avoiding such harms (Gogoshin 2023). When small mistakes are treated with punitive shaming then problems are hidden rather than discussed candidly and sympathetically, and small covered-up problems result in more serious moral flaws. Rather than internalizing moral values and nurturing a shared commitment to the importance of treating others with respect and concern, moral values become a set of rules to be followed at one’s peril. One need not be a clinical psychologist – struggling to help those hamstrung by guilt – to realize that shame and blame are poor tools for shaping positive appreciation of moral values. There is much to appreciate in McGeer’s scaffolding model of responsibility and her skill-based model of morality, including the essential role of others’ reactive attitudes in 373

Bruce Waller

shaping our moral skills. But we can embrace a wide range of valuable reactive attitudes – including resentment, anger, and remorse – while rejecting the blame and shame reactive attitudes associated with just deserts and moral responsibility. Resentment is a valuable reactive emotion that challenges disrespectful treatment and asserts one’s self-worth. Alicia can and should resent the biased jerk who treated her disrespectfully and might well reproach him; but she can do that  – and recognize his deep moral flaw and his vile behavior – while denying that he is morally responsible. Perhaps she believes that no one is ever morally responsible, or her denial of moral responsibility may be more localized: she knows this person was raised in a profoundly sexist home, that he joined a fraternity that reveled in its gross sexism, and his “old boys network” is sexist to the core; and this person’s early environment shaped him as a “cognitive miser” who finds deliberation aversive (Cacioppo & Petty 1982), so he has never questioned the values he absorbed from his relentlessly sexist environment. But his lack of moral responsibility does not change the fact that his behavior was repugnant, and resentment is appropriate. Alicia can and should be angry at his insulting behavior and her denial of moral responsibility in no way precludes such legitimate anger: anger that is vitally important in motivating opposition to immoral behavior (Damasio 1994). Alicia can also take a powerful attitude of reproach (Calhoun 1989) that informs the offender that he has committed a serious moral wrong, while still not believing that he is morally responsible for that wrong. If the sexist offender takes to heart Alicia’s expression of resentment and her sincere reproach, perhaps he will come to realize the nasty nature of his conduct and the harm it has caused many people and feel deep remorse for his destructive acts. He can and should feel this remorse even if he rejects the view that he is morally responsible for his harmful behavior: he feels genuine remorse on recognizing that his character contains a deep moral flaw, no matter how that flaw was implanted. He can and should sincerely apologize (Waller 2007) to those he now recognizes were harmed by his own morally deficient actions. The social environment – including others’ reactive attitudes – are vitally important in developing better moral skills; but moral skills development works better in the absence of blame and shame and just deserts. Blame and shame encourage the cover-up of our moral flaws, discourages us from looking deeper at social and psychological causes of behavior (causes that threaten belief in moral responsibility), justifies extreme social and economic inequity, focuses narrowly on the individual (especially of the “rugged” variety), and leaves a debilitating long-time burden of guilt.

Responsibility and belief in a just world Decades of research (Furnham 2003; Hafer & Bègue 2005) reveal a vitally important but largely unrecognized factor supporting the stubborn belief in just deserts and moral responsibility: the deep and generally nonconscious belief in a just world. A clear statement of the basic Belief in a Just World (BJW) is provided by researcher Adrian Furnham: “The BJW asserts that, quite justly, good things tend to happen to good people and bad things to bad people despite the fact that this is patently not the case” (2003: 795). Its obvious falseness notwithstanding, belief in a just world is powerful: Individuals have a need to believe that they live in a world where people generally get what they deserve. The belief that the world is just enables the individual to confront his physical and social environments as though they were stable and orderly. Without 374

Responsibility without blame

such a belief it would be difficult for the individual to commit himself to the pursuit of long range goals or even to the socially regulated behaviour of day to day life. . . . Since the belief that the world is just serves such an important adaptive function for the individual, people are very reluctant to give up this belief, and they can be greatly troubled if they encounter evidence that suggests that the world is not really just or orderly after all. (Lerner & Miller 1978: 1031) Belief in a just world is sometimes consciously endorsed: What goes around comes around; everything happens for a reason; “whatsoever a man soweth, that shall he also reap” (Galatians 6:7); karma brings ultimate justice. But the belief more often survives and flourishes by slipping beneath our conscious radar and exerting its influence in insidious and destructive ways. The evidence of profound injustice – innocent victims of war and abuse and ethnic prejudice, children dying of terrible diseases, hard-working morally decent people suffering poverty while those who exploit them live in wasteful luxury – is too obvious to consciously deny. Innocent victims challenge our comfortable belief in a just world, but there is a convenient solution to that problem: the “innocent” victims are not so innocent after all. In a just world no innocent person would suffer the brutal and traumatic crime of rape, so rape “victims” must become responsible for their own suffering by their provocative dress and reckless behavior: they are “loose” women who “led him on” and were “asking for it.” Strong belief in a just world encourages blaming victims (Wagstaff 1983; Furnham & Gunter 1984; Smith 1985; Harper & Manasse1992; Dalbert & Yamauchi 1994; Mohiyeddini & Montada 1998; Montada 1998; Appelbaum 2002; Appelbaum et al. 2006; SakalliUgurlu et al. 2007). The poor are lazy and the rich are virtuous (Furnham & Gunter 1984; Oldmeadow & Fiske 2007). Strong belief in a just world encourages blaming of those suffering from AIDS (Connors & Heaven 1990), eating disorders (Ebneter et al. 2011), mental illness (Bizer et al. 2012), and various forms of illness (Nudelman & Shiloh 2011). Abused children blame themselves to preserve their belief in a just world (Lerner 1980: 123–125). Moral responsibility protects belief in a just world: punishment becomes righteous retribution, and innocent victims are transformed into guilty villains who are getting their just deserts. Belief in a just world and belief in moral responsibility form a powerful mutually supportive alliance – though neither belief is plausible in a cruel world devoid of miraculous self-making powers. Belief in a just world is not a justification of belief in moral responsibility; to the contrary, it is an explanation for widespread belief in such an implausible doctrine. Philosophers may insist that their own allegiance to moral responsibility does not depend on BJW, which they reject; but the deep philosophical commitment to “ought implies can” – the world is morally ordered in such a manner that we can always fulfill our obligations – reveals the subtle but powerful influence of BJW on philosophical convictions. The belief that justice is built into the structure of the world has been a deep influence on philosophical thought. As Bernard Williams notes: Plato, Aristotle, Kant, Hegel are all on the same side, all believing in one way or another that the universe or history or the structure of human reason can, when properly understood, yield a pattern that makes sense of human life and human aspirations. (1993: 163) 375

Bruce Waller

Kant’s rationalistic moral philosophy offers an “equal capacity” justification of moral responsibility and a just world: The capacity for moral agency is supposedly present to any rational agent whatsoever, to anyone for whom the question can even present itself. The successful moral life, removed from considerations of birth, lucky upbringing, or indeed of the incomprehensible Grace of a non-Pelagian God, is presented as a career open not merely to the talents, but to a talent which all rational beings necessarily possess in the same degree. Such a conception has an ultimate form of justice at its heart, and that is its allure. Kantianism is only superficially repulsive – despite appearances, if offers an inducement, solace to a sense of the world’s unfairness. (1981: 21) Belief in a just world offers “solace to a sense of the world’s unfairness,” but the price is exacerbation of the world’s unfairness. Moral responsibility is the essential partner and protector of belief in a just world (Waller 2015: 61–68, 2018: 67–68). The depth and range of profound injustice in our world destroys belief in a just world unless we believe that the “injustice” is only apparent: rape victims brought their suffering on themselves, suffering infants are receiving their just deserts for their “original sin,” the poor are indolent, “victims” of ethnic bias are morally vile, cruel punishment becomes righteous retribution, the enormous gap between the impoverished and the wealthy marks the difference in their moral worth. It is hard to look at our unjust world without the rose-colored glasses of moral responsibility: glasses that make the victims of oppression into villains and transforms the powerful and privileged oppressors – from George Washington to Ronald Reagan – into paragons of virtue. Expressing resentment toward someone’s behavior involves discomfort for the person targeted, and experiencing genuine remorse involves suffering. You should feel and express resentment toward those who demean you, and that will cause those persons suffering; but in the absence of moral responsibility no one justly deserves to suffer, and in a just world no one can have an obligation to cause unjust suffering. With moral responsibility, the problem is solved: our resentment harms someone who justly deserves to be harmed, and we are free of evil taint. In a just world it is not possible for me to suffer remorse when I am not morally responsible and do not justly deserve such suffering. But we do not live in a just world. No one justly deserves to suffer the deep pangs of remorse, but such remorse is essential to moral development and reform. In like manner, no one justly deserves punishment: inflicting punishment is morally wrong. But it should not be surprising that in this unjust world sometimes we cannot avoid injustice. We ought not punish those who do not justly deserve punishment: inflicting such punishment should deeply disturb us and motivate us to minimize punishment and make unavoidable punishment as mild and as beneficial and respectful as possible, and also motivate efforts to change the destructive conditions that shape people in ways that require punishment. In Kant’s ideal world, if we ought not punish then we can avoid punishing; in our unjust world, we cannot eliminate unjust punishment. The blame and shame of moral responsibility blocks better methods of encouraging strong take-charge responsibility for our moral lives. Seeking to understand the social and psychological factors that shape our characters, striving to understand and change the systemic social factors that need changing, and promoting the cooperative open understanding 376

Responsibility without blame

of our moral mistakes and the harms they cause (rather than encouraging the denial and coverup of moral problems through shame and blame): these are better methods of developing a deep appreciation of and affective commitment to the essential moral values of respect and concern for the welfare of others. There are valuable forms of responsibility: basic desert moral responsibility is not among them.

Further reading (added by the editor) Caruso, Gregg D. (2021). Rejecting Retributivism: Free Will, Punishment, and Criminal Justice. New York: Cambridge University Press & Pereboom, Derk. (2014). Free Will, Agency, and Meaning in Life. Oxford: Oxford University Press, offer a detailed discussion of the desert-based model of responsibility. Waller, Bruce N. (2011). Against Moral Responsibility. Cambridge, MA: MIT Press & Smilansky, Saul. (2000). Free Will and Illusion. New York: Oxford University Press, present an alternative view and argue against desert for praise or blame. Hart, H. L. A. (1968). Punishment and Responsibility. Oxford: Clarendon Press, offers a discussion of responsibility in legal contexts and draws attention to a “welter” of different grammatical sense of the word responsibility, respectively responsible.

References Appelbaum, Lauren D. (2002). “Who Deserves Help? Students’ Opinions About the Deservingness of Different Groups Living in Germany to Receive Aid,” Social Justice Research, 15(3), 201–225. Appelbaum, Lauren D., Lennon, Mary Clare, & Aber, J. Lawrence. (2006). “When Effort Is Threatening: The Influence of the Belief in a Just World on Americans’ Attitudes Toward Antipoverty Policy,” Political Psychology, 27(3), 387–402. Arpaly, Nomy. (2003). Unprincipled Virtue: An Inquiry into Moral Agency. Oxford: Oxford University Press. Beckett, Katherine. (1997). Making Crime Pay: Law and Order in Contemporary American Politics. New York: Oxford University Press. Bizer, George Y., Hart, Joshua, & Jekogian, Allison M. (2012). “Belief in a Just World and Social Dominance Orientation: Evidence for a Mediational Pathway Predicting Negative Attitudes and Discrimination Against Individuals with Mental Illness,” Personality and Individual Differences, 52, 428–432. Bodek, Norman. (2011). “Zenjidoka, Solving Toyota’s Quality Problems,” Quality Digest, https:// www.qualitydigest.com/inside/management-article/zenjidoka-solving-toyotas-quality-prob lems-021411.html. Cacioppo, John T., & Petty, Richard E. (1982). “The Need for Cognition,” Journal of Personality and Social Psychology, 42(1), 116–131. Calhoun, Cheshire. (1989). “Responsibility and Reproach,” Ethics, 99(2), 389–406. Caruso, Gregg D. (2018). “Skepticism About Moral Responsibility,” in Edward N. Zalta (ed.), The Stanford Encyclopedia of Philosophy. Stanford, CA: Metaphysics Research Lab, Stanford University. ———. (2021). Rejecting Retributivism: Free Will, Punishment, and Criminal Justice. New York: Cambridge University Press. Connors, John, & Heaven, Patrick C. (1990). “Belief in a Just World and Attitudes Toward AIDS Sufferers,” The Journal of Social Psychology, 130(4), 559–560. Dalbert, Claudia, & Yamauchi, Lois A. (1994). “Belief in a Just World and Attitudes Toward Immigrants and Foreign Workers: A Cultural Comparison Between Hawaii and Germany,” Journal of Applied Social Psychology, 24(18), 1612–1626. Damasio, Antonio R. (1994). Descartes’ Error: Emotion, Reason, and the Human Brain. New York: Putnam. Davies, Paul Sheldon. (2019). “Darwinizing Debunking Arguments,” Ratio, 32(4), 275–289. Dehaene, Stanislas. (2014). Consciousness and the Brain. New York: Penguin. Dennett, Daniel. (1984). Elbow Room. Cambridge, MA: MIT Press.

377

Bruce Waller Ebneter, Daria S., Latner, Janet D., & O’Brien, Kerry S. (2011). “Just World Beliefs, Causal Beliefs, and Acquaintance: Associations with Stigma Toward Eating Disorders and Obesity,” Personality and Individual Differences, 51, 618–622. Frankfurt, Harry G. (1975). “Three Concepts of Free Action: Part 2,” Aristotelian Society: Supplementary Volume (supplement 49), 113–125. Furnham, Adrian. (2003). “Belief in a Just World: Research Progress Over the Past Decade,” Personality and Individual Differences, 34(5), 795–817. Furnham, Adrian F., & Gunter, Barrie. (1984). “Just World Beliefs and Attitudes Towards the Poor,” British Journal of Social Psychology, 23(3), 265–269. Gogoshin, Dane Leigh. (2023). “A  Challenge for the Scaffolding View of Responsibility,”  Ethical Theory and Moral Practice, 26(1), 73–90. Hafer, Carolyn L., & Bègue, Laurent. (2005). “Experimental Research on Just-world Theory: Problems, Developments, and Future Challenges,” Psychological Bulletin, 131(1), 128–167. Haidt, Jonathan. (2012). The Righteous Mind: Why Good People Are Divided by Politics and Religion. New York: Pantheon Books. Harper, David J., & Manasse, Paul R. (1992). “The Just World and the Third World: British Explanations for Poverty Abroad,” The Journal of Social Psychology, 132(6), 783–785. Harris, Don, & Muir, Helen C. (2005). Contemporary Issues in Human Factors and Aviation Safety. Aldershot: Ashgate. Hart, H. L. A. (1968). Punishment and Responsibility. Oxford: Clarendon Press. Hieronymi, Pamela. (2001). “Articulating an Uncompromising Forgiveness,” Philosophy and Phenomenological Research, 62(3), 529–555. ———. (2004). “The Force and Fairness of Blame,” Philosophical Perspectives, 18, 115–147. Kahneman, Daniel. (2011). Thinking, Fast and Slow. New York: Farrar, Straus and Giroux. Kane, Robert. (2007). “Libertarianism,” in Robert Kane, John Martin Fischer, Derk Pereboom, & Manuel Vargas (eds.), Four Views on Free Will (pp. 5–43). Malden, MA: Blackwell Publishing. Lerner, Melvin J. (1980). The Belief in a Just World: A Fundamental Delusion. New York: Plenum Press. Lerner, Melvin J., & Miller, Dale T. (1978). “Just World Research and the Attribution Process: Looking Back and Ahead,” Psychological Bulletin, 85(5), 1030–1051. Mason, Elinor. (2018). “Respecting Each Other and Taking Responsibility for our Biases,” in Marina Oshana, Katrina Hutchison, & Catriona Mackenzie (eds.), Social Dimensions of Moral Responsibility (pp. 163–184). New York: Oxford University Press. ———. (2019). “Between Strict Liability and Blameworthy Quality of Will: Taking Responsibility,” in David Shoemaker (ed.), Oxford Studies in Agency and Responsibility, vol. 6 (pp.  240–263). New York: Oxford University Press. McGeer, Victoria. (2019). “Scaffording Agency: A  Proleptic Account of the Reactive Attitudes,” European Journal of Philosophy, 27(2), 301–323. Mohiyeddini, Changiz, & Montada, Leo. (1998). “BJW and Self-Efficacy in Coping with Observed Victimization: Results from a Study About Unemployment,” in Leo Montada & Melvin Lerner (eds.), Responses to Victimizations and Belief in a Just World (pp. 41–54). New York: Plenum. Montada, Leo. (1998). “Belief in a Just World: A  Hybrid of Justice Motive and Self-Interest,” in Leo Montada  & Melvin Lerner (eds.), Responses to Victimizations and Belief in a Just World (pp. 217–245). New York: Plenum. Nudelman, Gabriel, & Shiloh, Shoshana. (2011). “Who Deserves to Be Sick? An Exploration of the Relationships Between Belief in a Just World, Illness Causal Attributions and Their Fairness Judgements,” Psychology, Health, & Medicine, 16(6), 675–685. Oldmeadow, Julian,  & Fiske, Susan T. (2007). “System-Justifying Ideologies Moderate Status=Competence Stereotypes: Roles for Belief in a Just World and Social Dominance Orientation,” European Journal of Social Psychology, 37, 1135–1148. Panksepp, Jaak,  & Biven, Lucy. (2012). The Archaeology of Mind: Neuroevolutionary Origins of Human Emotions. New York: W. W. Norton. Pereboom, Derk. (2014). Free Will, Agency, and Meaning in Life. Oxford: Oxford University Press. Raine, Adrian. (2013). The Anatomy of Violence: The Biological Roots of Crime. New York: Pantheon Books. Sabatini, Nicholas A. (2008). “Reaching the Next Level of Aviation Safety,” FAASTeam News (Federal Aviation Administration – FAASTeam – FAASafety.gov).

378

Responsibility without blame Sakalli-Ugurlu, N., Yalcin, Z., & Glick, P. (2007). “Ambivalent Sexism, Belief in a Just World, and Empathy as Predictors of Turkish Students’ Attitudes Toward Rape Victims,” Sex Roles, 57, 889–895. Scanlon, T. M. (1998). What We Owe to Each Other. Cambridge, MA: Harvard University Press. ———. (2008). Moral Dimensions: Permissibility, Meaning, Blame. Cambridge, MA: Harvard University Press. Sher, George. (2006). In Praise of Blame. Oxford: Oxford University Press. ———. (2009). Who Knew? Responsibility Without Awareness. New York: Oxford University Press. Smilansky, Saul. (2000). Free Will and Illusion. New York: Oxford University Press. Smith, Angela M. (2008). “Control, Responsibility, and Moral Assessment,” Philosophical Studies: An International Journal for Philosophy in the Analytic Tradition, 138(3), 367–392. Smith, Kevin B. (1985). “Seeing Justice in Poverty: The Belief in a Just World and Ideas About Inequalities,” Sociological Spectrum, 5, 17–29. Strawson, Galen. (1994). “The Impossibility of Moral Responsibility,” Philosophical Studies, 75(1–2), 5–24. Talbert, Matthew. (2012). “Moral Competence, Moral Blame, and Protest,” The Journal of Ethics, 16(1), 89–109. ———. (2019). “The Attributionist Approach to Moral Luck,” Midwest Studies in Philosophy, 43(1), 24–41. Vargas, Manuel. (2007). “Revisionism,” in Robert Kane, John Martin Fischer, Derk Pereboom, & Manuel Vargas (eds.), Four Views on Free Will (pp. 126–165). Malden, MA: Blackwell. Wagstaff, Graham F. (1983). “Correlates of the Just World in Britain,” The Journal of Social Psychology, 121(1), 145–146. Waller, Bruce N. (2007). “Sincere Apology Without Moral Responsibility,” Social Theory and Practice, 33(3), 441–465. ———. (2011). Against Moral Responsibility. Cambridge, MA: MIT Press. ———. (2015). The Stubborn System of Moral Responsibility. Cambridge, MA: MIT Press. ———. (2018). The Injustice of Punishment. New York: Routledge. Williams, Bernard. (1981). Moral Luck. Cambridge: Cambridge University Press. ———. (1993). Shame and Necessity. Berkeley, CA: University of California Press. Wilson, Timothy D. (2002). Strangers to Ourselves: Discovering the Adaptive Unconscious. Cambridge, MA: Harvard University Press. Zimmerman, Michael J. (2015). “Varieties of Moral Responsibility,” in Randolph Clarke, Michael McKenna, & Angela M. Smith (eds.), The Nature of Moral Responsibility (pp. 45–64). New York: Oxford University Press.

379

30 HOLDING RESPONSIBLE IN THE AFRICAN TRADITION Reconciliation applied to punishment, compensation, and trials Thaddeus Metz 1  Introducing African approaches to holding responsible This chapter draws on ideas salient in the African philosophical tradition to articulate principles about how to hold people responsible properly for wrongdoing, principles that would be found prima facie attractive by many ethicists, philosophers of law, and jurists not merely within that tradition, but also beyond it. In the works of African philosophers and related thinkers who have addressed the question of how to respond to wrongdoing, reconciliation (sometimes spoken of in terms of “repairing relationships” or “restoring harmony”) is often posited as the principal, if not sole, final aim. This chapter expounds a philosophical interpretation of reconciliation that is meant to have broad appeal, after which it is applied to punitive (criminal) justice, compensatory (civil) justice, and the kinds of trials that would be apt for both. In all three contexts, the chapter provides reason to take the African approaches seriously that will be found plausible (even if not convincing) by those from Western or otherwise non-African backgrounds, supporting the idea that reconciliation is at least one proper final aim when holding people responsible for wrongdoing. It was South Africa’s Truth and Reconciliation Commission (TRC), operating in the mid 1990s, that put reconciliation on the map for a global audience regarding how to hold people responsible for serious human rights violations (Truth and Reconciliation Commission 1998). Famously for the TRC, reconciliation consequent to apartheid-era political crimes meant that, in exchange for disclosing the full truth about them in a public forum, offenders would be free from any criminal and civil liability and expected to contribute to a new, democratic society. However, reconciliatory projects of various kinds had taken place previously in Zimbabwe in the early 1980s and have also taken place after the TRC in countries that include Rwanda and Sierra Leone, both largely in the early 2000s (for overviews, see Huyse & Salter 2008). The broad aim of seeking to repair broken relationships, whether for human rights violations or less severe wrongs, grows mainly out of an ethic widely shared by peoples indigenous to the African continent that prizes harmonious or communal relationships (and not really Christianity, on which see Tutu 1999; Krog 2008). Although a reconciliatory orientation naturally follows from such a relational ethic, this chapter submits that one need not hold such an ethic in order to find attractive a certain DOI: 10.4324/9781003282242-41 380

Holding responsible in the African tradition

conception of reconciliation and its implications for how to hold people responsible for wrongdoing. The next section spells out an interpretation of reconciliation that is intended to avoid problems with the TRC’s approach to it, viz., of having failed both to hold offenders accountable adequately and compensate victims decently (section 2). Roughly according to the favored account of reconciliation, there is a “backward-looking” element of disavowing past wrongdoing by offenders undergoing burdens and “forward-looking” elements of offenders compensating those they have wrongfully harmed, reforming their bad character, and thereby making it reasonable for society to accept them back into the fold. After having spelled out and motivated this account of reconciliation, this chapter applies it to three different questions about how to hold people responsible for wrongdoing. The first question is how to punish people for crimes in a just manner (section 3). Although reconciliation has often been pitched as an alternative to punishment, particularly by friends of the TRC, this essay contends that the concept of reconciliatory sentencing (or restorative sanctions) is a coherent and promising alternative to the penal values of desert and deterrence that are so prominent in Western legal theory. Then the essay considers the question of how victims of wrongful harm should be compensated for it (section 4). Repairing relationships with victims, according to the interpretation advanced here, normally involves offenders doing what would improve their victims’ quality of life, ideally in some kind of consultation with them. In seeking to make victims’ lives go better, the chapter argues that reconciliatory compensation would differ, and in a plausible way, from an approach salient in Western philosophy of law, which is instead to aim to return a stolen item or more generally to realize a state of affairs that would have occurred in the absence of injustice. Finally, the chapter invokes the favored account of reconciliation to address the question of precisely whom should be held responsible for wrongdoing in a criminal or civil trial (section 5). Obviously, a party guilty of having done a wrongful act should be, but a typically African approach would not limit liability to him. For many from sub-Saharan cultures, an offender’s family is sometimes thought to be rightly held responsible for the offense to some degree, even if not as much as the direct offender. In potentially holding some family members responsible for an offender’s crime, reconciliatory trials would apportion liability in less individualized ways than is typical in the West.1 Beyond noting this contrast, this chapter provides some grounds for taking the more inclusive, characteristically African approach seriously.

2  An interpretation of reconciliation The word “reconciliation” is sensibly defined as people reuniting after conflict. If two friends have had a fight, and immediately after it decide just to forgive, forget, and resume friendly relations, that would count as reconciliation. However, it would arguably not be a truly desirable kind of reconciliation, where a better form would include at least the two parties hashing things out and the one(s) who did wrong expressing remorse, supposing of course that a wrong had been done. This section articulates a conception of reconciliation that is meant to be attractive for a variety of settings, including between friends, but particularly at the social level when injustice has been done to a stranger (initially advanced in Metz 2015, 2022a).2 The rest of the chapter applies it to various proper ways to hold people responsible for wrongdoing. 381

Thaddeus Metz

The conception of reconciliation consequent to wrongdoing advanced here has two distinguishable facets, a “backward-looking” condition of acknowledging the wrong and a “forward-looking” condition of constructive interaction between those party to the conflict. Specifically, in this chapter reconciliation is construed as a morally desirable state in which: (a) those who have been substantially affected by conflict interact on a largely voluntary, transparent, and trustworthy basis for the sake of compossible ends that are expected to be good for one another (b) and any culpable wrongdoing has been disavowed, in the first instance by the wrongdoers. The (a) condition (or something like it) is essential for anything plausibly to fall under the heading of “reconciliation.” Reconciliation is more than just two parties removing themselves from each other’s sphere of influence and also more than mere peaceful coexistence between them. Instead, reconciliation involves some kind of cohesion consequent to conflict, construed here as basically cooperation and aid. Notice that mention of “trustworthy” interaction is meant to include the idea that the victim has some good reason to think that the wrongdoer will not reoffend. In the ideal case, that would be because the offender has had a change of heart or otherwise reformed his character, making it less likely that he would do the same wrong again. This element is one thing missing from the instance of reconciliation between friends mentioned previously; even if the victim succeeds in having “moved on,” he has no good reason to trust that the other party will not repeat the behavior. While something could logically be a kind of reconciliation without the (b) condition, it (or also something like it) is probably essential for a sort that is attractive. Disavowal of wrongdoing includes several elements that are also missing from the previous instance of reconciliation between friends. For one, since, in order to disavow wrongdoing, one must be aware of it, there is an implicit truth requirement. A desirable reconciliation includes an awareness of what happened between the parties, and hence ideally a willingness on the part of the wrongdoer to hear out how his behavior affected the victim. Second, moral language (or at least conceptualization) is used to appraise the conflict that took place. Instead of using neutral terms, what transpired is evaluated by the parties from a moral point of view. Third, where something is accurately labelled “wrong” or the like, it is disapproved of and in the best case by the wrongdoer himself. The optimal form of reconciliation is surely one in which the person who did wrong accepts responsibility for having done it, apologizes for it, and undergoes certain “productive burdens” as a way to express his remorse. Such burdens would typically include feeling guilty, compensating his victim, and doing the work of changing his character so that he will not do wrong again. Where the offender refuses to acknowledge his criminality, then a third party should step in to express disapproval for his behavior, in part by imposing such burdens on him. All these backward-looking elements are missing from the account of reconciliation between friends sketched previously, while some of them are also missing from the form of reconciliation that South Africa’s TRC adopted. The Commission did use moral language to appraise the violations of human rights, and it, as well as the broader public, did listen to many victims recount how they had been mistreated (see, e.g., Truth and Reconciliation Commission 1998). However, those who had violated people’s human rights did not systematically hear out their victims. In addition, there was no expectation that those who disclosed their human rights violations would express any contrition, let alone apologize directly to their victims. Still more, the TRC legislation (Republic of South Africa 1995), 382

Holding responsible in the African tradition

upheld by the Constitutional Court of South Africa (1996), forbade the imposition of any punitive or compensatory burdens on offenders who had fully disclosed their crimes. Friends of the TRC argue that it was only by forgoing the imposition of burdens on apartheid enforcers that they would have been willing to relinquish power and allow the transition to democracy (Constitutional Court of South Africa 1996: para 65; Burton 2000: 79; Lenta 2007: 158–159). That might well be true. The point, though, is that the TRC was then, as a result, poor in terms of the sort of reconciliation it effected. One need not appeal to another moral category such as retributive justice to criticize the “kid gloves” that the TRC used on wrongdoers, requiring them merely to confess in public. Reconciliation itself provides grounds to think that offenders should have been held accountable by, say, having labored in ways that would have improved their victims’ quality of life. The TRC did make recommendations to the government about how to compensate individual victims of human rights violations and about how to effect reparations more broadly to black South Africans. However, it was not offenders who were to do the hard work of effecting redress for victims, and, furthermore, subsequent governments have by and large failed to live up to the TRC’s prescriptions (on which see, e.g., Pather 2018).

3  Reconciliatory criminal justice During the heyday of the TRC, many South African public agents advanced reconciliation as an alternative to punishment. Parliament adopted a new Constitution that said in its postamble that “there is a need for understanding but not for vengeance, a need for reparation but not for retaliation. . . . In order to advance such reconciliation and reconstruction, amnesty shall be granted” (Republic of South Africa 1993: chap. 15). The Constitutional Court of South Africa likewise construed reconciliation as an alternative to punishment when it said that “the key elements of restorative justice have been identified as encounter, reparation, reintegration and participation,” where “(r)eparation focuses on repairing the harm that has been done rather than on doling out punishment” (2006: para. 114). The Chair of the TRC, Desmond Tutu, expressed a similar approach: “[T]he central concern is not retribution or punishment, but . . . the healing of breaches, the redressing of imbalances, the restoration of broken relationships” (1999: 51). In the light of the conception of reconciliation advanced in the previous section, this chapter argues in effect that the TRC’s non-punitive approach to it was lacking. The idea of reconciliatory sentencing (or restorative sanctions) is not merely coherent, but also philosophically attractive and merits consideration as a rival to the desert and deterrence conceptions of punishment’s point that have been characteristic of Western legal thought for hundreds of years. I now argue that the conception of reconciliation advanced in the previous section grounds pro tanto reason to punish those guilty of serious crimes, principally because of its backward-looking condition requiring the disavowal of wrongdoing. A wrongdoer who did not feel guilt, go out of his way to compensate his victim, and make strenuous efforts not to commit the wrong again would be failing to reconcile in the appropriate way. In contrast, for a guilty party to place such burdens on himself expresses remorse; it shows that he now distances himself from crime committed in the past. One additional sort of burden those who have committed weighty offenses should normally undergo includes submitting to penalties prescribed by an impartial third-party representing the public as a way for it, in turn, to express disapproval of their misdeeds.3 Supposing that what has been criminalized is behavior that is particularly wrong, expressing disapproval merely by saying certain 383

Thaddeus Metz

words or wagging a finger would not be sufficient. For serious offenses, a fitting disapproval of an offender’s misdeeds must take the form of frustrating his ends or reducing his quality of life. The worse the crime, the greater the offender’s remorse and the public’s disapproval should be, and hence the more severe the punishment should be. Commensurability of some kind between offense and penalty, which need not be a strict, cardinal proportionality between them (see Metz 2019: 125–126), is hence prescribed by the logic of reconciliation. This rationale for punishment is an expressive one, according to which showing disapproval of offenses with hard treatment is part of what it means to stand up for victims, to object to the flouting of just laws, and to treat offenders as agents responsible for their serious misbehavior. From this perspective, to fail to punish is disrespectful, at least to some degree, of victims, moral norms, and even offenders, a broad view that has had adherents in recent Anglo-American philosophy of punishment (e.g., Feinberg 1965; Hampton 1988; Duff 2001). However, the forward-looking condition of the conception of reconciliation normally prescribes “productive” penalties of certain kinds that are more unfamiliar to a Western audience and alien to a retributive standpoint. Consider what two philosophers of law have said of the ways that the Yoruba people, indigenous to what is now Nigeria, have often penalized offenders. According to one, the reconciliatory factor is lacking in Western theories of law and penology where the offender is punished without making restitutions; and emerging from prison, he is reconciled neither to himself, his victim, nor to society. . . . [W]hen a culprit is punished, such is done with the view to fine-tuning the character of the said offender in line with the communalistic ethos of the Yoruba culture. (Balogun 2018: 246, 311) According to the other, those who had committed crimes that did not pose an existential threat to the community were often punished by the Yoruba by “being forced to labor on community projects or those of their victims in reparation/restitution for the loss caused” (Bewaji 2016: 164). Given the logic of reconciliation, being punished should ideally not take the form of being killed, maimed, or incarcerated, and instead, at least when feasible, should consist of laboring in ways that would reform the characters of the guilty or compensate their victims, particularly the direct ones, but also potentially indirect ones such as the victims’ relatives or the broader community. Those kinds of penalties would improve relationships, and are characteristically what reconciliation would prescribe, as opposed to inflicting “unproductive” harm on an offender merely because he deserves it in retributive fashion. Furthermore, notice how the focus on making restitution to direct victims, and in ways that are burdensome to the offender, contrasts with both punitive community service (which need not benefit direct victims) and a civil suit (which need not be burdensome to offenders or even express disapproval) in the Western tradition. Such reconciliatory sentencing (or restorative sanctioning) is arguably suitable for any context of wrongdoing in which penalties are just. A student who has plagiarized an essay should not be suspended, but instead should be required to take an extra course on why plagiarism is wrong and then instruct first-year students on the topic (Metz 2022a: 272). Someone who has cheated on his taxes should, instead of paying a fine, be made to perform dull tasks for the state revenue service (Metz 2022b: 130). A doctor who used his medical 384

Holding responsible in the African tradition

knowledge to harm black people during apartheid should not simply rot in jail, but now put it to good use in healthcare clinics servicing black neighborhoods (Metz cited in Malan & Green 2013). A drunk driver might, in contrast to merely losing his driver’s license, be sentenced to work in a morgue (BBC 2016). Reconciliatory sentencing would focus on penalties that in effect make offenders clean up their own mess, whether that is the harm they have done to their victims or the poor state of their characters. In that regard, this approach to punishment differs from both retributivism and general deterrence theory and in attractive ways. From a retributive standpoint, punishment need not do good at all in the future to be morally justified; instead, the right penalty is whichever harm or submission fits the crime that was committed in the past. Punishment does not have to clean up anything. From a general deterrence approach, punishment should do some good, and so in a sense do some cleaning, but the offender would be penalized to prevent others from making a mess. Imposing commensurate penalties likely to compensate victims or reform offenders is prima facie appealing by comparison, although this chapter lacks the space to motivate the approach any more than this (cf. Metz 2019, 2022b).

4  Reconciliatory civil justice Strictly speaking, if the reconciliatory approach to criminal justice articulated in the previous section were adopted, it might be that no separate system of civil justice would be needed. After all, the point of a punishment would be to impose burdens on those who have done culpable wrongs that not only express commensurate disapproval on the part of the state, but also are productive for having offenders compensate their direct victims. It might be, however, that sometimes it is appropriate to make one party compensate another party who has been harmed even though the former is not culpable for the harm (say, because it is much easier for the former to bear the cost), in which case civil trials could still be apt. Setting aside the question of whether something other than a criminal trial would be needed in a reconciliatory system, this section focuses on the question of precisely how those culpable for wrongful harm should he held responsible to compensate for it in a just manner. By a reconciliatory approach, the right reparation is one that would improve the victim’s quality of life consequent to some kind of consultation with her, which differs from one prominent approach in the West, which is focused on the restoration of an original state, explained shortly. The forward-looking element of reconciliation involves guilty parties relating to victims in ways that support ends of theirs, specifically ones that would advance their good; that is the sort of relationship that is meant to be restored, or indeed established in the first place if it had never been present. That means compensating so as to make victims better off and in a manner they accept. For example, black people forced off land during the apartheid era in South Africa so that white-owned mining companies could secure minerals might, some 60 years later, be given cash if that is what they choose. Someone who stole a radio from someone might provide a television, if that is what the victim prefers. If a person negligently or recklessly broke another’s arm, then, in addition to paying for the hospital bills, he might perform tasks that the wounded person cannot undertake, even if he would not have undertaken them had his arm not been injured. Notice how these approaches to compensatory justice differ from an approach that is common in the Western tradition. As a first approximation, the thought is that one must 385

Thaddeus Metz

return an object that one stole. If a thief took land or a radio, he must give back the land or the radio, and in the condition it would have been in had it not been taken. However, this sort of compensation is an instance of a broader counterfactual principle, namely, that civil, compensatory, or rectificatory justice means “restoring the victim to the position she would have been in had the wrongful behavior not occurred,” as per the Stanford Encyclopedia of Philosophy entry on justice (Miller 2017: sec. 2.2), “returning victims to a condition they would have been in in the absence of the wrong” (Oyowe 2017: 239), or putting in place “the conditions that people were relying upon when framing their plans, and so allow them to carry on with their plans with minimal interruption” (Goodin 1991: 152; see also Goodin 1989). In Robert Nozick’s influential terms, The principle of rectification presumably will make use of its best estimate of subjunctive information about what would have occurred (or a probability distribution over what might have occurred, using the expected value) if the injustice had not taken place. (1974: 152–153) The over-arching theme here is that there is a specific path a person’s life would have taken without a wrongful taking or interference, where the right compensation would put her back on that path as much as possible. Some reason to prefer the reconciliatory approach of this chapter over this counterfactual one can be seen by reflecting on two kinds of situations. In one case, note that sometimes wrongdoing, for instance theft, turns out to provide a substantial net benefit to the victim. Suppose you have stolen a Ferrari from a person, and it so happens that, had you not done so, its owner would have died, say, from a collision or because she would have caught a plane that crashed. In that case, the logic of “restoring the victim to the position she would have been in had the wrongful behavior not occurred” entails that no compensation is owed at all (or perhaps only a compensation of a kind that would result in the victim’s death)! However, you, the thief, intuitively still owe compensation and indeed one that would make the victim’s life go well, as opposed to precisely however – potentially poorly! – it would have gone absent the theft. Less hypothetically, and without appealing to a statistical anomaly to make the point, consider that one encounters the claim made by some apartheid apologists or opponents of affirmative action in South Africa that no compensation is owed to black people, since, in the absence of apartheid, which brought industrialization to a degree noticeably greater than elsewhere in Africa, their economic livelihoods on average would have been worse than they were with apartheid. If the proper aim of compensatory justice were to give victims the lives they would have had in the absence of wrongdoing, then (granting the contestable claim that most black people’s economic lives would have been worse without apartheid) there would not be a justification for large-scale reparation programmes in South Africa. However, since large-scale reparation programmes in response to South African apartheid are indeed intuitively justified, the aim of compensatory justice is not to give victims the lives they would have had in the absence of wrongdoing. In reply, one might suggest that, when interpreting the idea that compensatory justice should give victims the lives they would have had in the absence of wrongdoing, we are to consider only harm the wrong did immediately and not any further, longer-term consequences of the wrong. In respect to the apartheid case, then, perhaps we are to compensate 386

Holding responsible in the African tradition

only for, say, the loss of a house upon being forced to relocate outside a city, and not any further results of that segregationist practice, whether good or bad. However, it is common to include longer-term consequences of a wrong when determining the right compensation for it. As Bernard Boxill, who composed the entry titled “Compensatory Justice” for the International Encyclopedia of Ethics, suggests, full compensation requires making up, not merely for the harm immediately caused by a wrong done to a victim, but also further harm that the wrong brought in its wake: Compensating him for his disability . . . requires making him no worse off than he would have been had he not been disabled; doing that requires counterfactual reasoning. We must consider for example whether he would have been a sure bet to win an Olympic gold medal if he had not been disabled. And if he was we must do what we can to make his condition reasonably close to what it would have been had he won that medal. (2013: 955) Now, by the same logic, if the consequences of a wrong did not bring about net costs to a victim, but instead produced net benefits to her, then the principle of seeking to “change the present so that it looks more like the present that would have obtained in the absence of the injustice” (mentioned, but not accepted, by Waldron 1992: 8) would prescribe no change (or perhaps even change that reduces the victim’s quality of life). If all the harm done in the wake of a wrong is in need of compensation to return a victim to the state that would have taken place, as per Boxill, then, by parity of reasoning, one should take all the benefit done in the wake of a wrong into consideration when calculating what is owed. However, that principle has counterintuitive implications about when and how much compensation is owed. A second telling situation is one in which returning a stolen object would harm the victim or is not something that she would in hindsight choose to have (from Metz 2020). Returning to the hypothetical case, suppose that, while the Ferrari was in your wrongful possession, the state adopted a heavy wealth tax specifically on those who own Ferraris but not Maseratis. Imagine, too, that you have a Maserati, one that is actually yours. The right form of compensation would surely be for you to offer the Maserati to the victim, even though returning the Ferrari would mean “restoring the victim to the position she would have been in had the wrongful behavior not occurred” or “making them no worse off than they would have been but for the misfortune.” What the two kinds of cases suggest is that one ultimate rationale for compensation is to mend a relationship between a guilty party and his victim. Instead of merely returning the victim to some original state that would have transpired without wrongdoing, a major point of civil justice is to hold those guilty of wrongful harm responsible in a way that expresses a suitable apology and shows respect for the victim as a person with a will and a good. That means that the guilty party should provide to the victim what she agrees would make her life go better, or so this chapter submits merits consideration by the field. An interesting implication of the reconciliatory approach is that negotiation is plausibly part of the right compensatory process, as it traditionally has been in indigenous sub-Saharan societies (see Masitera 2018: 114–116). It is not enough merely for the guilty party to improve the victim’s quality of life, but ideally he should also do so in a way that the victim accepts. That principle does not entail that the guilty party must provide whatever a victim 387

Thaddeus Metz

wants, regardless of how little or how much good it would do her. What it does mean that the guilty should seek to ascertain how a victim would like her life to be improved, and that the amount of improvement should be roughly proportionate to the wrong and harm done, as a way to express remorse fittingly.

5  Reconciliatory trials The previous sections discussed what a certain conception of reconciliation entails for how to hold responsible when it comes to punishing offenders and compensating victims. However, it has not addressed the question of exactly who counts as an offender or, closely related, whether it is only offenders that the state should hold liable for punishment and compensation. This section addresses the question of exactly who must undergo burdens that would compensate victims and reform offenders. Basically, whom should be put on trial? This section advances a more collectivist answer than is common to encounter in the West. It might seem obvious whom should be punished and who should make compensation – surely, all and only the ones who (have been fairly judged to) have culpably offended or wrongfully caused harm. However, this answer belies complexity. Suppose that another party X had encouraged a person Y to commit the offense against, or wrongfully cause the harm to, Z. Then, at least depending on the sort of encouragement, it would be natural to hold X somewhat responsible for the crime and harm (even if not as much as the immediate or direct wrongdoer Y). X would either himself count as an offender or should be held responsible even if he does not count as that. As per the previous two sections, both X and Y should, specifically, be made to undergo burdens that compensate Z by improving Z’s quality of life in a way she accepts (although normally Y’s burdens should be greater than X’s). So, here is one compelling exception to the idea that only those directly responsible for a crime or wrongful harm should be held liable for it. Reflection on the African tradition provides strong reason to consider more exceptions than this one. Reconciliatory processes salient among indigenous sub-Saharan peoples have often involved, not merely the direct offender and the direct victim, but also the families of both (Murithi 2009: 228; Elechi et al. 2010; Mangena 2015: 6, 7) as well as many members of the broader community (Elechi et al. 2010: 82; Mangena 2015: 3; Masitera 2018: 114–115). African societies often consider families of victims and even the society as a whole to be secondary victims who are owed at least an apology from the offender (Murithi 2009: 227–228; Elechi et al. 2010: 78, 81; Mangena 2015: 11). In addition, and of particular relevance to the present discussion, the families of offenders can be held liable to apologize and make compensation to victims, especially when the direct offenders lack the means to compensate fully on their own (Elechi et al. 2010: 77–79). Still more, the families of offenders can be tasked with helping to reform the offenders’ character, say, by remonstrating with or educating them (Elechi et al. 2010: 74, 78–79). Why might it be reasonable to hold offenders’ family members liable to undergo the burdens of compensating offenders’ victims and reforming offenders’ character? Given the above principle that punishment ought to be deployed so as to make people clean up their own mess, why think the mess partly belongs to the family or at least some, presumably older, members of it?

388

Holding responsible in the African tradition

Traditionally, one answer from sub-Saharan cultures has been that simply sharing the blood of the offender renders one liable to some degree for his misdeeds (Mangena 2015: 7). However, there are three other answers available that will appeal more strongly to a multicultural, contemporary audience of moral philosophers and ethicists. One argument is that it is reasonable to hold a family responsible for the wrongdoing of one of its members insofar as the family has benefited from the wrong, even if it did not encourage it. Perhaps if some family members genuinely could not have known they were receiving, say, stolen goods, it would be wrong to hold them responsible for the theft. However, if they did know, or if they did not know but should have known, then there is prima facie reason to enlist their help in effecting compensation and reform. The next two rationales have a broader scope, in that, if sound, they entail that it can be right to hold a family responsible for the actions of one of its members even when neither it nor any of its members benefited from the wrongdoer’s crime. Consider the idea that none of us has libertarian free will that would render us solely responsible for our choices; we are open to the causal influence of others such that there is a “shared responsibility” for crime (Mangena 2015: 6). Furthermore, it is plausible to think, with much of the African tradition, that we have weighty obligations to help one another, especially our family members and in respect of their moral character. Where older family members have failed in this obligation, say, to a teenager or young adult, they owe something to the one who has gone astray and of course to those harmed by that. From this perspective, we are indeed our “brother’s keeper” (Mangena 2015: 10) and liable to some burdens of punishment and compensation when our brother has done wrong. There will of course be cases where a family did all it could in respect of a member’s moral education, but he ultimately decided of his own volition to engage in criminal behavior. Even so, perhaps it could make sense for a trial judge to hold the family responsible for helping to reform his character, if it is now in a better position to do so. The family retains a strong obligation to care for its members, including care for their character, and, so, even if the family did not fail a given member in the past, it might sensibly be required to help him in the present. The last major rationale for deeming family to be liable to help compensate and reform is illustrated by a recent event in southern Africa, where it appears that Lesotho adult male citizens living illegally in South Africa committed a horrific mass gang rape there. After the facts looked well established, the Lesotho prime minister apologized to the president of South Africa, asked for forgiveness from him, offered to pay for the costs of the trial, and promised to try to prevent more illegal intrusions (Sibanda 2022). Here, the Lesotho government had not benefited from the actions of its citizens in South Africa. In addition, it had presumably done what it could have to educate them morally when they were young, and in any event had little reach given their residence outside the country. Even so, the prime minister deemed it appropriate to hold his country responsible for the wrongs of its citizens done in another country. Many readers will find the prime minister’s reactions to have been appropriate, if not morally required. The best explanation of why that approach to holding responsible is reasonable is probably that Lesotho has long identified with these citizens and they have done so with it. The operative principle is something like this: the more one has shared a sense of togetherness and engaged in joint projects with people, and the more they have done so with one, the more one should feel guilt and shame when they behave wrongly and the more one is obligated to act in accordance with these emotions. Being closely affiliated

389

Thaddeus Metz

with a guilty party is sufficient to be stained, such that it is not merely family, but also an offender’s “well-wishers” (Elechi et al. 2010: 74, 82), who in principle can be liable in a reconciliatory trial. When others share a sense of self with someone who has done wrong, perhaps there is a legitimate sense in which it is partially their mess despite not having done it. This chapter has not provided a systematic defense of the claim that it is right to take such a “wide,” “inclusive,” or “collectivist” approach to holding people responsible for wrongdoing, but it has articulated three different rationales for doing so that merit serious consideration, at least jointly and perhaps individually, too. It is yet another important approach to holding responsible that comes from the African context, but is arguably fitting not just for it.4

Notes 1 Cf. Jenny Steele’s “Responsibility for Others” and David Wong’s “Responsibility in Confucian Thought,” elsewhere in this Handbook. 2 For other accounts of reconciliation in the African context, see Villa-Vicencio and Verwoerd (2000); du Bois and du Bois-Pedain (2008); Huyse and Salter (2008); Villa-Vicencio (2009). 3 Exceptions are to be expected, say, in cases of double jeopardy or where punishing an offender for a relatively minor crime would foreseeably cause serious harm to innocent parties such as his young children. 4 I am grateful to Max Kiener for written comments on a prior draft and to Max Kiener, Matt King, Elinor Mason, Leo Menges, David Owens, Daniel Telech, and Andrea Westlund for oral comments on a talk that I gave based on this material at a mini-workshop on the Routledge Handbook of Responsibility organised by Max Kiener at the University of Oxford.

Further reading South Africa’s Truth and Reconciliation Commission (TRC) is responsible for having brought some African ideas about reconciliation to an international audience. Tutu, D. (1999). No Future Without Forgiveness. New York: Random House, by the former Chair of the TRC, addresses its ethical and historical background, as does Murithi, T. (2009). “An African Perspective on Peace Education: Ubuntu Lessons in Reconciliation,” International Review of Education, 55, 221–233. Villa-Vicencio, C., & Verwoerd, W. (eds.). (2000). Looking Back, Reaching Forward: Reflections on the Truth and Reconciliation Commission of South Africa. Cape Town: University of Cape Town Press as well as Swart, M., & van Marle, K. (eds.). (2017). The Limits of Transition: The South African Truth and Reconciliation Commission Twenty Years On. Leiden: Brill are collections of critical discussions about the TRC’s rationale, functioning, and influence. Mangena, F. (2015). “Restorative Justice’s Deep Roots in Africa,” South African Journal of Philosophy, 34, 1–12 draws on examples from four countries to sympathetically illustrate how reconciliation has been salient in the African context, while Villa-Vicencio, C. (2009). Walk with Us and Listen: Political Reconciliation in Africa. Cape Town: University of Cape Town Press defends a certain reconciliatory approach to conflict informed by a variety of practices and ideals indigenous to Africa. The foreward to this book, written by Tutu, D., supports a forgiveness-based approach to reconciliation, as does Krog, A. (2008). “ ‘This Thing Called Reconciliation . . . . ’: Forgiveness as Part of an Interconnectedness-towards-Wholeness,” South African Journal of Philosophy, 27, 353–366 by sophisticated appeal to characteristically African values. In contrast, works that advance forms of punitive reconciliation include Bewaji, J. A. I. (2016). The Rule of Law and Governance in Indigenous Yoruba Society. Lanham, MD: Lexington Books, which draws on traditional practices common among a major Nigerian people, as well as Wringe, B. “Political Apologies, Punishment, and Reconciliation Without Forgiveness” and Metz, T. (2022). “Why Reconciliation Requires Punishment but Not Forgiveness,” which both appear in Satne, P.  & Scheiter, K. (eds.), Conflict and Resolution. Cham: Springer and are less hermeneutic and more analytic in respect to method.

390

Holding responsible in the African tradition

References Balogun, O. A. (2018). African Philosophy: Reflections on Yoruba Metaphysics and Jurisprudence. Lagos: Xcel Publishers. BBC News. (2016). “Thailand Drunk Drivers Face Morgue Work as Punishment,” April 12. www. bbc.com/news/world-asia-36025937. Bewaji, J. A. I. (2016). The Rule of Law and Governance in Indigenous Yoruba Society: A Study in African Philosophy of Law. Lanham, MD: Lexington Books. Boxill, B. (2013). “Compensatory Justice,” in H. LaFollette (ed.), The International Encyclopedia of Ethics (pp. 953–959). Malden, MA: Blackwell Publishing Ltd. Burton, M. (2000). “Making Moral Judgements,” in C. Villa-Vicencio & W. Verwoerd (eds.), Looking Back, Reaching Forward: Reflections on the Truth and Reconciliation Commission of South Africa (pp. 77–85). Cape Town: University of Cape Town Press. Constitutional Court of South Africa. (1996). Azanian Peoples Organization (AZAPO) and Others v President of the Republic of South Africa and Others, (CCT17/96) [1996] ZACC 16. www.saflii. org/za/cases/ZACC/1996/16.html. ———. (2006). Dikoko v Mokhatla, 2006 ZACC 10. www.saflii.org/za/cases/ZACC/2006/10.html. du Bois, F., & du Bois-Pedain, A. (eds.). (2008). Justice and Reconciliation in Post-Apartheid South Africa. Cambridge: Cambridge University Press. Duff, R. A. (2001). Punishment, Communication, and Community. New York: Oxford University Press. Elechi, O. O., Morris, S., & Schauer, E. (2010). “Restoring Justice (Ubuntu): An African Perspective,” International Criminal Justice Review, 20, 73–85. Feinberg, J. (1965). “The Expressive Function of Punishment,” The Monist, 49, 397–423. Goodin, R. (1989). “Theories of Compensation,” Oxford Journal of Legal Studies, 9, 56–75. ———. (1991). “Compensation and Redistribution,” in J. Chapman (ed.), Compensatory Justice; Nomos, vol. 33 (pp. 143–177). New York: New York University Press. Hampton, J. (1988). “The Retributive Idea,” in J. G. Murphy & J. Hampton (eds.), Forgiveness and Mercy (pp. 111–161). New York: Cambridge University Press. Huyse, L., & Salter, M. (eds.). (2008). Traditional Justice and Reconciliation After Violent Conflict: Learning from African Experiences. Stockholm: International Institute for Democracy and Electoral Assistance. Krog, A. (2008). “ ‘This Thing Called Reconciliation . . . . ’: Forgiveness as Part of an Interconnectedness-towards-Wholeness,” South African Journal of Philosophy, 27, 353–366. Lenta, P. (2007). “In Defence of AZAPO and Restorative Justice,” in W. le Roux & K. van Marle (eds.), Law, Memory and the Legacy of Apartheid: Ten Years After AZAPO v President of South Africa (pp. 149–182). Pretoria: Pretoria University Law Press. Malan, M.,  & Green, A. (2013). “Basson’s Lawyer: There Are No Victims,” Mail  & Guardian, April 19. https://mg.co.za/article/2013-04-19-00-bassons-lawyer-there-are-no-victims/. Mangena, F. (2015). “Restorative Justice’s Deep Roots in Africa,” South African Journal of Philosophy, 34, 1–12. Masitera, E. (2018). “Ubuntu Justice and the Power to Transform the Modern Zimbabwean Rehabilitation Justice System,” in E. Masitera & F. Sibanda (eds.), Power in Contemporary Zimbabwe (pp. 109–120). Abingdon: Routledge. Metz, T. (2015). “A Theory of National Reconciliation: Some Insights from Africa,” in C. Corradetti, N. Eisikovits, & J. Rotondi (eds.), Theorizing Transitional Justice (pp. 119–135). Surrey: Ashgate. ———. (2019). “Reconciliation as the Aim of a Criminal Trial: Ubuntu’s Implications for Sentencing,” Constitutional Court Review, 9, 113–134. ———. (2020). “Must Land Reform Benefit the Victims of Colonialism?” in E. Masitera (ed.), Philosophical Perspectives on Land Reform in Southern Africa (pp.  145–160). Cham: Palgrave Macmillan. ———. (2022a). “Why Reconciliation Requires Punishment but Not Forgiveness,” in K. Scheiter & P. Satne (eds.), Conflict  and Resolution: The Ethics of Forgiveness, Revenge, and Punishment (pp. 265–281). Cham: Springer. ———. (2022b). “A Reconciliation Theory of State Punishment: An Alternative to Protection and Retribution,” Royal Institute of Philosophy Supplement, 91, 119–139.

391

Thaddeus Metz Miller, D. (2017). “Justice,” in E. Zalta (ed.), Stanford Encyclopedia of Philosophy. https://plato. stanford.edu/entries/justice/#CorrVersDistJust. Murithi, T. (2009). “An African Perspective on Peace Education: Ubuntu Lessons in Reconciliation,” International Review of Education, 55, 221–233. Nozick, R. (1974). Anarchy, State, and Utopia. New York: Basic Books. Oyowe, O. A. (2017). “Ubuntu, Rectification and the Land Question,” in D. Bilchitz, T. Metz, & O. Oyowe (eds.), Jurisprudence in an African Context (pp. 234–239). Cape Town: Oxford University Press. Pather, R. (2018). “Ramaphosa Under Pressure to Support Reparations for Apartheid Victims,” Mail & Guardian, December 17. https://mg.co.za/article/2018-12-17-ramaphosa-under-pressureto-support-reparations-for-apartheid-victims. Republic of South Africa. (1993). “Constitution of the Republic of South Africa Act No. 200.” www. gov.za/documents/constitution/constitution-republic-south-africa-act-200-1993. ———. (1995). “Promotion of National Unity and Reconciliation Amendment Act No. 87.” www. saflii.org/za/legis/num_act/ponuaraa1995557/. Sibanda, A. (2022). “Prime Minister Apologises for Basotho Crimes in SA,” Sunday World, August 4. https://sundayworld.co.za/news/prime-minister-apologises-for-basotho-crimes-in-sa/. Truth and Reconciliation Commission of South Africa. (1998). “Truth and Reconciliation Commission of South Africa Report.” www.justice.gov.za/trc/report/. Tutu, D. (1999). No Future Without Forgiveness. New York: Random House. Villa-Vicencio, C. (2009). Walk with Us and Listen: Political Reconciliation in Africa. Cape Town: University of Cape Town Press. Villa-Vicencio, C., & Verwoerd, W. (eds.). (2000). Looking Back Reaching Forward: Reflections on the Truth and Reconciliation Commission of South Africa. Cape Town: University of Cape Town Press. Waldron, J. (1992). “Superseding Historic Injustice,” Ethics, 103, 4–28.

392

Section 8: The ethics and politics of responsibility

31 ARTIFICIAL INTELLIGENCE AND THE IMPERATIVE OF RESPONSIBILITY Reconceiving AI governance as social care Shannon Vallor and Bhargavi Ganesh Introduction1 The accelerating development of artificial intelligence (AI) systems generates acute and interlinked challenges for responsibility ascription and governance. While today’s AI tools lack the type of agency that can bear normative (e.g., moral or legal) responsibility, they are increasingly deployed in ways that fail to map onto existing norms for ascribing moral and legal responsibility to humans. This failure of AI actions to map to our responsibility practices is commonly referred to as the problem of ‘responsibility gaps’ (Matthias 2004). Responsibility gaps are perceived to arise when an AI system action has moral stakes for which we normally hold one another accountable; typically, one that causes harm or offense. If the action can easily be blamed on a human’s poor design choices or failures of oversight, no gap will arise  – we know who is at fault. But increasingly, AI systems can produce actions that no human could have specifically anticipated or countered, and which were selected for reasons not easily attributed to human agency. Here it becomes less clear who can justifiably bear moral blame and answer for the harm. This is particularly true if we think, as philosophers since Aristotle have often argued, that moral responsibility requires having adequate knowledge and control of an action, and the power to have acted in a better way. Since Matthias (2004) first identified the problem, some have simply denied that such gaps exist, since all AI actions are ultimately enabled by human actions (Johnson 2015; Himmelreich 2019). But many more have acknowledged that the causal chains between human and AI actions are growing so diffuse that these gaps, ‘real’ or not, are increasingly perceived, creating a genuine problem for AI governance and public trust (Danaher 2016; Köhler et al. 2018; Nyholm 2018; Santoni de Sio & Mecacci 2021; Tigard 2021). To understand why responsibility for AI system actions is becoming so challenging to anchor, it helps to know a bit about what ‘AI’ systems today actually are, and what they do. ‘AI’ remains a term of art rather than a scientific reality. While definitions of artificial intelligence are hotly contested, the concept is rooted in a scientific research program dating

395

DOI: 10.4324/9781003282242-43

Shannon Vallor and Bhargavi Ganesh

from the 1950s that pursued machine simulation of the cognitive powers of human beings. But today’s AI systems do not create responsibility gaps by being ‘intelligent’ individuals in the sense that humans or other creatures are. The scientific concept of AI has been remade as a commercially attractive label for software tools that solve certain kinds of computationally intensive problems like prediction, classification, and optimization. Such tasks can now be performed without remaking our intellectual faculties of understanding, awareness, and reflection in silicon. Instead, the new powers of AI technology are mindless marvels of probabilistic analysis, realized through the generation of immensely complex mathematical models that map – in very non-human like ways – the statistical relationships within large sets of discrete data, from words to pixels to wavelengths. Some ‘AI’ systems employ machine learning algorithms that enable adaptive or ‘selflearning’ capabilities and new levels of autonomous functioning. While most AI systems are not adaptive or autonomous, that is, they don’t update their own internal model or independently take action in the world, an AI system that is built to exploit self-learning and autonomous capabilities can change itself and operate itself (within limits set by its human designers). These advanced capabilities can in principle yield many scientific, economic and social benefits; for example, self-driving vehicles that improve road safety, robotic surgery tools that reduce human error, and automated routing of shipping to optimize supply chain efficiency. However, such capabilities create especially thorny responsibility challenges. The mathematical complexities of today’s ‘deep learning’ AI models routinely impede human inspection of their decision chains, while adding an element of stochastic variability that makes their individual outputs unpredictable even when overall performance is stable and meets our expectations. Additionally, the speed and scale of operations enabled by AI can make real-time human oversight and steering of system actions practically infeasible. These AI features create gaps in traditional responsibility practices that rely upon responsible humans knowing what they are doing as they do it, and being able to exercise adequate control over that doing. The gaps appear in both backward-looking and forwardlooking frames: backward when we struggle to confidently assign culpability for harmful actions in which an AI system has already played a significant causal role, and forward when it is unclear how or whom to assign future obligations to mitigate the risks of AI, particularly where there is significant uncertainty about how a system will behave in a given environment, or the harms it might cause. It remains a matter of debate whether AI creates genuine absences of responsibility, or merely the appearance of absence (Himmelreich 2019; Tigard 2021; Tollon 2022). Yet the answer is largely moot, given that social trust in AI-driven actions and outcomes is at stake in either case. Social trust is threatened by any new form of agential power that evades the governing counterforce of effective responsibility practices – especially if this new, ungoverned power is growing. In what follows, we analyse our growing vulnerabilities to the new modes of power enabled by AI systems as a call to construct new practices of responsibility, embodied in AI governance as a creative act of social care. We draw upon Jonas’ (1984) account of responsibility as a caring response to others’ vulnerability to our power, as well as Vargas’ (2013) and McGeer’s (2015) instrumental account of moral responsibility as iterative and constructed. We conclude with some implications for AI governance and regulation, drawing insights from past regulatory efforts in steamboat engineering, consumer finance, and environmental health. 396

Artificial intelligence and the imperative of responsibility

1  Agential power of AI systems Today’s AI software tools for prediction, classification and automation enable new forms of agential power. What is agential power? We mean a power driven by the phenomenon of intentionality; a power that can be deliberately and non-automatically initiated, foregone, altered, steered, or terminated.2 Agential power stands in contrast to purely mechanical powers. A tidal wave unleashed by an earthquake carries tremendous kinetic power in the strictly causal sense; but it carries no agential power. While we might loosely call the wave an ‘agent’ of destruction, its power is driven by no intentionality of any kind. The same is true of meteor strikes, volcanic eruptions and solar storms. The clearest examples of agential power are voluntary, intentional acts undertaken by adult humans as a result of choice: stealing butter from a shop, or stopping a car to let someone cross the street. These stand apart from human acts involving no agential power, like the knee-jerk reflex produced by a physician’s hammer (the physician’s tap is agential.) Yet agential power need not be actively intentional; the relation to intentions can be tacit or counterfactual. For example, an unthinking harm (you sharply elbowing someone in the ribs as you mindlessly rush by them in a crowd) may not be your active intention, but it remains an exercise of your agential power insofar as a) the harm is partly explained by your standing quality of will, in this case a state of disregard for strangers’ physical boundaries and comfort; and/or b) you counterfactually could have enacted deliberate care in not elbowing people. Between such clear-cut cases of agential power fall many marginal or ambiguous cases; the actions of very young children, for example, or adults suffering severe intoxication or brain injury. Intelligent behaviors of nonhuman animals, and groups, form further classes of action that often appear robustly agential while remaining philosophically and scientifically contentious with respect to the nature of this agency. Why is agential power relevant for responsibility? Moral responsibility (and to a weaker extent, legal responsibility) is conventionally assumed to track agential power. Where such power is seen to be exercised by a moral agent, we typically hold that agent responsible for the act (or omission) in question. Yet the phrase ‘agential power’ is not commonly found in the literature on moral responsibility. Instead, philosophers tend to employ the atomistic concept of the ‘moral agent’. This agent-centred focus obscures the relational dimensions of the responsibility concept embedded in our social practices of holding others responsible. We argue that entities that are clearly not moral agents can nevertheless enact agential powers, and in doing so enter into moral relations that can and must be governed by responsibility practices. For example, an agent-centred focus fosters confusion about cases where agential power is enacted by a composite or distributed agent, such as a corporation. Much of the philosophical literature on the responsibility challenges presented by modern corporations is caught in the trap of trying to map their composite agency onto the concept of individual moral agency (Björnsson & Hess 2017). Fortunately, in the interest of responding to urgent moral and legal concerns about corporate power, we can employ the concept of agential powers rather than moral agency. Corporations are not moral agents, but nor are they mechanically compelled; they are not like the tidal wave. Nor are they, like the patient responding to the knee-tap, unable to modulate their responses to stimuli. Corporations can strategically select and steer their own actions. Their status as a font of agential power is not diminished even if a corporate 397

Shannon Vallor and Bhargavi Ganesh

will distributed among many hands lacks the internal coherence required to make sense of the corporation as a moral agent. And if responsibility practices can effectively govern agential power even when it does not flow directly from an identifiable moral agent, then we don’t need to prove the existence of corporate moral agency in order to demand corporate responsibility. Corporations must fit into our responsibility practices, but differently than individuals do. To put it another way: while the conduct of corporations is rightly subject to the modes of shared moral governance constituted by responsibility practices, the style of governance must be adapted to their unique kind of power: agential without being moral agents. Similarly, a focus on the agential power of AI systems allows us to see the need to incorporate AI into our practices of responsibility, regardless of the indeterminate status, or absence of, AI’s moral agency. This is important because of the broad scientific consensus that these systems, which lack sentience, common sense, and moral experience, are not personal agents that can bear moral or legal responsibility, notwithstanding their power to occasionally create the illusion of robust agency (Johnson 2022). But why should we accept the premise that AI systems, which lack moral agency, nevertheless have agential power? After all, in the software running on a given hardware configuration, we will find no agential power. The code can initiate or terminate a computational process given a certain data input, but this is not intentional or deliberate behavior; not even counterfactually. When the ‘system’ is defined to include only mechanical components, its agency remains of the merely causal kind we ascribe to an earthquake. It has no intentional states. However, as many have noted (Gray & Suri 2019; Crawford & Joler 2018), building on Deborah G. Johnson’s account of sociotechnical systems (2011), such a narrow definition fails to acknowledge that AI systems in the full sense are sociotechnical systems that execute designed objectives. And design entails intentional states. When taken as a complete sociotechnical system, the agential powers of AI systems appear; AI systems are thus far more similar to corporations than tidal waves, because their activity (designed, built, and deployed by humans) is initiated, steered, and terminated by factors that are intentionally driven, not unchosen or compelled. Of course, so are the powers of a conventional automobile, or a traffic light, yet we don’t take these to create deep problems for responsibility. The conventional automobile is under the direct control of a responsible driver, and any non-driver-initiated behaviors are either non-agential (caused by mechanical forces like wind), or result from willed acts or omissions of others, such as a negligent repair person. For automated systems like a traffic light, even the lack of a real-time controller is no grave obstacle to responsibility ascription. If it fails and a crash results, there are stable norms for assigning culpability to humans involved in its design, installation and maintenance. Marginal edge cases arise, but because the agential power of these artifacts is transparently derived from human agents who contribute whole intentions to the system’s operation, responsibility ascription is typically a tractable problem. Responsibility ascriptions for certain types of AI actions are more challenging. This is because sophisticated AI systems can effect a split of the intentional content of the action between the human and machine components of the system, leaving neither human nor machine in possession of all the essential elements of an intentional act. In a task like autonomous driving, the perception, classification, means-ends deliberation and strategic reasoning might all take place on the machine side of the system, obscured from human view. Yet the task’s meaning and context, the value of the goal, the interests at stake, and the morally 398

Artificial intelligence and the imperative of responsibility

salient consequences of a system act or policy remain entirely inaccessible to the machine, and embodied only on the human side of the system. This makes responsibility ascription a problem – while the driving system as a whole is clearly agential, neither the machine nor the human can form a complete intention corresponding to a given driving action. The ‘hybrid’ nature of AI action, split across machine mechanisms and human minds, bears some similarity to problems of corporate agency (Hakli & Mäkelä 2019). Both corporations and AI systems can form policies and initiate actions that are not reducible to or mirrored by the intentions of their individual human components. As such, both modern corporations and AI systems lie in the metaphysical lineage of families, communities, religious institutions and political states. AI systems represent a new branch of agential power on this tree. All entities in this lineage exercise their power through structured assemblages that blend multiple human agencies with artifacts that partially embody and enact components of intention (laws, rituals, policies, objects, codes, and processes, including algorithmic ones). However, AI systems automate far more of the intentional processes that in human agents play a central role in the selection (or termination) of an action. For example: previously, a state or corporation could use computational processes to gather surveillance data, but individual humans had to intentionally handcraft and apply selection criteria and action rules to use that data to curtail anyone’s freedoms. Today, AI systems can in principle be given a broad optimization task for a surveilled environment, such as ‘reduce ambient noise in the gallery’, and enabled to produce their own selection criteria and action rules for achieving that goal. Such a system might, for example, ‘learn’ to prohibit entry to a public gallery by children, or those in motorized wheelchairs, or those with a chronic cough. (Or, if the task parameters are under-specified, it might kill the power and lock people out of the gallery – an empty, dark gallery is very quiet!) By automating and altering the means by which intentions are executed, AI complicates traditional responsibility practices such as assignment, blame and praise. If a gallery AI system bars wheelchair users, it’s clearly a human design failure. The algorithm’s policy cannot be fully understood without reference to both human intentions (an aesthetic selection for quiet) and omissions of human concern (for disabled persons whose interests counterfactually could have been anticipated and judged more important than quiet). Yet no human wanted wheelchair users to be barred from the gallery! How then, do we assign blame or praise, which requires judging the quality and content of the will contributing to a specific action (Strawson 1962; McKenna 2012)? To know whether praise or blame is warranted, I must know if the actor’s will was good or bad. But how do I judge the moral quality of will behind an act of unjust discrimination, if that will was divided across the cognitive powers of a human and those of a non-sentient machine? Consider the even denser tangles of agential power embedded in an inverse reinforcement learning AI system designed by humans to derive its objective function (the formal definition of the task goal) from its real-world observations of other humans’ performances, adapting its own behavior over time to align with the inferred values and preferences of observed humans. Now consider a variant in which the AI system must choose whether to follow a driver’s current instructions (‘speed up!’) or its interpretation of their previously observed preferences (‘let children cross the road in good time’) (Milli et  al. 2017). Or consider a ‘swarm’ algorithm designed to adjust its behavior to align with the behaviors of thousands of other algorithmic agents engaged in the same task, while also adjusting to external changes in the task environment. These diffuse tangles of intention and causality 399

Shannon Vallor and Bhargavi Ganesh

make responsibility assignment, blame, and praise profound practical challenges, even if in principle some set of humans are morally responsible for any AI system action. Moreover, AI systems make possible new and more powerful types of action, creating new human vulnerabilities for which we lack shared responsibility scripts, norms, or intuitions. Previously, to be vulnerable to most surveillance harms one had to be targeted, noticed, and selected, by someone. Even with troves of publicly accessible Internet data about me, I could remain obscure unless someone were motivated to search for my data and analyze it. Today, I can be superficially indistinguishable from a million others, observed by literally no one, and still be flagged, restricted, or penalized by a ‘smart’ algorithm for who I am, who I know, what I earn, where I was born, or what I say online. Here, I am exposed in a radical new way – for algorithmic automation can erase the shelter of public obscurity (Selinger & Hartzog 2017) that has long been a protective side-effect of the cognitive bottleneck in human surveillance data analysis and decision-making. We are beyond Foucault’s Panopticon, subject to a new kind of power: networked AI systems that can in principle collect, analyze and act on everything. AI systems can also see and act upon what humans cannot, creating new asymmetries of power. For example, AI systems often find proxies for sensitive characteristics like race in unlabeled data, even where humans cannot (Adam 2022). AI can thus act on our worst biases in ways we could not if we tried, opening new paths to punishing the already vulnerable or marginalized. With no safe harbor from scrutiny or hope of obscurity, the difference in vulnerability is not simply of degree, but of kind. In the next section, we examine how our vulnerability to these new agential powers highlights the nature of responsibility as relational, pointing to a practice of social care as a remedy for AI responsibility gaps.

2  Responsibility as social care in the use of agential power AI responsibility gaps may seem intractable if we, like many philosophers, treat moral responsibility as a rigid concept defined by certain objective qualities of individual moral agents. For if the cognitive elements targeted by conventional responsibility ascriptions are increasingly smeared across discrete machine and human components of AI systems, how can we possibly put ‘responsibility’ back together? Fortunately, there’s another way to think about moral responsibility, one developed in Hans Jonas’ The Imperative of Responsibility: In Search of an Ethics for the Technological Age (1984). In this landmark work, Jonas describes our unprecedented responsibilities for the fate of nature and future generations, responsibilities that emerged only when 20th century technologies made us capable of planetary destruction. Jonas notes the utter incapacity of traditional moral theories to competently address these new risks and their attendant responsibilities, which were literally inconceivable only a century ago. What’s important for our purposes is that the obligating force of moral responsibility is seen by Jonas to arise from the meeting of agential power with another subject’s vulnerability: The “what for” [of being responsible] lies outside me, but in the effective range of my power, in need of it, or threatened by it. It confronts this power of mine with its right-to-be and, through the moral will, enlists it for itself. The matter becomes mine because the power is mine. (92) 400

Artificial intelligence and the imperative of responsibility

Jonas goes on to describe the experience of the responsible agent, “the subject who, in virtue of his power, is called to its [the vulnerable subject’s] care.” Agential power creates the caring duty, and the “exercise of the power with disregard of the obligation is, then, ‘irresponsible,’ that is, a breach of the trust-relation of responsibility” (93). What is of crucial importance for Jonas is that the shape of this moral responsibility is materially conditioned on the type and extent of the power, and its encroachment upon the vulnerable: responsibility is a correlate of power, so that the scope and kind of power determine the scope and kind of responsibility. When power and its constant exercise grow to certain dimensions, then not only the magnitude but also the qualitative nature of responsibility changes. (128) Like nuclear weapons, AI embodies new dimensions of agential power and vulnerability. This calls for a new kind of responsibility to meet it. Jonas’ account provides a criterion by which to assess the suitability of new responsibility practices: they must enact and fulfill the obligation to take care for that which our new powers makes vulnerable. Responsibility safeguards the integrity of that which calls out from a place of danger. Jonas’ account of responsibility is not further developed, however, as the rest of the work explores eschatological concerns. Fortunately, there are other philosophical threads that we can pull on. One prospect is to link up Jonas’ account of technological responsibility as social care with feminist care ethics (Held 1993). In this tradition, too, moral responsibility is not a brute fact about individual agents, but a kind of relation: a caring response to another’s vulnerability. ‘Being responsible’ establishes to others our worthiness to be trusted with our own power, our commitment to taking care for them in the act of power’s use. In demonstrating that care, I take part in a social practice of trust-making and trust-maintaining, one that must continually be negotiated and reconstructed as our powers and vulnerabilities to one another shift. Rather than say that responsibility is something, we should thus say that we make responsibility something, in order that it do something. While much of the philosophical literature still takes the ‘internalist’ view that responsibility lies within qualities of the individual agent, others have more recently endorsed the latter, ‘instrumentalist’ view. Victoria McGeer (2015) and Manuel Vargas (2013) each offer naturalized accounts of moral responsibility as a relational phenomenon, constituted by social practices and norms that deliver the desirable or socially necessary outcomes which justify these practices. On this view, responsibility practices are not primarily justified by individuals getting what they ‘deserve’, but are rather about building and sustaining a certain kind of ‘moral ecology’ (Vargas 2013: 246). That ecology is enabled by the cultivation of responsible moral agents; we are justified in being, and holding others, responsible because that is how this moral ecology endures. This moral ecology is itself a justifiable end, because it alone makes social trust possible and rational despite our profound vulnerability to one another. Human society is not sustainable in the total absence of social trust; we require social bonds to survive. This way of framing responsibility is uniquely suited to the challenge of AI responsibility gaps. For one thing, AI systems today, and for the foreseeable future, will not satisfy the internalist’s criteria. AI systems lack the reasons–responsiveness and sensitivity to moral considerations that both internalists and instrumentalists agree are characteristic of morally 401

Shannon Vallor and Bhargavi Ganesh

responsible individuals. On the internalist view, then, AI systems are ‘outsiders’ to responsibility practices, since those practices are only meaningful in their application to moral agents. But this makes the AI responsibility gap a permanent, intractable problem! More helpful is the instrumentalist view, on which AI systems can be novel participants in our evolving moral ecology of responsibility, without being responsible agents. As parts of this moral ecology that enact agential power, AI systems rightly come under effective governance by responsibility practices, since any part of the ecology can justifiably be governed to promote the ecology’s purpose: sustaining social trust and cooperation among beings vulnerable to one another’s agential power, through the exercise of social care in its use. The instrumentalist view encourages us to be creative and constructive in the development of new responsibility practices. These can be explicitly designed to help AI play a healthier role in the moral ecology, and to help us remain responsible agents despite their novel presence mediating our decisions and actions. As McGeer (2015: 273) observes, responsible human agency is always a work in progress. Responsibility practices already shape agents who don’t perfectly meet the conditions for knowledge and control of our actions, which is to say, all of us. They shape us to learn to be more responsible than we would otherwise be. So, AI further muddying the waters of our agency does not forestall the work of cultivating responsible agency; it makes it all the more urgent. This dictates the proper aim of governing AI systems: ensuring that their use promotes the continued cultivation of responsible agency in us, sustaining the broader moral ecology of social care and trust. We suggest this entails two primary tasks for AI governance: 1) ensuring that AI systems do not diminish the incentives for, or present new obstacles to, the cultivation of responsible human agency; and 2) ensuring that the agential powers these systems enact do not exploit our vulnerabilities in ways that undermine social care and trust. This account steers between the two dangers that traditional accounts of responsibility would force us to choose from: the Scylla of treating AI systems as if they were moral agents, or the Charybidis of leaving AI systems awkwardly outside the sphere of moral responsibility even as they mediate our own execution of it. Our account allows us to create an appropriate place for AI in our moral ecology of responsibility.

3  Constructing AI governance as social care: lessons from history Such construction of responsibility through technology governance is hardly a new task. Lessons for AI governance can be drawn from histories of steam power engineering, public health, and consumer finance that demonstrate the ability of governments and institutions to creatively exercise social care for our vulnerability to the power of new technologies. In the 1800s, steam power generation was an exciting new innovation. Steamboat travel, however, quickly became deadly, as high-pressure steam boilers exploded, causing thousands of casualties. The gory accidents, which displayed the vulnerability of both crews and passengers to this new technological power, made governing steamboats a popular political platform, with support from members of the press, citizens, and advocacy groups. In response to these calls to action, the US government undertook an iterative process between 1838 and 1871 to improve the safety of steamboats, revising and constructing new regulatory tools when initial efforts proved inadequate. Along the way a range of methods was used to stabilize a new moral ecology of maritime safety. Information gathering and safety device testing led to the development of 402

Artificial intelligence and the imperative of responsibility

engineering standards. Novel licensing requirements for pilots and engineers were combined with stricter liability standards for both steamboat operators and owners. Finally, the Act of 1871 created the federal Steamboat Inspection Service and with it a comprehensive marine safety code. Through this progressive regulatory effort, the government created new kinds of responsible agents (such as federal inspectors) as well as new governing forces (standards, requirements, devices, codes and penalties) to curb the destructive force of the steam boiler in the face of human vulnerabilities, rebuilding social trust in the exercise of this new power. The UK government faced a similar challenge after the Great Smog of 1952. Coal-fired power stations and automobiles, while exciting innovations at the time, produced an excess amount of pollution, shrouding London’s sky in a thick layer of smoke exacerbated by a local climate inversion. The unprecedented event caused thousands of casualties – the worst air pollution event in the UK’s history. In response, Parliament passed the Clean Air Act of 1956, building upon the foundations of the 1936 Public Health Act. The 1956 Act stipulated that smokeless fuels be used in both domestic and industrial contexts, funded research on air pollution, and incentivized the first national surveys of air quality (Brimblecombe 2006). Of course, air pollution was only partially and imperfectly addressed by these measures. Even so, the government formed a new moral ecology of environmental health, by investing in novel efforts to inform the public about the relationship between pollution, industrial and domestic choices, and public health vulnerabilities. As Brimblecombe notes, the Act’s greatest legacy was that ‘it created a belief that a better environment was possible and worthwhile despite the fact that at times it would restrict our personal freedom’ (2006: 313). It transformed the view of air pollution from a passive “inevitability” to an active sociotechnical force, bringing to light technology’s agential power that gave people new responsibilities to constrain that power for the sake of others, through personal choices, industrial compliance and public policy. The 2008 mortgage crisis in the US, which plunged millions into foreclosure and financial ruin, offers a more recent and cautionary example of the fragile political task of stabilizing a moral ecology from shocks of unregulated innovation. One factor behind the crisis was mortgage securitization, a financial innovation from the 1960s that had previously been lauded for its revolutionary role in freeing up funds for banks to lend to consumers. While it is disputed whether securitization was directly responsible for the crisis, researchers agree that this innovation generated additional risk in an already speculative environment, in which vulnerable households were encouraged to take on more debt than they could afford (Been et al. 2011). Further exacerbating the risks were fraudulent business practices on the part of credit rating agencies and mortgage insurers. In 2010, the US government passed the Dodd-Frank Act, creating the Consumer Financial Protection Bureau: the first US government agency tasked with protecting vulnerable consumers in the financial marketplace from deceptive and exploitative practices. The regulation also created enhanced reporting and testing requirements, prohibited certain types of risky trading activities, and increased oversight of mortgage insurance companies and credit rating agencies. In 2018 the Trump-led Congress rolled back key provisions of the act, while Trump’s appointee to the Consumer Financial Protection Bureau sought to defang the regulator (Klein 2018). This reminds us that establishing a moral ecology of responsible innovation is not a technocratic or legalistic exercise, but a collective political and moral task of social care that is never finished. 403

Shannon Vallor and Bhargavi Ganesh

AI is one more innovation that amplifies and transforms our agential power while producing harmful outcomes that disproportionately endanger the most vulnerable groups, leading to calls for government intervention. While some impacts of AI merely add to existing harms, AI does present novel governance challenges. Because AI is increasingly ‘general purpose’ technology used across many domains and regions, impacted communities are far more diverse and dispersed than in the previous cases. Machine learning algorithms also increasingly generate individualized content, creating more individualized harms. This makes it harder for activist/political movements to mobilize around shared disasters in the same way as steamboat accidents, smog events, and foreclosures. It also makes it more challenging to design regulatory interventions that can effect harm reduction at sufficient scales to protect those vulnerable. In each of the policy examples highlighted previously, the iterative nature of policymaking contributed to the negotiation and reconstruction of trust-making and trust-maintaining. In these examples, we also see the lasting impact of policy failures on that trust. For example, mistrust in financial institutions stemming from the US foreclosure crisis has been theorized as being at least partially responsible for low millennial homeownership rates. Similarly, public health campaigns such as vaccination efforts have struggled to earn back the trust of minoritized groups who were experimented on and mistreated by healthcare institutions in the past. If we would like the benefits of AI to be realized, we must treasure public trust, rather than trifle with it. Once it is lost, trust does not magically reappear; in some cases it may never be re-earned. We acknowledge, then, the limitations and fragilities of regulatory efforts to stabilize moral ecologies under pressure from new sociotechnical powers. Yet the instrumentalists’ focus on the construction of responsible agency helps us reframe the task of governance so that we see these not as legacies of failure or futility, but as iterative experimental works to support and nurture our (always imperfect) capacity for responsible agency. Rather than counsels of despair, we take the challenges for AI governance as encouragement, calling for greater adaptability and creative ambition in governance as a perpetual task of social care for our imperfect human agency and mutual vulnerability. In conclusion, the responsibility gap challenge is, on our view, not about locating AI’s ‘missing agency’. Our moral responsibility practices always confront missing, imperfect, and broken agencies. The real gap is a governance gap – a challenge to society to exercise appropriate care for those members of our moral community who are vulnerable to AI’s novel agential powers. Even where responsibility assignment and blame are challenged by AI automation, effective AI governance can enact responsible constraint of these systems’ agential powers. One strategy involves restrictive constraints (for example, prohibitions on AI use or features). The other strategy involves enabling constraints that mandate responsible, trustworthy human use of these systems’ power. The latter has the advantage of preserving AI’s potential benefits, but requires incentivizing social care by creating new positive duties, roles and liabilities of human agents (developers, users, and regulators) who can answer to vulnerable others for uses of this power.3 We urgently need an ethic of social care in the design of government interventions, that is, an ethic of joint attentiveness and responsiveness to the widening human vulnerabilities to AI now being exposed by civil society, activists, scholars, and AI professionals. An ethic of social care moves the goal of AI regulation and governance beyond risk management or even individual harm reduction, to the more ambitious aim of establishing and maintaining a stable and healthy moral ecology for responsible and trustworthy AI innovation. In that new 404

Artificial intelligence and the imperative of responsibility

moral ecology, which has yet to coalesce, the agential powers of AI systems will not weaken or replace, but support and extend our efforts to become responsible to, and for, one another.

Notes 1 This research was supported by the UKRI Engineering and Physical Sciences Research Council (grants EP/V026607/1 and EP/W011654/1) and Arts and Humanities Research Council (grant AH/ X007146/1). Stuart Anderson and Tillmann Vierkant provided detailed comments and insights on earlier drafts. 2 Our notion of agential power is broader than those in Gilabert (2018) and Lazar (2022), which involve political power over another. Ours also covers, for example, the agential power of a person on a deserted island who enacts intentionality in constructing a shelter, or carving on cave walls. 3 Here we follow Coeckelbergh (2020) and Tigard (2021) on the need to enhance our answerability for AI systems.

Further reading For a rich and comprehensive exploration of the literature on responsibility gaps and artificial intelligence, see Nyholm, S. (2020). Humans and Robots: Ethics, Agency and Anthropomorphism. London: Rowman and Littlefield. Jonas’, H. (1984). The Imperative of Responsibility: In Search of an Ethics for The Technological Age. Chicago: University of Chicago Press is a classic and foundational text in the ethics of technology that offers a philosophically vital and urgent perspective on responsibility and what is now commonly called ‘existential risk’; while its conclusions are challenging and rightly controversial, it is a call to action too long neglected. Care ethics provides another lens on responsibility that is too often overlooked in the literature on responsibility gaps and AI ethics; see Held, V. (1993). The Ethics of Care: Personal, Political, and Global. Oxford: Oxford University Press, and Tronto, J. C. (1993). Moral Boundaries: A Political Argument for an Ethic of Care. New York: Routledge.

References Adam, H., Yang, M. Y., Cato, K., Baldini, I., Senteio, C., Celi, L. A., Zeng, J., Singh, M., & Ghassemi, M. (2022). “Write It Like You See It: Detectable Differences in Clinical Notes by Race Lead to Differential Model Recommendations,” in Proceedings of the 2022 AAAI/ACM Conference on AI, Ethics, and Society (AIES ’22) (pp. 7–21). New York: Association for Computing Machinery. https://doi.org/10.1145/3514094.3534203. Been, V., Chan, S., Ellen, I. G., & Madar, J. R. (2011). “Decoding the Foreclosure Crisis: Causes, Responses, and Consequences,” Journal of Policy Analysis and Management, 30(2), 388–396. Björnsson, G., & Hess, K. (2017). “Corporate Crocodile Tears? On the Reactive Attitudes of Corporate Agents,” Philosophy and Phenomenological Research, 94, 273–298. Brimblecombe, P. (2006). “The Clean Air Act After 50 Years,” Weather, 61(11), 311–314. Coeckelbergh, M. (2020). “Artificial Intelligence, Responsibility Attribution, and a Relational Justification of Explainability,” Science and Engineering Ethics, 26, 2051–2068. Crawford, K., & Joler, V. (2018). “Anatomy of an AI System.” https://anatomyof.ai/. Danaher, J. (2016). “Robots, Law and the Retribution Gap,” Ethics and Information Technology, 18(4), 299–309. Gilabert, P. (2018). “A Broad Definition of Agential Power,” Journal of Political Power, 11(1), 79–92. Gray, M. L., & Suri, S. (2019). Ghost Work: How to Stop Silicon Valley from Building a New Global Underclass. New York: Houghton Mifflin Harcourt. Hakli, R., & Mäkelä, P. (2019). “Moral Responsibility of Robots and Hybrid Agents,” The Monist, 102(2), 259–275. Held, V. (1993). The Ethics of Care: Personal, Political, and Global. Oxford: Oxford University Press. Himmelreich, J. (2019). “Responsibility for Killer Robots,” Ethical Theory and Moral Practice, 22(3), 731–747.

405

Shannon Vallor and Bhargavi Ganesh Johnson, D. G. (2011). “Software Agents, Anticipatory Ethics, and Accountability,” in G. Marchant, B. Allenby, & J. Herkert (eds.), The Growing Gap Between Emerging Technologies and LegalEthical Oversight. The International Library of Ethics, Law and Technology, vol. 7. Dordrecht: Springer. Johnson, D. G. (2015). “Technology with No Human Responsibility?” Journal of Business Ethics, 127, 707–715. Johnson, K. (2022). “LaMDA and the Sentient AI Trap,” Wired, June 14, 2022. www.wired.com/ story/lamda-sentient-ai-bias-google-blake-lemoine/. Jonas, H. (1984). The Imperative of Responsibility: In Search of an Ethics for The Technological Age. Chicago: University of Chicago Press. Klein, A. (2018). “No, Dodd-Frank Was Neither Repealed nor Gutted: Here’s What Really Happened,” The Brookings Center on Regulations and Markets, May 25, 2018. www.brookings.edu/ research/no-dodd-frank-was-neither-repealed-nor-gutted-heres-what-really-happened/. Köhler, S., Roughley, N., & Sauer, H. (2018). “Technologically Blurred Accountability,” in C. Ulbert, P. Finkenbusch, E. Sondermann, & T. Diebel (eds.), Moral Agency and the Politics of Responsibility (pp. 51–68). New York: Routledge. Lazar, S. (2022). “Power and AI: Nature and Justification,” in J. Bullock et al. (eds.), The Oxford Handbook of AI Governance. Online May 2022. DOI: 10.1093/oxfordhb/9780197579329.013.12. Matthias, A. (2004). “The Responsibility Gap: Ascribing Responsibility for the Actions of Learning Automata,” Ethics and Information Technology, 6, 175–183. McGeer, V. (2015). “Building a Better Theory of Responsibility,” Philosophical Studies, 172(10), 2635–2649. McKenna, M. (2012). Conversation and Responsibility. Oxford: Oxford University Press. Milli, S., Hadfield-Menell, D., Dragan, A.,  & Russell, S. (2017). “Should robots be obedient?” in Proceedings of the 26th International Joint Conference on Artificial Intelligence (IJCAI'17) (pp. 4754–4760). AAAI Press. Nyholm, S. (2018). “Attributing Agency to Automated Systems: Reflections on Human-Robot Collaborations and Responsibility-Loci,” Science and Engineering Ethics, 24(4), 1201–1219. Santoni de Sio, F., & Mecacci, G. (2021). “Four Responsibility Gaps with Artificial Intelligence: Why They Matter and How to Address Them,” Philosophy & Technology, 34(4), 1057–1084. Selinger, E., & Hartzog, W. (2017). “Obscurity and Privacy,” in J. C. Pitt & A. Shew (eds.), Spaces for the Future: A Companion to Philosophy and Technology (pp. 119–129). New York: Routledge. Strawson, P. (1962). “Freedom and Resentment,” Proceedings of the British Academy, 48, 187–211. Tigard, D. W. (2021). “There Is No Techno-Responsibility Gap,” Philosophy and Technology, 34, 589–607. Tollon, F. (2022). “Responsibility Gaps and the Reactive Attitudes,” AI and Ethics. https://doi. org/10.1007/s43681-022-00172-6. Vargas, M. (2013). Building Better Beings: A Theory of Responsibility. New York: Oxford University Press.

406

32 MORAL RESPONSIBILITY FOR HISTORICAL INJUSTICE Michael Schefczyk

1 Introduction According to prevalent usage in the philosophical literature, historical injustices (HI) are large-scale wrongs in the distant past. It is generally acknowledged that agents who are morally responsible for injustices are obliged to perform acts of rectification vis-à-vis the wronged parties. Rectification consists, among other things, in apologies or measures to correct the effects of wrongdoing. In the case of HI, however, those who have committed the wrong, and those who endured it, are dead.1 How does the death of perpetrators and victims affect the moral situation? Do rectificatory obligations simply perish with their demise? Or do some members of later-born generations have to take responsibility for past wrongdoing, while other later-born people have corresponding historical claims? Under which conditions do such historical claims and obligations exist,2 if they exist at all, and in what precisely do they consist? These are core questions of the philosophical literature on HI. As this chapter deals with moral responsibility, it approaches the topic of HI from an uncommon, but instructive angle. The angle is uncommon insofar as it will give more than usual attention to issues crucial for the moral responsibility of historical agents  – namely, citizens’ culpability and moral awareness;3 the angle is instructive, however, insofar as moral responsibility is arguably the primary grounding of remedial responsibility in standard cases.4 It seems natural to assume that something similar holds in the context of HI. I call remedial first-hand responsibility any responsibility for HI which is predicated on moral responsibility; and I call second-hand responsibility any remedial responsibility which is predicated on morally relevant relations to culpable individuals or collectives. The chapter proceeds as follows: Section 2 gives a highly condensed overview of topics and trends in the HI literature. I distinguish between interactional and structural approaches to HI, opt for a comprehensive understanding making room for both and critically assess, very briefly, reductionist attempts to bring together demands of reparative and distributive justice.5 Section  3 proposes a standard procedure to examine remedial responsibility for HI. It reports on three models linking the culpability of states and (parts of) the population; on this basis, it shows the effect of different constellations of first-hand responsibility 407

DOI: 10.4324/9781003282242-44

Michael Schefczyk

of states and populations on second-hand responsibility. In the final part of the chapter, section 4, I examine the cultural ignorance excuse for HI and contend that a subjective element of “normative taste” for more or less forgiving standards of moral competence seems unavoidable.

2  (Very brief) overview One may distinguish between a “first wave” and a “second wave” of HI literature. The understanding of injustice in the first wave is interactional. First wavers conceptualize HI as (complex) events with sufficiently definite temporal boundaries. The account is interactional insofar as it refers to unjust acts that are committed by (a complex network of) perpetrators and done to (a group of) victims. Examples for such interactional injustices are the Armenian genocide, the Shoa, European colonialism, US slavery, sexual slavery in the Imperial Japanese Army, the predation of the Benin Bronzes, breaches of the Treaty of Waitangi and of the Indian Nonintercourse Act. While perpetrators and original victims are dead, contemporary people who are classified as members of the wronged social group may qualify as potential claimants if they seem to suffer from the continuing effects of past injustices. Pioneering contributions to the first wave emerged in the 1970s and include Boxill’s “The Morality of Reparation” (1972) and Lyons’ “The New Indian Claims” (1977). It is instructive to contextualize these contributions. While Rawls (1971/1999) had devised, in 1971, an (enormously influential) ahistorical ideal theory of a just basic structure that naturally “has nothing to say about racial injustice” (Mills 1997/1999: 140) or genocides of indigenous peoples, Robert Nozick (1974), his best-known critic, proposed a historical entitlement approach to just holdings more congenial to the first wave literature, but highly critical of the idea of distributive justice. Lyons’ contribution deserves credit for drawing attention to the potential tension between demands of reparative and distributive justice when resources like land are involved. Waldron’s “Superseding Historic Injustice” (1992), one of the best-known contributions to the debate, takes up this thread, generalizing Lyon’s core point that property rights are not stable over time and that, due to changing circumstances, historical demands may be superseded by other claims. Meyer (2004) called Waldron’s supersession thesis one of the “two main sources of theoretical doubt” regarding reparations. The other source of doubt was also proposed in the 1970s. The famous non­ identity argument, independently formulated by at least three philosophers (Boonin 2008: 129), radically questions for a broad range of cases that later-born people can make valid reparative demands. In one application, the argument purports to show that descendants of slaves in the United States have not been harmed by slavery (Herstein 2008: 505). This is because, some argue, harming someone requires making them worse off relative to how they would otherwise have been. But the descendants might not have been worse off in this relative sense because, without slavery, they might not have been born at all. The circumstances of their parents would have been so different that they reproduced at different time and with a different combination of egg and sperm cells, leading to completely different children being born. But if so, slavery did not make the current descendants worse off and, thus, also did not harm them in the relevant sense. In contrast to the interactional understanding of the first wave, second-wave literature places special emphasis on the concept of structural injustice.6 Structural injustices consist in institutional arrangements, culturally transmitted modes of behavior, attitudes and 408

Moral responsibility for historical injustice

mental models that distribute access to opportunities, wealth, power or well-being in an unjust manner, as specified by some theory of distributive justice or theory of rights. Frequently those who are privileged and powerful due to unjust structures cannot be blamed for producing injustice. They act within their rights and – like the oppressed and disadvantaged – according to a social–structural script. From a structuralist point of view, the distinction between perpetrators and victims, vital for the interactional account, is seen as mistaken moralism. Thus, the structural injustice literature was considered largely incompatible with the first wave’s framing of the “rectification project” (Butt 2009).7 In recent years, however, there are some signs of a sea change (Young 2011; Lu 2017; Nuti 2019; Page 2021). Whereas the first wave presupposes that injustices must be past in order to count as historical, new-wave structural theorists assume the opposite; they take the decisive point to be that an injustice began in the (distant) past but has not ended yet. If disadvantaged positions in present societies are relevantly similar to disadvantaged positions in the past, and if the latter are connected to the former via mechanisms that update relevantly similar unjust social structures, then people in these disadvantaged positions can be said to be “structural descendants” (Nuti 2019: 162). While this approach makes room for historical obligations within the structuralist account, Nuti favors the framing of interactional and structural justice as competing paradigms. Butt (2021) objects that both approaches are complementary and should be integrated in a pluralist theory. Interactional and structural injustices can (and typically do) intersect and interact. Not least because historical demands frequently involve substantial transfers of resources, the question arises for structural and interactional accounts alike how historical obligations (as requirements of remedial responsibility) and the demands of distributive justice (as specified by a distributive principle) are related. Perhaps the most popular answer is reductionist: Apart from what is necessary to achieve a just pattern of holdings by means of just social structures, no reparative measures are justified (Wenar 2006). From a liberal egalitarian viewpoint, Pierik (2006: 435) points out that reductionist approaches have considerable advantages. First, they bypass the notorious nonidentity problem because the rectification of past wrongdoing becomes now the side product of an essentially forwardlooking account. Second, reductionist accounts dispense with the need to obtain historical information on the legitimacy of ancient holdings; third, they render the construction of complex counterfactuals on how claimants would fare without the injustice redundant.8 Reductionism has the disadvantage, on the other hand, of not capturing the intuition that wrongs must be rectified by reaffirming and restoring the group’s status as a subject of moral rights.9 Land claims by First Nations in North America, for instance, are not predicated on some distributive principle but on the proposition that the land was taken from them by the use of force and fraud; the claim is that the moral status of the wronged communities was violated and must be restored.10 Moreover, the philosophical debate on HI is not restricted to cases in which material transfers are on the agenda. Some past injustices must be remembered in order to reaffirm and restore the moral status of a historical group and its members (Bluestein 2008); and they must be remembered appropriately (Altanian 2022). It is not obvious how to make sense of this within a reductionist framework.

3  Historical injustice and moral responsibility In light of the idea that a person cannot be morally responsible without making a causal contribution (“moral entails causal”) (Sartorio 2004: 317; see Kaiserman in this volume), 409

Michael Schefczyk

later-born people cannot be morally responsible for HI. But they may be responsible in the remedial sense. Unrectified HIs are states in need of remedy. There are two ways to relate remedial and moral responsibility. First, moral responsibility implies remedial responsibility in the fullest sense. The wrongdoer is obliged to rectify the injustice via acknowledgement of wrongdoing, apology, correcting the effects of wrongdoing and taking measures to avoid wrongdoing in the future. To be sure, remedial responsibility can also be implied by other roles.11 What might be named the “principle of perpetrator priority” states that moral responsibility has, as a basis of remedial responsibility, lexical priority over “backup principles.”12 Second, failing to perform a remedial obligation is obviously morally wrong. In the case of historical obligations, it is an “act of injustice” (Butt 2009: 177). Thus, a collective or individual agent can be morally responsible for the wrong of failing to perform a remedial obligation without being morally responsible for the original wrong which must be remedied. As mentioned in the introduction, I call remedial responsibility predicated on moral responsibility first-hand responsibility; and I call remedial responsibility predicated on morally relevant relations to culpable individuals or collectives second-hand respon­ sibility. In the light of the principle of perpetrator priority, the following standard procedure with regard to remedial responsibility for HI seems natural: Examine first the moral situation at t soon after the event; specify rectificatory obligations and claims at t (note that first-hand responsibility and second-hand responsibility may go hand in hand). Second consider the moral consequences of the passage of time: specify how deaths, births, immigration, emigration, end of states and the foundation of states affect first-hand responsibility and second-hand responsibility. Third, in the case that firsthand and second-hand responsible agents are unable or unwilling to (fully) rectify HI, check the applicability of backup principles: Specify how historical obligations based on first-hand responsibility on the one hand and on backup principles on the other hand are related to each other. The literature on HI presupposes that states are collective moral agents. They can be culpable for the commission of injustices and can thus bear first-hand responsibility. In order to fulfil their remedial responsibility, states use their power to tax the population (in the case of reparations) and to speak for the country as a whole (in the case of apologies). This is unproblematic if a nexus between state culpability and population culpability exists. There are three main models for such a nexus. First, Stilz’ authorization model (AM) conceptualizes the population as a unified political body consisting of the citizens of a democratic legal state. They “share in the liability for what their state does” (2011: 191) because they have reason to authorize their state. Authorization, in turn, means that the state is entitled to act in the name of its citizens. States are so entitled when they rule in a way that citizens have “reason to accept” (p. 198). If the state acts as the ideal agent of the people (the principal), they have “reason to ‘own up’ to what an authorized state does” (p. 198). One limitation of this authorization account is its narrow scope. States that commit large-scale injustices are often not democratic legal states. For example, the authorization model does not apply to a regime like the German Kaiserreich. Another limitation is that democratic citizens can authorize only state acts that qualify as “a reasonable interpretation of rights.” An unreasonable interpretation of rights creates no liability for a state’s citizens. The de facto support for the criminal activities of their government abroad would not count as authorization in the technical sense and therefore, according to the model, would not create liability. 410

Moral responsibility for historical injustice

A de facto consent-based liability of residents is found in both Miller’s (2004) likeminded group model (LMGM) and Pasternak’s (2021) participatory intent model (PIM). In essence, LMGM argues that an approving population influences government action, either by providing an enabling condition (the government would stop its criminal activities if they were sufficiently unpopular) or by actively incentivising it; the population is in this way complicit in state injustice, even if it does not authorize it in Stilz’ sense. LMGM delivers the intuitively plausible result that even disunited and politically disenfranchised populations of autocratic or oligarchic states can be blameworthily involved in state crimes. PIM is more demanding than LMGM insofar as it requires the population’s identification with the state and the intention of the people to do their bit in upholding order and keeping the state active. If people show such participatory intent, they conceive of themselves and act as citizens, as constituent parts of the state. This “renders one the inclusive author of what the group does” (Pasternak 2021: 71). This form of “intentional citizenship” is possible in autocratic or oligarchic regimes (p. 120). PIM covers cases in which a state’s criminal activities push beyond what the population would have been willing to accept in other circumstances. These cases tend to fall through the cracks of LMGM.13 AM, LMGM and PIM allow us to construct a link between the first-hand responsibility of states and that of populations. But we have not yet turned to the question how to connect first-hand responsibility and second-hand responsibility. Let us now consider different constellations of the first-hand responsibility of states and of populations and their effects on second-hand responsibility. As mentioned, the proposed standard procedure starts with an examination of the moral situation at t. This means that we have to ask whether a state or its population (or both) is (are) culpable for an injustice. We consider the moral situation soon after the event and ignore, in the first step, the moral consequences of the passage of time. Note that AM and PIM conceptualize the population as an independent political body. In ideal democracies, the political body is perfectly and fully represented by the state; in other regimes, the state represents the political body partially, imperfectly or not at all. The existence and persistence of a political body does not depend on the existence and persistence of a particular state or form of government. There is thus room for divergence in both directions. A  present state may be culpable and bear first-hand responsibility, while the population is not culpable and does not bear first-hand responsibility; or a present political body may be culpable and bear first-hand responsibility, while the state does not. As an example for the latter, think of the Federal Republic of Germany, which was founded after the Second World War and has no first-hand responsibility for crimes committed under the National Socialist regime. As a democratic successor state, however, it represents the political body, the German people. On the uncontroversial assumption that the German people were culpably involved in NS crimes; they thus bear first-hand responsibility; the Federal Republic stands in a morally relevant relation to a culpable political body and therefore carries second-hand responsibility. As an example for the former, think of the German Kaiserreich. On the perhaps problematic conjunction that the Kaiserreich was morally co-responsible for the Armenian genocide and that its subjects were not,14 one may argue that neither the Weimar Republic nor its citizens bore first-hand responsibility. The remedial responsibility of the Republic and its citizenry at t had then to be based on backup principles, most plausibly on outcome responsibility (Miller 2007: 100–101). As outcome responsibility is not the only backup principle, its application requires further argument to the effect that it is better to oblige the innocent population of a defunct culpable state than, 411

Michael Schefczyk

for instance, innocent beneficiaries. Depending on the circumstances, this argument may be difficult to make. First-hand responsibility and second-hand responsibility of the political body typically go hand in hand. The reason is that neither authorization, nor participatory intent or likemindedness presuppose unanimity in the population. If the majority authorized, willingly helped or enthusiastically approved of an injustice, this typically suffices to support the claim that the population as a whole was culpable.15 On closer inspection, to be sure, the members of the minority are not first-hand responsible as they did not authorize, willingly help or approve the injustice; but they bear second-hand responsibility because of their moral relation, to wit shared citizenship, to those who did. With the passage of time, the composition of first-hand and second-hand responsible people in the population changes. The absolute and relative number of those who authorized injustices, participated in them or applauded them constantly decreases due to deaths, births, immigration and emigration. At some point, the number of first- or second-hand responsible people falls below a critical threshold. Reminiscent of Parfit’s view of personal identity, one may argue that there comes a moment at which the political body is not the same anymore. The collective memory, and other characteristics, are not suitably continuous and connected to former states of the political body. The present political body may then take an “attitude of nonidentification” (Parfit 1971: 687) to the political body which was co-responsible for HI. Possibly, backup principles of remedial responsibility have to be activated then; possibly, the historical obligation vanishes. These considerations help to understand why, as a general rule, HI start to lose absolute and relative moral weight at some point in time. Relevant factors for the absolute weight of an injustice are the identity (in the reductionist sense) of collectives to which victims or perpetrators belonged and the importance of an injustice for the collectives’ present well-being and self-understanding. The latter two (must) dwindle with the passage of time. Simultaneously, new obligations emerge so that the relative moral weight of historical obligations typically also declines. Although both effects usually seem to act unidirectionally, it is important to be aware that the relative weight of historical obligations might also increase in the case that other obligations lapse.16 Butt (2009: 183–188) argues that in certain circumstances a kind of normative linking offsets the diminishing absolute weight of HI. He discusses the case of restitution; a nation N has at t the moral obligation to return some material stuff X (artworks, land). If N refuses to return X at t, it breeds a second injustice, the failure to fulfil a rectificatory duty. Thus, N is at t1 not only obliged to return X but also to compensate the claimants for the “standing time.” If there is a continuous chain of refusing referenda over the return of X and if the composition of the population does not change dramatically between the referenda, the burden of rectification is linked and may become greater, not lesser, over time.

4  Historical injustice and the cultural ignorance excuse Can states and political bodies be first-hand and second-hand responsible when they are morally ignorant of the injustice? A theory of moral responsibility which Levy (2005) calls “attributionism” denies the exculpatory force of moral ignorance.17 According to this view, historical agents who were convinced that colonization or slave-owning or mass murder were morally acceptable and acted accordingly would be blamed for these injustices, no matter whether these agents could have avoided being so convinced or not. 412

Moral responsibility for historical injustice

In the following discussion, I  consider approaches that  – in contrast to strict attributionism  – assume that moral ignorance can have exculpatory force when agents did not wrongfully contribute to their ignorance. One version of this view is provided by Wolf (1987/2003: 379–380), who gives the fictitious example of an evil dictator’s son, JoJo, who was educated in a way that effectively prevents him from responding in the right way to morally relevant features. Having been morally incapacitated by his educators, JoJo is, according to Wolf, innocently ignorant and thus excused for derivative wrongdoing.18 The ignorance excuse thus opens the door for a distinction between (extremely) bad historical perpetrators and those who are also blameworthy. Some of them might have been horrible human beings, but they possibly did not and could not know what they were doing. Excusing agents for failing to realize the relevant moral reasons because they were incapacitated by their socio-epistemic environment goes by the name of cultural ignorance. In a naïve application of the cultural ignorance excuse, historical agents are excused when certain practices or institutions were not illegal  – or were even required by law  – at the time. Although simplistic, the naïve application is not without an element of truth. Legality is insufficient to excuse historical perpetrators, but it can work as a component in a plea for cultural ignorance. People in many societies use the law as at least one important source of moral orientation.19 If its laws prescribe p or support p or tolerate p, members of this society are likely, and sometimes even justified, to consider this as evidence for p’s moral permissibility. Unjust laws can thus act as epistemic impediments. As this effect can be offset by alternative sources of moral orientation or by independent reasoning, a plea for cultural ignorance would have to make the claim that historical agents, given their socio-economic environments and their level of epistemic competence, had no fair chance to realize the wrongness of p. But what does a “fair chance” require? The more difficult epistemic tasks are, the more likely agents will fail at solving them. But how difficult is too difficult? Whether epistemic impediments are sufficient to make a case for a mitigating or an exempting excuse not only depends on the difficulty of the task – given a complex set of socio-epistemic factors – but also on the fixing of a normative standard by present observers. A subjective element of “normative taste” for more or less forgiving standards might be hard to avoid here. Socio-economic environments can be more or less favorable to the acquisition of moral awareness. I  shall mention three dimensions of socio-economic environments. The first consists of the ideational resources of a society. Imagine a scale for the statement “slavery is wrong” that ranges from “unheard-of” to “everyone heard of it.” The closer a past society is to the left end of said scale, the more difficult it is for its members to realize that some practice – like slavery – is wrong. Rosen, for instance, assumes chattel slavery was “simply taken for granted” (2003: 65) until “quite late in antiquity” (2003: 64). If members of such societies thought about slavery, they could not draw on circulating views that were critical. That made it more difficult for them to realize its wrongness. Another dimension is the level and distribution of epistemic skills and virtues in a society. A high degree of epistemic competence can arguably offset the pro tanto exculpatory effect of lacking ideational resources. One may wonder, for instance, how an intellectually agile and accomplished culture – like Athens in the time of Socrates and Plato – could fail to see the wrongness of slavery. Although epistemic resources can be very unevenly distributed in society, members of an epistemic elite were, in principle, fully capable of realizing the injustice of certain practices, even if common people were not. Provided that Aristotle did indeed defend the existence of natural slaves and justified the enslavement of such naturals 413

Michael Schefczyk

(some scholars disagree with both claims), he could and should have realized that his arguments were incoherent (again, some scholars disagree).20 Ordinary Athenians without the time and education to competently reflect on such issues, however, were possibly excused for their ignorance. A third dimension consists in the availability of real-world examples for the possibility of abolishing or living without some unjust practice or institution p, say, feudal oppression and disenfranchisement. It is one thing to imagine a world without p or to argue against the moral permissibility of p; it is quite another to be able to point to working alternatives. The intuition here is that a lack of known viable alternatives presents an epistemic impediment to realizing the injustice of an established institution.21 Uncontroversial real-world examples of completely morally ignorant cultures (with regard to p) are hard to come by. Rosen gives the largely fictional case of an “ordinary Hittite lord” (Rosen 2003: 64), a member of an ancient slaveholding society. Even if we apply demanding standards of epistemic responsibility, Rosen argues, “it seems clear that the ancient slaveholder need not be responsible for this moral ignorance” (p. 65). More relevant for real-world debates about HI – and certainly more controversial – is Wolf’s tentative application of the cultural ignorance excuse to “the slaveowner of the 1850s, [and] the Nazis of the 1930s” (1987/2003: 382). Wolf does not claim that even the epistemically most virtuous Americans (1850s) or Germans (1930s) were incapable of arriving at the right moral conclusions or that both societies lacked ideational resources, epistemic competence or real-world examples in general. She does make a case for partial cultural ignorance. Some Americans who were slaveowners in the 1850s and some Germans who were Nazis in the 1930s might have lived in a socio-epistemic subculture that excused them for their moral ignorance.22 Since in both societies the right moral answers were available, an attempt to construct a partial cultural excuse must emphasize the difficulty in appropriating them in particular subcultures. Such “difficulties in trying” (Guerrero 2017: 211–212) result from the interplay of social and psychological factors associated with the specifics of their socio-economic environments. As a general rule, it becomes more difficult to appropriate a right moral answer, the more costly the appropriation is in one’s life situation. One may distinguish here between purely intellectual, emotional and practical costs. If, say, accepting “slavery is wrong” would require a deep revision of a person’s belief system, the purely intellectual costs would be relatively high. This could be the case for a member of US society in the 1850s who had no personal stake in slave-owning but was heavily invested in Aristotelian justifications for slavery (see Campbell 1974). The difficulty in trying also increases when a right answer involves a departure from norms and values accepted in one’s social network, leading to possible social exclusion and loss of economic opportunities. Avoiding such practical costs by concealing one’s beliefs in public or by abstaining from acting upon them, would still involve emotional costs like guilt feelings, shame or fear of discovery. The (unconscious) awareness of such emotional problems may motivate rationalization and denial strategies. It remains controversial how strong the exculpatory force of difficulties in trying actually is. Mason and Wilson (2017) plausibly argue that one has to consider the moral quality of explanatory factors affecting such difficulties. In cases in which ignorance very likely results from the interaction of self-interest and a lack of moral-epistemic virtues, one may presume that the agents could and must have known better and that their ignorance was manufactured or “affected,” to use Moody-Adams (1994) term. But there appears to be no way to make a watertight, uncontroversial case for such a claim. As mentioned before, 414

Moral responsibility for historical injustice

judging reasonable expectations regarding the moral competence of other cultures unavoidably involves the normative taste for more or less forgiving standards.

5  Concluding remarks Questions of first-hand and second-hand responsibility are highly relevant for the debate on historical injustices. It matters whether a state was culpable or not, whether a population was complicit in crimes or not, whether historical actors were morally ignorant or not, and whether a complicit population formed and forms an intergenerational political body or not. Answers to these questions influence the content of historical obligations; they help us to justify why “some occurrences [are] picked out as demanding redress, while others are ignored” (Vernon 2012: 12).

Notes 1 Thompson defines historical injustice disjunctively as “a wrong either done to or by past people” (2002: x). Ivison presents a slightly different conjunctive variant and refers to injustices committed “by people now dead against people who are also dead” (Ivison 2009: 508). This “death requirement” is not fulfilled in all cases of HI discussed in the literature. But as the death requirement gives rise to specific philosophical problems, it is commonly treated as a waypost of the debate. 2 Historical obligations are obligations of later-born people for HI. 3 I use the term “culpable” as synonymous with “morally responsible.” 4 Miller defines remedial responsibility as the obligation to put right a state of affairs in need of remedy (2007: 98). 5 For the distinction between “reparative” and “corrective justice,” see Platz and Reidy (2006: 361– 362). The authors contend that reparative justice is geared towards the repair of the relationship. I propose to interpret this in the sense that reparative justice demands the wrongdoer reaffirm and restore the victim’s status as a subject of moral rights. The authors, in contrast, make the stronger claim that reparative justice “demands of victims a willingness to venture forgiveness or at least reconciliation in response to wrongdoer’s reparative efforts at making amends” (p. 362). 6 For an excellent exposition of the literature on structural injustice, see McKeown (2021). 7 Even though Young conceded that “where individual victims and perpetrators are still alive, application of the liability model is appropriate” (2011: 173), the general thrust was to redirect attention to the so-called forward-looking “social connection model of responsibility” (p.  173). As Young recommended the social connection model for cases in which perpetrators and victims are deceased, she rejected in effect the rectification project for historical injustice. 8 Interestingly, Nozick (1974) arrives at a similar position from the opposing starting point of a historical entitlement theory of just holdings. Lacking much historical information, Nozick (1974: 231) recommends the temporal application of Rawlsian redistribution policies as a proxy in order to rectify the cumulated effects of historical injustices. The difference principle would work as a means of transitional justice. 9 On the idea that rectification is owed to the dead, see Ridge (2003) and Stemplowska (2020). 10 Non-reductionists, like Waldron, concede that reparative claims based on the original acquisition of land have a standing of their own. Reparative justice may thus involve the satisfaction of material demands as long as they do not conflict with the requirements of distributive principles. 11 Apart from moral responsibility, Miller (2007: 100–104) mentions causal responsibility, benefit, capacity and community. 12 Backup principles may be needed in cases like these: The perpetrators are dead or do not exist anymore (groups, corporations, states); the resources of the perpetrator are insufficient to rectify the material effects of injustice; the moral opportunity costs of full rectification are too high for the perpetrator (for instance, if full rectification would cause severe hardship in the perpetrator state); the perpetrator is unwilling to rectify and too powerful to be forced to do so. In such cases, backup principles of remedial justice may come into play. In our context, the most important of those principles are the beneficiary principle and the principle of outcome responsibility.

415

Michael Schefczyk 13 The German population, for instance, did not show much resistance to the Nuremberg Laws and the blatant disenfranchisement of German Jews in the 1930s. Like-mindedness is less obvious with regard to the extermination program, in which Germans intentionally participated in sufficient numbers when it finally came to that. In contrast to LMGM, PIM seems to deliver the intuitively correct result here. 14 The complicity of the Kaiserreich and of German state officials in the Armenian genocide are not controversial (Schefczyk 2018). But it is less clear whether the German population was complicit in the sense of LMGM. 15 Why this is supposed to be so is surprisingly difficult to show. See Risse (2004) for the related problem of justifying the majority rule. 16 Max Kiener called my attention to this very interesting point. 17 With some variation in terminology, attributionists argue that moral blame refers to a person’s character, attitudes, quality of will or real self. If Q’s moral ignorance regarding some wrong-making feature is expressive of who Q is, one blames Q for being morally ignorant and, derivatively, for the morally wrong actions that result from it. 18 For a critical engagement with Wolf, see Levy (2003). 19 Other sources of moral orientation are one’s education, social milieu, religious teachings, ideologies and so forth. Our more or less spontaneous emotional responses to thoughts and events can also be included in this list. For the consistently spontaneous decision to rescue Jews in Nazi Europe, see Monroe et al. (1990). 20 See Lindsay (1994). 21 Mill seems to ponder on a related point at the end of Utilitarianism: “All persons are deemed to have a right to equality of treatment, except when some recognized social expediency requires the reverse. And hence all social inequalities which have ceased to be considered expedient, assume the character not of simple inexpediency, but of injustice, and appear so tyrannical, that people are apt to wonder how they ever could have been tolerated” (Mill 1863/1969: 258). 22 As far as the Nazis of the 1930s are concerned, the case for innocent moral ignorance looks weak. The best candidates for the “toxic education excuse” are those born after 1915 because they were young enough for Nazi brainwashing in schools, etc. The more exposed they were (for birth cohort 1927, the first day at school was in 1933), the less involved they were in Nazi crimes (being 30 or younger at the end of the war).

Further reading Booth, W. J. (2020). Memory, Historic Injustice, and Responsibility. New York and London: Routledge defends the idea that caring for the dead is an essential element of historical justice. Butt,  D. (2006). Rectifying International Injustice. Principles of Compensation and Restitution Between Nations. Oxford: Oxford University Press and Lu, C. (2017). Justice and Reconciliation in World Politics. Cambridge: Cambridge University Press are mandatory reading for those who are interested in historical injustice on an international scale. Thompson, J. J. (2002). Taking Responsibility for the Past. Reparation and Historical Injustice. Cambridge: Polity Press is one of the most influential books of the “first wave” of historical injustice literature. Vernon, R. (2012). Historical Redress. Must We Pay for the Past? London: Continuum provides a well-written and thorough introduction that covers central topics of the debate. Spinner-Halev,  J. (2012). Enduring Injustice.  Cambridge: Cambridge University Press argues that certain present injustices will persist if we fail to understand that they are enduring past injustices.

References Altanian, M. (2022). “Genozidleugnung: Organisiertes Vergessen oder Substantielle Erkenntnispraxis?” Zeitschrift für Praktische Philosophie, 9(1), 251–278. Bluestein, J. (2008). The Moral Demands of Memory. Cambridge: Cambridge University Press. Boonin, D. (2008). “How to Solve the Non-Identity Problem,” Public Affairs Quarterly, 22(2), 129–159. Boxill, B. (1972). “The Morality of Reparation,” Social Theory and Practice, 2(1), 113–123.

416

Moral responsibility for historical injustice Butt, D. (2009). Rectifying International Injustice: Principles of Compensation and Restitution Between Nations. Oxford: Oxford University Press. ———. (2021). “What Structural Injustice Theory Leaves Out,” Ethical Theory and Moral Practice, 24(5), 1161–1175. Campbell, M. (1974). “Aristotle and Black Slavery: A  Study in Race Prejudice,” Race, XV(3), 283–301. Guerrero, A. (2017). “Intellectual Difficulty and Moral Responsibility,” in P. Robichaud & J. W. Wieland (eds.), Responsibility: The Epistemic Condition (pp. 199–218). Oxford: Oxford University Press. Herstein, O. (2008). “Historic Justice and the Non-Identity Problem: The Limitations of the SubsequentWrong Solution and Towards a New Solution,” Law and Philosophy, 27(5), 505–531. Ivison, D. (2009). “Historical Injustice,” in J. S. Dryzek, B. Honig, & A. Phillips (eds.), The Oxford Handbook of Political Theory (pp. 507–525). Oxford: Oxford University Press. Kaiserman, A. (this volume). Levy, N. (2003). “Cultural Membership and Moral Responsibility,” Monist, 86(2), 145–163. ———. (2005). “The Good, the Bad, and the Blameworthy,” Journal of Ethics & Social Philosophy, 1(2), 2–16. Lindsay, T. (1994). “Was Aristotle Racist, Sexist, and Anti-Democratic? A Review Essay,” Review of Politics, 56(1), 127–151. Lu, C. (2017). Justice and Reconciliation in World. Cambridge: Cambridge University Press. Lyons, D. (1977). “The New Indian Claims and Original Rights to Land,” Social Theory and Practice, 4(3), 249–272. Mason, E., & Wilson, A. (2017). “Vice, Blameworthiness, and Cultural Ignorance,” in P. Robichaud & J. W. Wieland (eds.), Responsibility: The Epistemic Condition (pp. 82–100). Oxford: Oxford University Press. McKeown, M. (2021). “Structural Injustice,” Philosophy Compass, 16(7). Meyer, L. H. (2004). “Historical Injustice and the Right to Return,” Theoretical Inquiries in Law, 5(2), 305–316. Mill, J. S. (1863/1969). “Utilitarianism,” in Collected Works of John Stuart Mill, Volume 10 (Essay on Ethics, Religion and Society) (pp. 203–259). Toronto: University of Toronto Press. Miller, D. (2004). “Holding Nations Responsible,” Ethics, 114, 240–268. ———. (2007). National Responsibility and Global Justice. Oxford: Oxford University Press. ———. (this volume). Mills, C. (1997/1999). The Racial Contract. Ithaca and London: Cornell University Press. Monroe, K., Barton, M., & Klingemann, U. (1990). “Altruism and the Theory of Rational Action: Rescuers of Jews in Nazi Europe,” Ethics, 101(1), 103–122. Moody-Adams, M. (1994). “Culture, Responsibility, and Affected Ignorance,” Ethics, 104(2), 291–309. Nozick, R. (1974). Anarchy, State, and Utopia. Oxford: Blackwell. Nuti, A. (2019). Injustice and the Reproduction of History: Structural Inequalities, Gender and Redress. Cambridge: Cambridge University Press. Page, J. (2021). “Contributing to Historical-Structural Injustice via Morally Wrong Acts,” Ethical Theory and Moral Practice, 24, 1197–1211. Parfit, D. (1971). On “The Importance of Self-Identity,” Journal of Philosophy, 68(20), 683–690. Pasternak, A. (2021). Responsible Citizens, Irresponsible States. Should Citizens Pay for Their State’s Wrongdoings? Oxford: Oxford University Press. Pierik, R. (2006). “Reparations for Luck Egalitarians,” Journal of Social Philosophy, 37(3), 423–440. Platz, J., & Reidy, D. (2006). “The Structural Diversity of Historical Injustices,” Journal of Social Philosophy, 37(3), 360–376. Rawls, J. (1971/1999). A Theory of Justice, Revised Edition, Cambridge, MA: Belknap of Harvard University Press. Ridge, M. (2003). “Giving the Dead Their Due,” Ethics, 114, 38–59. Risse, M. (2004). “Arguing for Majority Rule,” Journal of Political Philosophy, 12(1), 41–64. Rosen, G. (2003). “Culpability and Ignorance,” Proceedings of the Aristotelian Society, 103, 61–84. Sartorio, C. (2004). “How to be Responsible for Something Without Causing It,” Philosophical Perspectives, 18, 315–336.

417

Michael Schefczyk Schefczyk, M. (2018). “Modern Germany and the Annihilation of the Ottoman Armenians: A Note on the Political Avowal of Shame and Guilt,” in M. Altanian (ed.), Der Genozid an den ArmenierInnen. Beiträge zur wissenschaftlichen Aufarbeitung eines historischen Verbrechens gegen die Menschlichkeit (pp. 85–109). Cham: Springer VS. Stemplowska, Z. (2020). “Duties to the Dead: Is Posthumous Mitigation of Injustice Possible?” in D. Sobel, P. Vallentyne, & S. Wall (eds.), Oxford Studies in Political Philosophy, vol. 6 (pp. 32–60). Oxford: Oxford University Press. Stilz, A. (2011). “Collective Responsibility and the State,” Journal of Political Philosophy, 19(2), 190–208. Thompson, J. (2002). Taking Responsibility for the Past: Reparations and Historical Justice. Oxford: Polity Press. Vernon, R. (2012). Historical Redress: Must We Pay for the Past? London and New York: Continuum. Waldron, J. (1992). “Superseding Historic Injustice,” Ethics, 103(1), 4–28. Wenar, L. (2006). “Reparations for the Future,” Journal of Social Philosophy, 37(3), 396–405. Wolf, S. (1987/2003). “Sanity and the Metaphysics of Responsibility,” in G. Watson (ed.), Free Will (pp. 372–387). Oxford: Oxford University Press. Young, I. M. (2011). Responsibility for Injustice. Oxford: Oxford University Press.

418

33 CORPORATE DIGITAL RESPONSIBILITY Alexander Filipović

1 Introduction We are all required to “act responsibly” and, in our ordinary lives, this requirement points to four different (normative) expectations. First, there is the expectation that we act in caring, good, fair, or socially acceptable ways. This sense of responsibility coincides with the meaning of moral action more generally, and requires that we act in accordance with moral norms. Second, there is, or at least can be, an expectation that we act on behalf of others and their (moral) interests, leading to some form of human advocacy or representation. Third, there is the expectation that we prospectively assess the potentially versatile consequences and ramifications of our conduct in sufficient detail. And fourth, acting responsibly comes with the normative expectation that, whenever we act wrongly, we should be prepared to admit our guilt and apologize. These four normative expectations play an essential role in applied ethics too, that is the field of ethics that deals with morally relevant problems in specific fields of action (Frey & Wellman 2003). Sub-areas of applied ethics include medical ethics, media ethics, financial market ethics, social work ethics, or business ethics. In this chapter, I will address the topic of responsibility within one area of applied ethics, namely business ethics, and I will specifically focus on corporate digital responsibility. I will subscribe to the aforementioned four-expectations account of responsibility, alongside the important distinction between responsibility for future action (prospective responsibility) and responsibility for past action (retrospective responsibility). In applied ethics, one can distinguish between a bottom-up approach and a top-down approach (Beauchamp 2003: 7–12). The bottom-up approach starts from specific judgments and experiences. It focuses on specific acts, cases, or situations and uses an appreciation of these specific instances to derive its ethical reflection and some more general ethical judgments. Proponents of this position “point to our use of existing social agreements and practices, insight-producing novel cases, and comparative case analysis as the starting-points from which we commonly make moral decisions” (Beauchamp 2003: 8). For example, medical action is characterized by a broadly consensual, normative practice

419

DOI: 10.4324/9781003282242-45

Alexander Filipović

of action that can be surveyed, systematized, and generalized, and then used as “principles” for cases and their ethical evaluation (so paradigmatically Beauchamp & Childress 1979). By contrast, the top-down approach proceeds in the opposite direction. It starts from general principles and theories (such as autonomy or justice) that are then “applied” to specific cases (cf. Beauchamp 2003: 7). However, neither bottom-up approaches nor top-down approaches explicitly feature aspects of responsibility. This is because both types of approaches primarily focus on substantive moral considerations, such as specific moral norms or intuitions (bottom-up) or general moral principles (top-down), whereas responsibility appears, if anything, as a formal moral principle, and requires an additional substantive component to guide action. While the distinction between “substantive” (or material) and “procedural” (or formal) theory is often too simplistic, I think it can help to elucidate this important aspect about the concept of responsibility and its difficult position in applied ethics. Yet, despite the somewhat unclear position of the concept of responsibility in applied ethics, various scholars and practitioners in applied ethics, including in business ethics, use the term “responsibility”. The corporate world uses the concept of responsibility in the everyday sense mentioned before and relates it to specific practical challenges. What is more, business ethics, as a form of applied ethics, is not only a philosophical discipline in the strict sense (i.e., practiced by philosophers at universities) but it is also developed, systematized, and discussed in various practical areas (sometimes with the help of philosophers). In this chapter, I  will focus on concepts of responsible corporate action, assess them philosophically, and form them into a concept of applied digital ethics in the corporate context. Business ethics as a subfield of economic ethics is an intensively researched area. The relevance of this area is clear: historically, companies have become central factors of social development over the last 100 years (French 1995: VII); today, in the age of globally operating companies in a digital context, problems of corporate power are intensifying. In business ethics, the concept of responsibility has always played a central role. In addition to general corporate responsibility (CR) toward other institutions, groups and individuals, the focus of business ethics is on the responsibility of managers, employees, company owners, etc. Corporate responsibility (CR) generally concerns the positive contribution business enterprises make to society. Most definitions of corporate responsibility (CR) share “the belief that companies have a responsibility for the public good” (Blowfield & Murray 2019: 7). The contexts for responsible corporate governance are diverse and concern not only abstract ideas about the social benefits of a company but also specific responsibilities, for example, for the living conditions of people in the local neighborhood of the company’s headquarters. In this respect, Blowfield & Murray also use the term as an umbrella term that captures the variety of ways in which business’s relationship with society is being defined, managed, and acted upon. Therefore  .  .  . Corporate responsibility (CR) comprises (a) the responsibilities of business in the context of wider society, (b) how those responsibilities are defined and negotiated, and (c) how they are managed and organized. (Blowfield & Murray 2019: 11) Looking at the changes in the ideas of CR since the 1950s, for example, it becomes clear how much social, economic policy and cultural contexts influence corporate responsibility 420

Corporate digital responsibility

(CR) in each case. For some time now, people have been talking about corporate social responsibility (CSR) in academia, corporate policy and corporate management. Particularly in the context of the 2008 financial crisis, the social demands on corporate action generally increased. Above all, however, in view of the climate crisis, CSR concepts have focused on the environmental compatibility of corporate action. With the rapid development of digital technologies (computer technologies based on algorithms, data, and artificial intelligence), digitalization now represents a central and new context of corporate responsibility (CR). Corporate digital responsibility (CDR) is a concept of responsibility specific to organizations (such as companies and civil society actors) and institutions in the digital age. The concept is recognizably related to corporate social responsibility (CSR). CDR updates and specifies CSR for structures and processes in which society, the economy, politics, media and culture are changing in the wake of technological innovations (especially digital communications infrastructure).

2  Business ethics as ethics of responsibility The topic of responsibility raises three central questions: (i) Who is responsible? (Subject of responsibility) (ii) To whom is one, e.g., a corporation, responsible? (Addressee of responsibility) (iii) For what is one, e.g., a corporation, responsible? (Object of responsibility) In the context of business ethics, a first answer to the “who” of responsibility could be that the company itself bears responsibility and is understood as a moral agent. This is not an unusual idea, given that companies are already legal subjects across the criminal, civil and public law, and, on this basis, they are held legally responsible, which includes retrospective responsibility for the damage they cause and thus need to answer for, as well as prospective responsibility for the welfare of the people (and animals and the environment) affected by their actions (in the future). One may therefore extend this view to morality and regard companies not only as legal but also as moral agents. But once we do so, we may face the objection that moral agency as well as moral responsibility is tied to the personal status of humans, so that only (individual) persons but not companies can bear genuine moral responsibility. Although the dispute has not yet been decided, the answer to the question plays a vital role in the status of business and corporate ethics as responsibility ethics. In the formation of the ethics of responsibility in the 20th century, scholars, including Hans Jonas, held that the capacity of human agency is systematically limited (Jonas 1984). Therefore, considering only individuals as potential bearers of moral responsibility would lead to fundamental gaps in responsibility and leave unaccounted for many instances of collective action in their inherently social and organizational forms. To avoid these gaps, Neuhäuser (2011) plausibly argues (following Peter A. French, e.g., French 1979) that corporations should be assigned the status of moral agents (cf. Bazargan-Forward  & Tollefsen 2020). In business ethics, companies are usefully described as (collective) moral agents. More specifically, their moral agency rests on a discursive, communication-based concept of responsibility, with a special focus on accountability. Buddeberg and Hecker provide such a view. They argue that many accounts of responsibility, both of the individual and collective 421

Alexander Filipović

type, are often based on a truncated notion of responsibility and that they overlook a central aspect, namely that bearing responsibility does not only mean someone can be ascribed authorship, can be blamed, punished, or subject to appraisal and held liable for the consequences of his or her action, but that someone can (and under specific conditions is obliged to) answer to queries and account for his or her actions by providing good reasons. (Buddeberg & Hecker 2018: 468) This view is based on Rainer Forst’s account of the right to justification (Forst 2002: 256, 2012), and it renders responsibility a genuinely discursive and communication-based idea, within which companies can play a distinctive role too and exercise their own form of discursive practical reason. Such a view also fits nicely with business practices. Companies already constantly purport to justify their actions. There are justification strategies for every corporate decision, such as reports, memos, board papers, supervisory board reports, shareholder information, annual reports and so on. In addition, highly professionalized communications departments play a central role in corporate success. Taken together, this shows that companies already have significant experience and capacities for justifying their actions to stakeholders, policymakers, or the wider public. Thus, it is not far-fetched to consider companies as bearers of a discursively constructed responsibility, and explain their corporate responsibility (CR) in terms of discourse and communication. So understood, corporate responsibility not only requires the ability to act as a collective moral agent but also the capacities for communication and discourse, along the lines of Forst’s fundamental right to justification (Forst 2012). But note that Forst’s concept of justification requires more than giving a mere (causal) explanation of one’s actions. It requires giving justificatory or normative reasons, which are also intelligible to others, which can be part of a meaningful discourse or negotiation, and which potentially allow the participants of such an exchange to converge on shared moral views. Arguably, corporations can give such justifications and thereby become bearers of responsibility (Buddeberg & Hecker 2018: 470–471). Thus, so far, I have explained the first question mentioned previously, i.e., who is responsible, and I did so in terms of communication: a subject or bearer of responsibility is any entity capable of engaging in the relevant discursive or communicative practice of responsibility. But on this basis, we can also make progress on the second and third questions, i.e., to whom one is responsible and what one is responsible for. Once responsibility is understood discursively, the addressee(s) of responsibility (relating to the second question) are already in the picture as well. These addressees are all those who are either affected by the action or conduct under consideration or who represent those that are affected. Thus, addressees include individuals, stakeholders, society, and potentially also the representatives of future generations, animals, or even the inanimate environment (Freeman 1984; Blowfield & Murray 2019: 194–217). Moving on to the third question, regarding what one is responsible for, we can refer to different types of corporate responsibility (CR), relying on Carrol’s canonical work (Carroll 1979). Carrol distinguishes between economic responsibility (responsibility for providing society with desirable goods), legal responsibility (responsibility for compliance with the law), ethical responsibility (responsibility for what is ethically expected from corporations 422

Corporate digital responsibility

and goes beyond the law), and discretionary responsibility (responsibility for further desirable and discretionary actions, such as philanthropy). These categories, which have been further developed in the academic literature (Blowfield & Murray 2019: 17–23), set out what companies are responsible for and support corporate responsibility (CR) in practice (Blowfield & Murray 2019: 125–153). Finally, what companies are responsible for has a temporal perspective too. Once responsibility is understood discursively, companies are not only retrospectively responsible for the things they did in the past but also prospectively responsible for the things they ought to do in the present or future (Neuhäuser 2011: 46). Prospective responsibility could be as broad as retrospective responsibility and could include animals and the environment. What is more, it can be understood as a form of care-responsibility which creates a diverse network of agents, premised on mutual solidarity.

3  Corporate Digital Responsibility (CDR) – a concept under debate Companies face different challenges over time and corporate responsibility (CR) depends strongly on the context. For instance, in the aftermath of the 2008 financial crisis, many companies (especially banks) faced the challenge of overcoming previous irresponsible conduct related to financial products. In the context of global warming, companies face challenges related to environmental sustainability. In the context of social progress, they have to fight for diversity, equality, and inclusion. And when it comes to globalization, there are yet more sets of distinct challenges. Of particular importance, however, are the challenges surrounding the increasing digitalization of our world. Digital transformation fundamentally changes politics, the economy, society, globalization, law, etc., as well as people’s individual lives, raising diverse problems at all levels. And since companies are key drivers of this transformation, digitization is moving into the focus of corporate responsibility (CR) too. As result, the idea of Corporate Digital Responsibility (CDR) has emerged and shaped ethical and political debates since around 2016 (for an introduction, see, e.g., Dörr 2021; Lobschat et al. 2021).

3.1  First orientation: working definition of Corporate Digital Responsibility (CDR) CDR is still a fairly recent concept in business ethics and public debates. To increase terminological clarity, I will provide a working definition now, before going into greater detail later on. The consequences of digital technologies for politics, business, and society and most people’s everyday lives are immense, and often perceived as problems. The idea of corporate digital responsibility (CDR) places particular emphasis on describing the digital transformation in problem-oriented terms and presenting it as an opportunity for corporate action. But we need to ask what exactly is changing in the context of this transformation, what causes such change, and which problems arise as a result. Definitions of corporate digital responsibility (CDR) provide a first orientation. Some meta-studies and recent articles collect such definitions (Herden et al. 2021: 16; Mueller 2022: 692). For initial guidance, I define CDR as follows: Corporate Digital Responsibility refers to the incorporation of social and environmental considerations into a company’s digital strategy, which is done voluntarily. Acting in a digitally responsible manner requires 423

Alexander Filipović

not only adhering to legal norms, but also promoting ethical and sustainable business practices that generate social, environmental, and economic benefits that go beyond mere legal compliance (cf. Koska & Filipović 2019: 31). Some scholars use the concept of sustainability too and unite a set of social, economic and ecological aspects (Dörr 2021: 36). Furthermore, existing definitions of CDR refer to a set of norms, either specifically to a “set of values and specific norms that govern an organization’s judgments and choices” (Lobschat et al. 2021: 876) or more broadly to a “set of practices, policies, and governance structures of corporations” (van der Merwe & Al Achkar 2022: e12–e15). Finally, scholars also emphasise the context of digitalization as an important consideration.

3.2  The concept of corporate digital responsibility (CDR) According to my definition, the concept of corporate digital responsibility (CDR) involves different elements that require different assessments. This section attempts to answer the question of how a systematic conceptualization of corporate digital responsibility (CDR) can succeed. Building on the idea of business ethics as responsibility ethics, the concepts of corporate digital responsibility (CDR) refer a) to the actions of companies and their morally relevant effects on others, b) to a specific context of these actions, which is described as “digitization” or “digital transformation”, and c) to the respective sectors, because the configuration of responsibilities and their relevance may differ in the (digital) media sector (Altmeppen & Filipović 2019) from, for example, the responsibilities in the (digital) automotive, financial market or medical sectors. In this respect, I propose to conceptualize corporate digital responsibility (CDR) as being shaped by the following three fields within ethics:     i. Business ethics as ethics of responsibility   ii. Digital ethics iii. Domain Ethics The first two areas are necessary for any concept of CDR, insofar as they relate to the key questions of “How do I understand business ethics as responsibility ethics?” and “How do I  understand digital ethics?” Answers to these questions can differ and draw on philosophical research as well as practical experience. The respective approaches and their combination can be very complex. Moreover, if specific fields of economic activity are also considered for CDR, such as the media, medicine, banking, mobility, energy, etc., the situation becomes even more complicated. For example, the transformative dynamics of the medical sector (including, among others, health data and remedies) are immense and lead to a wide variety of highly relevant social and individual problems that must also be considered for CDR in the medical sector. Medical ethics should then be used as a source for this, which, combined with business ethics and digital ethics, results in an appropriately comprehensive concept. In this respect, CDR combines various fields of applied ethics in a sophisticated manner. The complexity of the concept is no coincidence: it reflects the 424

Corporate digital responsibility

complexity of our world and can thus be a means of providing ethical support for problems of societal development.

3.3  Digital ethics Corporate digital responsibility (CDR) builds not only on corporate ethics, as already explained, but also on digital ethics, which this section will now enlarge on. Already in the 1980s, Richard Mason presented digital ethics as some form of information or computer ethics, when he said: “Our moral imperative is clear. We must ensure that information technology, and the information it handles, are used to enhance the dignity of mankind” (Mason 1986: 11). Digital ethics takes as its main task the critical reflection of the effects of digital technologies. Since digital ethics is fundamentally based on technology ethics and media philosophy, different concepts of digital ethics differ from each other insofar they are also based on different paradigms in technology ethics or media philosophy. Examples of different paradigms can be found in the work of Luciano Floridi (e.g., Floridi 1999, 2014), Charles Ess (Ess 2014), Rafael Capurro (e.g., Capurro 2006, 2017), Shoshana Zuboff (Zuboff 2019), and Christopher Koska (Koska 2023). Nevertheless, scholars converge on the view that digital technologies not only change possibilities of action and evoke specific moral problems (for example, problems of privacy, public communication, or the future of work), but that these technologies also change human self-understanding in a fundamental way, as it is also discussed in relation to questions surrounding transhumanism and posthumanism (Puzio 2022). At all levels, digital technologies raise questions about the good and just life of human beings, alongside questions of sustainability and climate justice, which concern human flourishing as well. Given the variety of different philosophical approaches and the dynamic development of technologies, it is key to clarify what digital ethics ought to focus on and take as its object, and also how digital ethics could thereby incorporate a range of critical perspectives beyond narrow disciplinary boundaries. Whittlestone and colleagues can help here. In their literature review, they examine the “ethical and societal implications of algorithms, data, and artificial intelligence” (Whittlestone et al. 2019). The term “implications” is deliberately chosen to convey the idea that we need to look beyond a set of seemingly obvious “negative effects” of digital technologies and also take into account “changes” in a more general and morally neutral sense. Doing so is important because it is hard to predict from the outset which developments of digital technologies, with their immense innovative dynamism, ultimately turn out to be “good” or “bad” (Whittlestone et al. 2019: 6–7). Therefore, Whittlestone and colleagues are concerned with a wider range of potential consequences or tensions among different values, as they may arise in the course of the development, implementation, use, and constant changes of digital technologies. Before briefly explaining these tensions, let me address the concept of “digitality” that sits at the center of ethical considerations in this context. As has already become apparent, the term “digitization” is a fuzzy description for a range of technologically mediated phenomena that require new ethical thinking. Whittlestone and colleagues speak of ADAbased technologies “to capture a broad range of ethically and societally relevant technologies based on algorithms, data, and AI, recognizing that these three concepts are not totally separable from one another and will often overlap” (Whittlestone et al. 2019: 2). I find this account very helpful but I shall retain the term “digital technologies” (and “digitization” 425

Alexander Filipović

for their implementation) for simplicity here. It is essential to add, however, that technologies themselves do not determine our practices and actions. Rather, there is a complex reciprocal relationship between technology, societal practices, and action. It is also important to note that certain societal practices already existed before some technological developments arose and then became intertwined with them, such as the recording of health data to improve individual performance. The aforementioned ADA (algorithms, data and AI)-based technologies lead to various tensions in today’s world, which can be conceptualized as value conflicts or ethical challenges in general. Whittlestone et al. list as central tensions: • “Using algorithms to make decisions and predictions more accurate versus ensuring fair and equal treatment. • Reaping the benefits of increased personalization in the digital sphere versus enhancing solidarity and citizenship. • Using data to improve the quality and efficiency of services versus respecting the privacy and informational autonomy of individuals. • Using automation to make people’s lives more convenient versus promoting self-actualization and dignity” (Whittlestone et al. 2019: 27). These tensions affect not only companies but also all other agents that use ADA-based technologies, including the state and civil society organizations. However, for companies, the challenges from these new technologies become especially apparent. Thus, the arising tensions cannot simply be eliminated but must be directly and responsibly addressed.

3.4  Corporate Digital Responsibility (CDR) as a response to new ethical and social implications of digitalization This chapter asks how and why digital technologies necessitate a new way of thinking about corporate responsibility (CR). Lobschat et al. offer three reasons as to why a new form of corporate responsibility is needed, namely because of • “exponential growth in technological development, • malleability of technologies and data in use, and • pervasiveness of technology and data” (Lobschat et al. 2021: 876). All three aspects illustrate the uncertainties which companies face in the context of digital technologies and, on these grounds, Lobschat et al. (like Elliott et al. 2021; van der Merwe & Al Achkar 2022: e12–e15) strongly argue for a new and independent concept of corporate digital responsibility (CDR), rather than for a substantive extension of the already existing concept of corporate social responsibility (CSR), which Herden et  al. support (Herden et al. 2021). The exact relationship between general corporate responsibility (CR) and corporate digital responsibility (CDR) depends on whether digitization is seen as a quantum leap or “merely” as a gradual development (see also Mihale-Wilson et  al. 2022: 127–129). Considering the advances in artificial intelligence from around 2020 onwards, it seems plausible that there will be entirely new responsibilities and thus the need for a genuinely distinct approach to corporate responsibility too (also Elliott et al. 2021). 426

Corporate digital responsibility

4  How Corporate Digital Responsibility (CDR) works in practice 4.1  Normative requirements The concept of responsibility itself does not yet include any criteria as to why responsibility is and should be assumed. Nor does it provide guidance on substantial moral questions. This is also true in the context of corporate digital responsibility (CDR). CDR only states, in a formal rather than substantive sense, that those who are affected by corporate conduct in certain ways have a right to demand a justification or explanation from the respective companies. On some proposals, however, companies are expected to cover certain aspects and explain, for instance, how they respect human dignity or how they make sure to act sustainably. But the literature on CDR still rarely features specific moral norms. Rather, its main focus rests on the first sense of “acting responsibly”, mentioned at the beginning of this chapter, namely the requirement to act in accordance with moral norms in general, irrespective of which exact norms there are. Yet, I claim that we ought to support the development of explicit moral norms in the areas of general business ethics, digital ethics, and other respective fields of action. A pharmaceutical company has different responsibilities and different value conflicts to deal with than a media company, as they represent not only different industries but also use different digital technologies. Thus, corporate digital responsibility (CDR) requires, not only a formal obligation to justify entrepreneurial action, but also the addition of various substantive moral norms, ranging from general moral orientations (human dignity and human rights), sustainability and the Sustainable Development Goals (United Nations 2016–2030), to concepts of citizenship, good work, privacy, and solidarity (see also Leipziger 2017).

4.2  Agents and their areas of responsibility The use of ADA (i.e., algorithms, data and AI)-based technologies affects not only companies, but also governments, various forms of organizations, and almost all individuals. They all act in the context of digital technologies and have their specific yet potentially also overlapping areas of responsibility. Governments, states and supranational regulation remain important as they often set the legal framework for digital technologies. Yet, companies remain the key driving force behind new development (Koska & Filipović 2019). For this reason, companies are also the key bearers of responsibility in a global capitalist system. This is not to say, of course, that policymakers and lawmakers are absolved of their respective responsibilities. They also face increasing responsibilities. Yet, they need to acknowledge the role of companies and find ways of responsible interaction with the corporate world. What is more, ADA-based technologies also necessitate a change in the concept of responsibility. Earlier, I argued that the responsibility of companies, as well as their status as moral agents, is essentially linked to their capacity to engage in discourse and communication. In the age of automation, AI systems may obtain a similar status, assuming that they become able to participate in such discourse and communication too. But if so, non-human agents can be among the bearers of responsibility (Coeckelbergh & Loh 2020). Lobschat and colleagues suggested exactly that: We suggest that CDR goes beyond that point, to include guidelines for the development and deployment of artificial and technological actors in an organization. 427

Alexander Filipović

Algorithmic decision-making, machine learning, and AI involve non-human and nonsocial entities, so a key question is whether and how we can delegate digital responsibility to artificial actors and take responsibility for their actions. (Lobschat et al. 2021: 878)

4.3  Practical implementation of corporate digital responsibility (CDR) But why would some agents acknowledge or even seek their digital responsibility, given that responsibility is a burden? In the case of companies, the motivation could be either moral (i.e., based on a commitment to moral values), legal (i.e., based on the determination for legal compliance), or economic (i.e., based on the prospect of higher corporate profits through responsible action). However, questions about the motivations behind taking responsibility seriously are not new. They also arise in the context of general corporate responsibility (CR) and corporate social responsibility (CSR). To make sure that digital responsibility is supported and implemented in practice, it can help to develop CDR into a broad movement so that more companies are incentivized to introduce CDR into their strategies. Social pressure from the public, customers, shareholders, respective employees and in general, the company’s stakeholders (van der Merwe & Al Achkar 2022) could play a vital role here. Particularly in Europe (and especially in Germany), there are already several initiatives working in this direction, including awards for corporate digital responsibility (CDR) or cross-industry collaborations. Lobschat et al. see the establishment of a CDR corporate culture as the key to implementing digital responsibility. They claim that there are three fundamental levels of CDR culture (Lobschat et  al. 2021: 880–883). These layers are (1) shared values, (2) specific norms, and (3) artifacts and behaviors. The particular type of CDR culture relates to the strategic framework of digital responsibility within an organization. It encompasses presumptions and shared values (layer 1) from which specific CDR norms are deduced (layer 2), which then leads to specific artifacts and behaviors related to CDR (layer 3). Corporate digital responsibility (CDR) culture can be effective and a broad CDR movement is highly desirable. But in reality, companies still often act irresponsibly and even barely comply with applicable law, devoting much of their attention to the search for loopholes. In this context, van der Merwe et al. point to the power struggle that often takes place between government regulation and entrepreneurship: As governments reckon with the continued digital transformation and look for new ways of retaining their control and power, tech companies will continuously push against regulations while lobbying for concessions or reach settlements that do not impact their profit margins. Concurrently, tech companies may pursue strategies to lobby governments to impose certain regulations to undermine a competitor. (van der Merwe & Al Achkar 2022: e12–e19)

5 Conclusion The need to demand and promote Corporate Digital Responsibility (CDR) stems from the enormous transformative potential of general computer technologies (processors, memories, networks, sensors  .  .  .), data processing, algorithms, and artificial intelligence technologies. Digitalization, understood in this way, is characterized by its profound impact on 428

Corporate digital responsibility

the whole world, the exponential growth of technological development, the malleability of technologies and data used, and their ubiquity. Companies have a crucial role in this transformation and, therefore, bear responsibility. Corporate digital responsibility (CDR) is a concept that identifies companies as moral agents and caring responsibility bearers, develops moral principles, norms, and values within the framework of digital ethics, and spells out such principles, norms, and values in specific corporate contexts. CDR involves the voluntary integration of social and environmental concerns into the economic actions of companies in the context of digital opportunities. CDR requires that companies develop the capacities to evaluate actions in the context of digital opportunities, adhere to moral norms, identify and settle value conflicts, and justify their conduct to all stakeholders with an open mindset. So understood, digitally responsible corporate action means not only complying with legal requirements but also creating social, ecological, and economic added value beyond mere legal compliance.

Acknowledgments I would like to thank Alexandra Palkowitsch for essential support in the preparation of this text.

Further reading Corporate Digital Responsibility (CDR) is a new concept; therefore, there is no canonical literature yet. Insofar as the paper argues for combining corporate ethics and digital ethics, the first recommendations go in these two directions. A good introduction to the question of corporate responsibility (CR) is Blowfield, M., & Murray, A. (2019). Corporate Social Responsibility (4th ed.). Oxford and New York: Oxford University Press. A good introduction to the questions of digital ethics is Floridi, L. (2014). The 4th Revolution: How the Infosphere Is Reshaping Human Reality. Oxford: Oxford University Press. Zuboff, S. (2019). The Age of Surveillance Capitalism: The Fight for a Human Future at the New Frontier of Power. New York: PublicAffairs carefully examines new economic models and vividly and critically describes the transformative power of digital enterprises. For literature on the concept of CDR, a helpful first entry point is Dörr, S. (2021). Corporate Digital Responsibility: Managing Corporate Responsibility and Sustainability in the Digital Age. Berlin: Springer. The book combines the topics of sustainability and digital responsibility and is therefore highly relevant. Otherwise, the research literature on CDR takes place in journals. I recommend the comprehensive and explanatory paper Lobschat, L. et al. (2021). “Corporate Digital Responsibility,” Journal of Business Research, 122, 875–888, which also lists other recent literature on this topic.

References Altmeppen, K.-D., & Filipović, A. (2019). “Corporate Digital Responsibility. Zur Verantwortung von Medienunternehmen in digitalen Zeiten,” Communicatio socialis, 52(2), 202–214. Bazargan-Forward, S., & Tollefsen, D. (eds.). (2020). The Routledge Handbook of Collective Responsibility. New York: Routledge. Beauchamp, T. L. (2003). “The Nature of Applied Ethics,” in R. G. Frey & C. H. Wellman (eds.), A Companion to Applied Ethics (pp. 1–16). Oxford, UK: Blackwell Publishing Ltd. Beauchamp, T. L.,  & Childress, J. F. (1979). Principles of Biomedical Ethics. New York: Oxford University Press. Blowfield, M., & Murray, A. (2019). Corporate Social Responsibility (4th ed.). New York and Oxford: Oxford University Press. Buddeberg, E., & Hecker, A. (2018). “Justification Incorporated: A Discursive Approach to Corporate Responsibility,” Ethical Theory and Moral Practice, 21(3), 465–475.

429

Alexander Filipović Capurro, R. (2006). “Towards an Ontological Foundation of Information Ethics,” Ethics and Information Technology, 8(4), 175–186. ———. (2017). Homo Digitalis. Beiträge zur Ontologie, Anthropologie und Ethik der digitalen Technik. Wiesbaden: Springer VS. Carroll, A. B. (1979). “A Three-Dimensional Conceptual Model of Corporate Performance,” Academy of Management Review, 4(4), 497–505. Coeckelbergh, M., & Loh, J. (2020). “Transformations of Responsibility in the Age of Automation: Being Answerable to Human and Non-Human Others,” in B. Beck & M. Kühler (eds.), Technology, Anthropology, and Dimensions of Responsibility (pp. 7–22). Stuttgart: J. B. Metzler. Dörr, S. (2021). Corporate Digital Responsibility: Managing Corporate Responsibility and Sustainability in the Digital Age. Berlin: Springer. Elliott, K., Price, R., Shaw, P. et al. (2021). “Towards an Equitable Digital Society: Artificial Intelligence (AI) and Corporate Digital Responsibility (CDR),” Society, 58(3), 179–188. Ess, C. (2014). Digital Media Ethics (2nd, rev. a. updated ed.). Cambridge and Malden, MA: Polity Press. Floridi, L. (1999). “Information Ethics: On the Philosophical Foundation of Computer Ethics,” Ethics and Information Technology, 1, 37–56. ———. (2014). The 4th Revolution: How the Infosphere Is Reshaping Human Reality. Oxford: Oxford University Press. Forst, R. (2002). Contexts of Justice: Political Philosophy Beyond Liberalism and Communitarianism. Berkeley and Los Angeles: University of California Press. ———. (2012). The Right to Justification: Elements of a Constructivist Theory of Justice. New York: Columbia University Press. Freeman, R. E. (1984). Strategic Management: A Stakeholder Approach. Boston, MA: Pitman. French, P. A. (1979). “The Corporation as a Moral Person,” American Philosophical Quarterly, 16(3), 207–215. ———. (1995). Corporate ethics. Fort Worth: Harcourt Brace College Publishers. Frey, R. G., & Wellman, C. H. (eds.). (2003). A Companion to Applied Ethics. Oxford, UK: Blackwell Publishing Ltd. Herden, C. J., Alliu, E., Cakici, A. et al. (2021). “Corporate Digital Responsibility: New Corporate Responsibilities in the Digital Age,” NachhaltigkeitsManagementForum, 29(1), 13–29. Jonas, H. (1984). The Imperative of Responsibility: In Search of an Ethics for the Technological Age. Chicago and London: University of Chicago Press. Koska, C. (2023). Ethik der Algorithmen. Auf der Suche nach Zahlen und Werten. Berlin, Heidelberg: J.B. Metzler. Koska, C., & Filipović, A. (2019). “Blackbox AI – State Regulation or Corporate Responsibility?” Digital World, 3(4), 28–31. Leipziger, D. (2017). The Corporate Responsibility Code Book (3rd ed.). Boca Raton, FL: Routledge. Lobschat, L., Mueller, B., Eggers, F. et al. (2021). “Corporate Digital Responsibility,” Journal of Business Research, 122, 875–888. Mason, R. O. (1986). “Four Ethical Issues of the Information Age,” MIS Quarterly, 10(1), 5. Mihale-Wilson, C., Hinz, O., van der Aalst, W., & Weinhardt, C. (2022). “Corporate Digital Responsibility: Relevance and Opportunities for Business and Information Systems Engineering,” Business & Information Systems Engineering, 64(2), 127–132. Mueller, B. (2022). “Corporate Digital Responsibility,” Business & Information Systems Engineering, 64(5), 689–700. Neuhäuser, C. (2011). Unternehmen als moralische Akteure. Berlin: Suhrkamp. Puzio, A. (2022): Über-Menschen. Philosophische Auseinandersetzung mit der Anthropologie des Transhumanismus. Bielefeld: Transcript Verlag. van der Merwe, J., & Al Achkar, Z. (2022). “Data Responsibility, Corporate Social Responsibility, and Corporate Digital Responsibility,” Data & Policy, 4, e12–1–12. Whittlestone, J., Nyrup, R., Alexandrova, A., Dihal, K.,  & Cave, S. (2019). Ethical and Societal Implications of Algorithms, Data, and Artificial Intelligence: A Roadmap for Research. London: Nuffield Foundation. Zuboff, S. (2019). The Age of Surveillance Capitalism: The Fight for a Human Future at the New Frontier of Power. New York: PublicAffairs.

430

34 RECKLESS COMPLICITY International banks and future climate Henry Shue*

I am responsible for what I do, and you are responsible for what you do. But on any credible view I need to give attention, in what I do, to what you will do in consequence. And you need to give attention, in what you do, to what I will do in consequence. In that sense, there are two parts of morality. There is what I should do simpliciter, and then there is what I should do by way of contribution to what you do. If I fail in the first I am a principal. If I fail in the second I am an accomplice. . . . For my own actions inevitably include my actions of contributing to your actions. (Gardner 2007: 132)

Introduction I want to uncover some philosophical issues embedded deep in contingent empirical features of the current global energy regime. Corporations that persist in exploring for and extracting additional supplies of fossil fuel are accumulating additional ammunition for the disruption of the climatic constituents of the physical preconditions for civilized society, thereby worsening the risk that weather phenomena on this planet will become increasingly inhospitable to peaceful and cooperative human social life (Kemp et  al. 2022). Fossil fuel companies are among the principal evil-doers in this arena (Ekwurzel et al. 2017; Shue 2017), while governments consistently fail in their duty to rein in the corporate harm (Speth 2021). Indeed, the largest fossil fuel companies are governmentowned and -controlled (Grasso 2022). However, loans to fossil fuel companies, both the investor-owned ones  – like ExxonMobil, Shell, and TotalEnergies  – and the stateowned ones – like Saudi Aramco, China National Petroleum Corporation [CNPC (PetroChina & SinoPec)], and Gazprom – from banks (both investor-owned and state-owned) enable continuing exploration for and extraction of more and more fossil fuels. The decisions to grant these loans may constitute wrongs of complicity at least equally as serious as the fossil fuel companies’ wrongs. This is the judgment whose supporting arguments I will explore here.

431

DOI: 10.4324/9781003282242-46

Henry Shue

Why searching now for additional fossil fuels to extract is wrong It has been perfectly clear for years that it would be destructive to the planet’s climate to continue to use fossil fuels for unabated combustion as in the past and that large percentages of the reserves of fossil fuel already identified for extraction must instead be left in the ground (or extracted only after techniques for carbon removal from the atmosphere have, if they ever do, become mature technologies sustainable for the long-term at large scale).1 A widely influential 2015 article, adopting the relatively lax goal of keeping average global temperature rise to 2° C, concluded “that a third of oil reserves, half of gas reserves and over 80 per cent of current coal reserves should remain unused from 2010 to 2050” (McGlade & Ekins 2015). It is now evident that the requirements for avoiding the combustion of existing reserves are much more stringent, especially if the temperature rise is to be kept to 1.5° C, as it ought to be (Allen et al. 2018). This is not the place to amass the abundant empirical evidence for the greater stringency, but a startlingly illustrative example is the precise specification recently of the largest 495 extraction projects in the world already underway or planned, labelled ‘carbon bombs’, which, combined, would release double the amount of CO2 still compatible with restricting the temperature rise to 1.5° C (Kühne et al. 2022). And since this is also not the place for engineering details, I shall mostly refer simply to ‘extraction’ as shorthand for the prior exploration required to find and confirm the fossil fuel reserves (which is damaging to the areas explored), the actual extraction from under the earth or the sea by well-drilling (which is further damaging and polluting), and the subsequent transport into the global distribution system that, as we shall see, involves extensive and long-lasting infrastructure like pipelines that are once again destructive of land and often polluting of rivers, lakes, and ground-water. In the current circumstances in which it is urgent to bring carbon emissions to net zero by 2050, which requires cutting them in half by 2030, launching yet more projects to find additional reserves of fossil fuels is at the very best a wildly irrational waste of resources like socially valuable engineering and managerial expertise and a cause of pointless additional pollution of air, land, and water – thus wrong, firstly, as wasting valuable social resources and, secondly, as further damaging valuable natural resources. The ready availability of additional fossil fuels would substantially increase the risk that the climate will be catastrophically undermined by their unabated combustion, and virtually guarantees that the climate will be more seriously undermined than already fated to be by the combustion of the fossil fuels already being exploited – thus wrong, thirdly, as worsening the dangers threatening the security and lives of current and future generations (Shue 2021). The extraction of additional fossil fuel increases the risk that more fossil fuel will be burned in ways that emit CO2. Because much CO2 released into the air by combustion remains in the atmosphere for centuries, and some persists for millennia (Ciais et al. 2013: 472–473), additional extraction imposes exacerbated risks on numerous future generations. This is the case, as we will see shortly, unless the CO2 is promptly extracted from the atmosphere with experimental techniques that are in their infancy and may or may not ever mature into sustainable technologies at scale. An especially egregious case of the first wrong of wasting valuable social resources is that persisting pursuit of extraction consumes corporate resources needed to fulfill legally mandatory, but in fact totally ignored, asset retirement obligations (AROs) to seal abandoned oil and gas wells, now egregiously leaking methane, that will otherwise have to be sealed at enormous public expense (Schuwerk & Rogers 2020) – thus wrong, fourthly, as ruthlessly 432

Reckless complicity

profiting by imposing further great expense on people generally through aggressive externalization of corporate costs. ARO for a single deep fracked well, for example, can cost $1 million, and more fracking wells are being drilled every day. In 2022 proponents of expanded extraction and its supporting infrastructure were claiming that the disruption in the energy markets resulting from the Russian attack on Ukraine and the sanctions placed against Russia in response meant that as temporary emergency measures new extraction and infrastructure are not only permissible but obligatory as part of the defense of Ukraine. This is, fifthly, wrong as attempted deception (Halper 2022b). Most extraction and infrastructure projects take several years to complete, thus providing additional supplies only several years from now, and would then operate for multiple decades – well beyond 2050 when net carbon emissions must be zero. Renewable energy can be added more quickly and more cheaply and will be sustainable.

How banks are complicit in the fossil fuel companies’ wrongs Thomas E. Hill, Jr. can be counted upon to have provided a crisp but jargon-free account of complicity: The paradigm is a direct and substantial contribution to wrongdoing, made freely with acceptable options available and with the knowledge that one is so contributing. Aside from being a willing and known contribution to a wrong, the act in question may not itself be wrong. (Hill 1979: 87) The contribution of bank loans to exploration for and extraction of fossil fuels is entirely direct. How substantial is it? Energy projects for which bank loans cannot be arranged must be abandoned, as business is ordinarily conducted. Of course, in theory an oil company like Shell could save its own pennies until it had enough funds put aside in its piggy-bank to finance the next exploration and extraction project that appeals to it. This is, however, never how the international energy business operates, for many reasons, including the facts that it would mean forgoing the financial leverage that comes from operating with other people’s money as well as one’s own, that it would greatly slow down the ruthlessly competitive hunt for additional assets, and that it would require fewer stock buy-backs, lower executive remuneration, and smaller dividends in order to save for new extraction projects. Obviously, then, bank loans are not necessary in a strict logical sense. One can easily imagine possible worlds in which the energy business is conducted differently. No ‘but-for’ test of causal contribution would be satisfied by bank loans. And no particular bank loan would satisfy even the test of being a necessary element in one of perhaps many possible sufficient sets of conditions. Some other bank might have made the required loan in those same circumstances. Philosophers can become fascinated by questions about ‘necessary condition’ and ‘sufficient condition’ and their role, if any, in the explication of causal contribution. I believe, however, that while these notions are at home in the logical world, they are ill at ease and quite unhelpful in the social world. Chiara Lepora and Robert E. Goodin’s study noted a crucial feature of complicity: “Complicity is not an all-or-nothing matter. Instead, it is a graded moral notion. Its badness 433

Henry Shue

comes in degrees” (Lepora & Goodin 2013: 130). Since complicity comes in degrees, complicity must be a scalar notion, not a binary notion like necessity and sufficiency.2 One can strain toward some qualified sense of ‘necessary’ and ‘sufficient’: Must the cause be a necessary condition, a sufficient condition or a necessary member of a set of conditions that are together sufficient for the outcome? In law these terms, much discussed in the philosophical literature, are interpreted as meaning “necessary or sufficient in the particular circumstances in issue.” (Honoré & Gardner 2019: Sect. 3) But I am not sure that qualified necessity, or qualified sufficiency, make sense; something with qualified necessity strikes me as something that is not quite necessary. In any case, we do much better, I  think, to follow Hill and simply ask in commonsense terms, was the contribution “direct and substantial”? Might these oil wells have been drilled without this loan from this bank? Sure – it is certainly possible. But if this loan at this time enabled this fossil fuel company to conclude that this project was feasible and to launch it, this loan made a direct and substantial contribution to this project. This loan enabled this project to move forward at this time. Some other loan might have done the same, but this loan is the one that did it in fact. This loan, then, is heavily responsible for the launch of this extraction project. The useful question to ask about a contribution to a wrong is not “is it necessary or not”, and not “is it a necessary member of a sufficient set or not”, but: how much of a contribution was it in reality? For most extraction projects, for example, a pool of banks is assembled by one or more ‘lead’ banks to finance the project jointly. The lead arrangers thereby make more of a contribution in creating the coalition than the other banks – their organizational role is central. Their role may be, in the ordinary senses, crucial and vital – not strictly necessary, but in fact key. The lead banks make an important difference. If their leading role constitutes complicity in a wrong, their complicity may be greater organizationally. Other banks in the pool obviously make more or less substantial contributions by the amount of money they provide and so may be more or less key financially. As an example consider one of the most horrific fossil fuel projects in the world currently undergoing financing: the East Africa Crude Oil Pipeline (EACOP). This nightmare pipeline will begin in Uganda at the Tilenga Oilfield near Lake Albert – and partly inside Murchison Falls National Park  – run through the basin of Lake Victoria (Africa’s largest lake, serving over 40  million people), cross 200 rivers, displace thousands of native villages, and meet supertankers on the Tanzanian coast of the Indian Ocean, after 1,443 kilometres (Hook 2022). Because the oil from Tilenga is too thick to flow otherwise, the entire pipeline will constantly be heated to 50° C, making this by far the longest heated oil pipeline ever constructed (StopEACOP 2022). The majority owners (70%) of the pipeline are French carbon major TotalEnergies (Total) and China National Offshore Oil Corporation (CNOOC), with the minority ownership of 30% split evenly between Uganda and Tanzania, who are providing the pipeline company both a ten-year corporate tax holiday and a VAT exemption. The leading banks arranging the financing are Standard Bank (South Africa – through its subsidiary Stanbic Uganda), Industrial and Commercial Bank of China (ICBC), and Sumitomo Mitsui Bank (SMBC). A  number of banks have announced that they will refuse to make a project specific loan for this looming human and ecological disaster, and then joined in a massive general purpose loan – “revolving credit facility” – of 434

Reckless complicity

$8 billion to TotalEnergies itself (BankTrack 2022). The banks trying to have it both ways in this duplicitous manner include JPMorgan Chase (Tett 2010), Barclays, BNP Paribas, Crédit Agricole, and Société Générale; the hypocrisy here is staggering. EACOP, for instance, is designed to function for a minimum of 25 years (StopEACOP 2022). Since it cannot be completed for a few years from now, its lifetime will run past 2050, the year by which climate scientists believe carbon emissions must reach net zero (a goal to which all the banks providing Total with the revolving credit facility loudly claim to be committed). CNOOC and Total will need to re-pay any project specific loans that Standard Bank, ICBC, and SMBC can arrange directly for the pipeline, and Total will need to satisfy the terms of its revolving credit facility from the pool of other banks that pretend not to support the pipeline. So by making either kind of loan banks lock-in additional financial pressures to keep the oil flowing through the pipeline for at least a quarter of a century, which is in effect pressure to raise the risks of disaster for future generations.

How banks can try to defend their loans for additional fossil fuel The banks can offer several kinds of defenses for their policy of continuing to loan for additional extraction in spite of the worsening climate change highly likely to be caused by combustion of what is extracted and the other kinds of wrongful damage listed earlier. Some of these defenses are worth only a passing reference, but one becomes of philosophical interest. First, any given bank may of course attempt to excuse its loan with what John Gardner dubbed the “arms dealers’ defence” (Gardner 2007): if we don’t do it, someone else will. One might also call this cynical fatalism: what we do makes no difference (so we might as well do as we please). We cannot be held responsible for a causal contribution to a harmful outcome because our contribution was not necessary. Any initial effects of our actions on others fade away in the glare of everything else that is happening, leaving no identifiable shadow on the others. Meanwhile, we earn some interest and keep our bank growing. This may strike many as more of an excuse than a justification, and it obviously rests on the controversial presupposition that a contribution is significant only if it is necessary. One admirable line of response, offered by both John Gardner and Tom Hill, is that one ought not to engage in wrongs oneself even if others otherwise will – expected effects are not normally to be decisive. If, on the other hand, expected effects are to be taken to be decisive, a particular difficulty for this notorious move here is its resting on an empirical hypothesis that is simply false in the case at hand. Will enough other banks make loans, and will whether they do be affected by whether this given bank does? In the case of EACOP activists around the globe are campaigning to convince banks not to join the pool funding the pipeline and to shame those who do. At this writing the issue is wide open. A number of major banks have publicly declared the pipeline to be irresponsible and refused to fund it (although, as indicated previously, some of them have turned around and provided a general line of credit to one of the pipeline owners, Total). Activists in France have meanwhile sued Total for rights violations (Hook 2022). In a case like this, one bank might be either the one to put the funding for the pipeline over the top, eliminating the last chance to stop it, or the one to keep the funding short – in sum, the one to make the difference. The arms dealers’ defense is inappropriate to the actual situation, which this defense mistakenly assumes is a settled outcome, when the outcome here is open to influence. In fact, the 435

Henry Shue

shadow of a bank’s choice to support the project may endure as a lost opportunity to act decisively to prevent several calamitous harms. A second defense attempts, in effect, to buttress the first into more of a justification by providing moral grounding for its indifference to harmful effects on other people and the shared environment. One might call this the justification of fiduciary exclusivity. This contention is that it is not determinative if the decision to provide the loan negatively affects people unaffiliated with the bank because the bank, like any capitalist enterprise, has responsibilities only to its shareholders. This exclusive internal responsibility entails that it is not merely permissible to take into account only the interests of the shareholders, but obligatory to do so. Failing to maximize the interests of the shareholders within the limits of the law is a betrayal of the trust at the heart of the enterprise by its management. In the US at present this simplistic proposition is highly influential politically and is frequently declaimed as if it were a self-evident truth. When it is taken to need any further grounding, its basis may be a kind of crude version of the invisible hand: if all firms (and individuals and families and nations) simply pursue their own interest, the common interest will be best served. Unfortunately, this empirical hypothesis is notoriously false in the specific case of the common interest in protecting the environment (Sagoff 1988). A slightly more sophisticated basis is that if firms are to be expected to concern themselves with, in addition to profit, the well-being of the environment or the health and wellbeing of persons other than shareholders, it is the prior responsibility of the government to impose such requirements through regulation and to impose them uniformly on all firms. However, even if such a division of labor on which company directors maximize profits and government officials regulate to correct ‘market failures’ like environmental destruction were justified in economic ideal theory, it is hopelessly irrelevant to the real world in which corporate fossil fuel interests have politically captured major regulators (Mildenberger 2020; Stokes 2020) and can emit greenhouse gases at will and without sanction. In fact, the legislatures of several individual states in the US have chosen to take the offensive against proponents of climate action and begun to pass laws penalizing all enterprises – including banks – that choose to limit in any way the interests of shareholders in order to protect the climate. The State of West Virginia has gone to the extreme of banning from doing any business with the state government banks that have refused to make loans for coal projects – including some of the same banks providing the credit line to the TotalEnergies oil company; and the state governments of West Virginia, Louisiana, and Arkansas have withdrawn state funds from management by BlackRock on the fanciful grounds that BlackRock gives too much weight to the environment rather than to the interests of the investors whose funds it manages (Gelles 2022). Now, it is by no means clear that these state officials are acting on the basis of a considered and principled view about fiduciary responsibility rather than merely promoting the narrow interests of politically powerful corporations within their own territory. An appeal to something like fiduciary exclusivity would, however, be the only way I can see to try to give some kind of principled defense that might justify their policy of hostility to firms that choose to exercise responsibility toward the environment that may not be legally mandated. But while I think it is important to realize that some such vaguely articulated view is rhetorically potent at present, any position asserting that corporations are morally bound to pursue only self-interest constitutes either the rejection out of hand of social responsibility; the embrace of the mistaken empirical hypothesis that if each agent pursues self-interest, 436

Reckless complicity

common interests will reliably advance; or the appeal to an ideal economic theory that ignores political reality. A third defense of continuing bank loans for additional extraction of fossil fuels, however, does raise a significant philosophical issue about the imposition of risk on future generations. This defense can be called the one-sided gamble, and it rests on a contention to which I have already alluded a couple of times. The combustion of fossil fuels would cease to worsen climate change if either the carbon emissions are prevented from escaping into the atmosphere by some version of Carbon Capture and Storage (CCS) or the escaped emissions are retrieved by some version of Carbon Dioxide Removal (CDR). Obviously much turns on empirical issues about these immature technologies that we cannot pursue. However, whatever the prospects for the various versions of each of the two kinds of potential technology, a couple of empirical points are settled. First, CCS could approximate equivalence to the result of emissions eliminated by avoiding combustion of fossil fuel only to the degree to which it was 100% efficient in capturing both the emissions from the ultimate combustion and the emissions from the extraction, transport, and refining of the fuel. Otherwise the emissions from CCS are not net zero (Halper 2022a; Surapaneni 2022). Second, and very importantly, no conceivable form of CDR can possibly be equivalent to emissions not released. CO2 released into the atmosphere can certainly be removed from the atmosphere – trees and plants do it all the time, and some man-made technologies are succeeding in removing small amounts, currently at great expense (Rawnsley 2022). But throughout the years or decades during which the CO2 is in the atmosphere prior to its removal, it is driving climate change. CO2 released and CO2 removed can be net zero, but the climate after the CDR removes the CO2 will be worse, other things being equal, than the climate when the CO2 was released and worse than if the CO2 had not been emitted. If the quantity of CO2 is significant, it could have forced the climate to pass a tipping point that would not otherwise have been passed (Shue 2021: 23–26 and 111–113). The presence of the CO2 in the atmosphere is reversible, but the effects of its transient presence there may be irreversible. And an important philosophical point will emerge. Given the importance of bringing carbon emissions to net zero no later than 2050, additional extraction and additional associated infrastructure can be justified only if some form of either CCS or CDR will succeed in taking out of the picture a very high percentage – preferably 100% – of the additional emissions made possible by the additional extraction and additional infrastructure.3 Any attempted justification rests, then, on the following gamble: if banks proceed at present with enabling the fossil fuel business as usual, some other agents at some date in the near future will advance CCS and CDR adequately to capture all emissions meanwhile released (the highly effective CDR) and to prevent any further releases of emissions (the highly effective CCS). For banks to continue to finance additional extraction of fossil fuels is, then, for them to choose to increase the risk of severe climate change from additional emissions on the gamble that unspecified others will provide the CDR and CCS required to reverse the risk. This choice by the banks makes matters worse in at least three respects. Firstly, as I have just noted, even the achievement of this technologically highly optimistic scenario would not be equivalent to not having allowed in the first place any emissions later removed because, for as long as those emissions are in the atmosphere, they will be driving worsening climate change and possibly driving it beyond points of no return. Secondly, there is obviously the very real empirical possibility that the one-sided gamble will 437

Henry Shue

simply turn out badly because the CDR and CCS will never be affordable and sustainable at sufficient scale to deal with the scale of the emissions that will meanwhile have poured out from the continuing extraction. Both technologies may fail. Thirdly, however, a deeper, less obvious moral problem is created by the risk taken by the banks.4 Every risk involves “three roles, namely the risk exposed, the decision-maker and the beneficiary” (Hermansson & Hansson 2007: 131). If I bet $5 of my own money in hope of winning $25 for myself, I occupy all three roles: I am the decision-maker, the potential beneficiary, and the risk exposed. When a bank decides to pursue short-term gains (interest to be received and business-as-usual) for itself by enabling additional extraction of fossil fuels on the gamble that unspecified others will develop CDR or CCS at adequate speed and scope to prevent the additional emissions made possible by the additional extraction from leading to additional climate change, the decision-maker and the potential beneficiary are the bank itself. The risk exposed, by contrast, are all future persons in danger of suffering from worsening climate change if highly successful CDR and CCS are not, for whatever reason, in fact promptly developed and widely employed. For the sake of a relatively small potential benefit for themselves the bank imposes an unknown and unlimited risk of great danger upon countless future persons. If an individual risked a great danger of unknown probability for herself in pursuit of a small gain of unknown probability, we would judge her to be imprudent – indeed, perhaps a fool. If a bank imposes an analogous risk upon innumerable future persons, it is unjustifiable  – indeed, a moral outrage. If it would be imprudent to choose such a risk for oneself, it is wrong to impose it on others. The others in this case are future generations. They can only lose and have nothing to gain – hence, this is a one-sided gamble imposed by banks utterly without justification.

Conclusion It is evident that banks are complicit in the wrongs being committed by fossil fuel firms. With reference again to Tom Hill’s account of complicity, a bank’s loan for extraction or infrastructure is, first, exactly the kind of intervention into the course of events that can make a direct and substantial contribution to that kind of fossil fuel project. Second, such a loan is undoubtedly made freely with acceptable options available. Indeed, many of the other options open are vastly superior to underwriting the continuation of the fossil fuel energy regime (Garg 2022). Opportunities for loans to the building of an alternative global regime of alternative sources of energy abound. Some such loans may be riskier than some loans to an established giant oil company, although in the face of the acceleration of the deterioration of the climate, loans over multiple decades to unreconstructed fossil fuel companies that need to make radical changes in their business plans are unlikely to remain nearly as safe as they once were (Allen & Coffin 2022; Sanzillo & Mattei 2022). The banks could also be financing the research and development of the CDR (or, much less promisingly, CCS) that they now appear to be tacitly assuming unspecified others will finance. Third, the major international banks that make fossil fuel loans act with extensive knowledge, drawing on the sophisticated work of their own large internal research departments. They have no excuse whatsoever for not understanding the well-known dynamics of the climate change into which they are intervening and the necessity of a rapid cessation of unabated combustion of fossil fuel which they are delaying. Most especially, they can hardly fail to be aware of the great uncertainty about whether CDR or CCS, without at least one of which continued combustion of fossil fuels will accelerate and exacerbate 438

Reckless complicity

climate change, will be developed and implemented at global scale and sufficiently soon, particularly since they themselves are failing to invest in either. This uncertainty is uncertainty about whether the projects in which they are investing will make direct and substantial contributions to the undermining of the climate for future generations. Knowingly running a risk with an unknown probability of worsening destruction and disruption of a global extent, an unknown duration, and a currently unlimited seriousness is a paradigm of recklessness. Overall, a bank loan to a project like EACOP makes this planet a worse – specifically, a more dangerous – place for humans and many other living things, present and future. Governments should announce now that no bailouts will be provided in future to banks that gamble today on such risks. If there are any categorical moral imperatives, one is the imperative to protect the physical preconditions of civilized human society, which in turn is the social precondition of functioning individual morality. This is, then, a moral imperative to protect the social possibility of morality. Bank loans for the extraction of yet more fossil fuel that cannot safely be burned with current technology constitute complicity in wrongdoing that is reckless in its threat to the physical preconditions of civilized human society for future generations.

Further reading Cripps, E. (2022). What Climate Justice Means and Why We Should Care. London: Bloomsbury Continuum. A brief, accessible, and provocative introduction to moral issues concerning climate change. McKinnon, C. (2022). Climate Change and Political Theory. Cambridge: Polity Press. An advanced and comprehensive analysis of moral issues concerning climate change. Thompson, H. (2022). Disorder: Hard Times in the 21st Century. Oxford: Oxford University Press. A political and economic analysis of the role of the fossil fuel energy regime in recent international politics. Wenar, L. (2016). Blood Oil: Tyrants, Violence, and the Rules That Run the World. New York: Oxford University Press. An ethical critique of the role of the fossil fuel energy regime in recent international politics with recommendations for change.

Notes * For helpful comments I am grateful to Paula Casal, Clemens Pinnow, Michael Schefczyk, and the Editor. 1 Abated combustion would be combustion whose CO2 emissions were prevented from escaping by some form of Carbon Capture and Storage (CCS) technology, which may or may not ever be feasible at large scale – probably only if required by law. For advocacy of CCS, see CarbonTakeBack.org. 2 In spite of this insight they worry inordinately, and inconsistently, about whether contributions satisfy the but-for test and introduce unhelpful notions of ‘definitely essential’ and ‘potentially essential’ (Lepora & Goodin 2013: 60–63). 3 For a favorable account of CCS, see www.globalccsinstitute.com/. For more critical views of CCS, see (Mikulka 2019) and (Halper 2022a). For CDR, see (Shue 2021: 89–116) and (Temple 2021). 4 The wrongfulness of the choice to run this risk is explored more fully in (Shue 2021: 103–111). Here the analysis is extended to a new case. I contend there that all “decisions on mitigation [of emissions] in the present are decisions on risk distribution across generations” (108).

References Allen, M. R. et al. (2018). “Framing and Context,” in V. Masson-Delmotte et al. (eds.), Global Warming of 1.5° C (pp. 49–91). www.ipcc.ch/sr15/chapter/chapter-1/.

439

Henry Shue Allen, T., & Coffin, M. (2022). Unburnable Carbon: Ten Years On – The Financial Markets Are Still Enabling a Carbon Bubble. London: Carbon Tracker Initiative. https://carbontracker.org/reports/ unburnable-carbon-ten-years-on/. BankTrack. (2022). “12 banks lend $8 billion to oil and gas expansionist TotalEnergies,” May 13. www.banktrack.org/article/12_banks_lend_8_billion_to_oil_and_gas_expansionist_totalenergies. Ciais, P. et al. (2013). “Carbon and Other Biogeochemical Cycles,” in T. F. Stocker et al. (eds.), Climate Change 2013: The Physical Science Basis. Cambridge: Cambridge University Press. Ekwurzel, B., Boneham, J., Dalton, M. W. et al. (2017). “The Rise in Global Atmospheric CO2, Surface Temperature, and Sea Level from Emissions Traced to Major Carbon Producers,” Climatic Change, 144, 579–590. DOI: 10.1007/s10584-017-1978-0. Gardner, J. (2007). “Complicity and Causality,” Criminal Law and Philosophy, 1, 127–141. DOI: 10.1007/s11572-006-9018-9. Garg, V. (2022). “Ahead of COP27, ‘STEPS’ in the Right Direction but More Clean Energy Finance Needed,” Institute for Energy Economics and Financial Analysis, November 3. https://ieefa.org/ resources/ahead-cop27-steps-right-direction-more-clean-energy-finance-needed. Gelles, D. (2022). “How Republicans Are ‘Weaponizing’ Public Office Against Climate Action,” New York Times, August 5. www.nytimes.com/2022/08/05/climate/republican-treasurers-climatechange.html? Grasso, M. (2022). From Big Oil to Big Green: Holding the Oil Industry to Account for the Climate Crisis. Cambridge, MA: MIT Press. Halper, E. (2022a). “How a Pricey Taxpayer Gamble on Carbon Capture Helps Big Oil,” Washington Post, October 9. www.washingtonpost.com/business/2022/10/09/carbon-capture-oil-gas/. Halper, E. (2022b). “Fossil Fuel Projects Were Stalled a Year Ago. Now They’re Making a Comeback,” Washington Post, November  3. www.washingtonpost.com/business/2022/11/03/ fossil-fuel-cop27-russia/. Hermansson, H., & Hanson, S. O. (2007). “A Three-Party Model Tool for Ethical Risk Analysis,” Risk Management, 9, 129–144. Hill, T. E., Jr. (1979). “Symbolic Protest and Calculated Silence,” Philosophy  & Public Affairs, 9, 83–102. Honoré, A.,  & Gardner, J. (2019). “Causation in the Law,” in E. N. Zalta (ed.), The Stanford Encyclopedia of Philosophy (Fall 2019 ed.). https://plato.stanford.edu/archives/fall2019/entries/ causation-law/. Hook, L. (2022). “The Oil Giants Drilling Among the Giraffes in Uganda,” Financial Times, April 12. www.ft.com/content/e1670042-11bd-4c68-9bde-a599d94bd8c0? Kemp, L., Xu, C., Depledge, J. et al. (2022). “Climate Endgame: Exploring Catastrophic Climate Scenarios,” PNAS [Proceedings of the National Academy of Sciences of the United States of America], 119, 34. Published online August 1, 2022. DOI: 10.1073/pnas.2108146119. Kühne, K., Bartsch, N., Tate, R. D. et al. (2022). “Carbon Bombs” – Mapping key fossil fuel projects. Energy Policy, published online. DOI: 10.1016/j.enpol.2022.112950. Lepora, C., & Goodin, R. E. (2013). On Complicity and Compromise. Oxford: Oxford University Press. McGlade, C., & Ekins, P. (2015). “The Geographical Distribution of Fossil Fuels Unused When Limiting Global Warming to 2°C,” Nature, 517, 187–190. DOI: 10.1038/nature14016. Mikulka, J. (2019). “Stanford Study Says Renewable Power Eliminates Argument for Using Carbon Capture with Fossil Fuels,” DeSmog, November  21. www.desmog.com/2019/11/21/ jacobson-stanford-carbon-capture-fossil-fuels-renewables/. Mildenberger, M. (2020). Carbon Captured: How Business and Labor Control Climate Politics. London: MIT Press. Rawnsley, J. (2022). “Racing Against the Clock to Decarbonise the Planet,” Financial Times, August 11. https://channels.ft.com/en/rethink/racing-against-the-clock-to-decarbonise-the-planet/. Sagoff, M. (1988). The Economy of the Earth: Philosophy, Law, and the Environment. New York: Cambridge University Press. Sanzillo, T.,  & Mattei, S. (2022). “Oil and Gas Profits Driven by Ukraine Conflict, Not Financial Skill,” Institute for Energy Economics and Financial Analysis, November  1. https://ieefa.org/ resources/oil-and-gas-profits-driven-ukraine-conflict-not-financial-skill?

440

Reckless complicity Schuwerk, R., & Rogers, G. (2020). It’s Closing Time: The Huge Bill to Abandon Oilfields Comes Early. London: Carbon Tracker Initiative. https://carbontracker.org/reports/its-closing-time/. Shue, H. (2017). “Responsible for What? Carbon Producer CO2 Contributions and the Energy Transition,” Climatic Change, 144, 591–596. DOI: 10.1007/s10584-017-2042-9. ———. (2021). The Pivotal Generation: Why We Have a Moral Responsibility to Slow Climate Change Right Now. Princeton: Princeton University Press. Speth, J. G. (2021). They Knew: The US Federal Government’s Role in Causing the Climate Crisis. London: MIT Press. Stokes, L. C. (2020). Short Circuiting Policy: Interest Groups and the Battle Over Clean Energy and Climate Policy in the American States. New York: Oxford University Press. StopEACOP. (2022). “Seven Financiers Abandon TotalEnergies’ EACOP Pipeline in a Week,” May 21. www.stopeacop.net/our-news/seven-financiers-abandon-totalenergies-eacop-pipeline-in-a-week. #StopEACOP. (2022). “StopEACOP Campaign Calls on Standard Bank to Come Clean About Its Funding of the East African Crude Oil Pipeline,” August 22. www.banktrack.org/article/stopea cop_campaign_calls_on_standard_bank_to_come_clean_about_its_funding_of_the_east_african_ crude_oil_pipeline. Surapaneni, L. (2022). “Carbon Capture’s ‘Healthy’ Cigarette,” Harvard Public Health, August 17. https://harvardpublichealth.org/carbon-capture-climate-changes-healthy-cigarette/? Temple, J. (2021). “The UN Climate Report Pins Hopes on Carbon Removal Technologies That Barely Exist,” MIT Technology Review, August  9. www.technologyreview.com/2021/08/09/1031450/ the-un-climate-report-pins-hopes-on-carbon-removal-technologies-that-barely-exist/? Tett, G. (2010). Fool’s Gold: The Inside Story of J.P. Morgan and How Wall St. Greed Corrupted Its Bold Dream and Created a Financial Catastrophe, Updated with New Preface and Afterword. New York: Free Press.

441

35 RESPONSIBILITY AND GENDER Paula Casal

Men are afraid that women will laugh at them. Women are afraid that men will kill them. Margaret Atwood, based on a student survey. Second Words: Selected Critical Prose 1960–1982

Introduction Thomas Scanlon (1998: 248) distinguishes attributive responsibility, which concerns moral appraisal and attributions of praise and blame, from substantive responsibility, which concerns what we owe to each other. These come apart, for example, when we criticize others for being work-shy but help them bear the consequences of this conduct. Cost-bearing or liability is what matters to distributive justice and is the focus here in discussing behaviors that are “gendered” or characteristic of one sex.1 Social constructivists attribute gender patterns to socialization alone, while others attribute them to evolutionary pressures, because, as Frans de Waal (2022: 25) stresses in his recent book on gender, the same patterns are found in (i) most other mammals, (ii) young children, (iii) diverse cultures, and (iv) diverse millennia. Evolutionary explanations include not just anatomical variation but behavioral changes. For example, to be less visible, a species may become darker, nocturnal, or both. Some such behaviors are not transmitted genetically but learned. This makes socialization an important part of some evolutionary explanations. What makes them evolutionary is the mechanism: chance variation and natural (and sexual) selection cause certain genes to spread. Our genes affect our conduct our entire lives, and social facts, such as stress in pregnancy, may affect individuals before birth. To simplify matters, however, philosophers often call “natural” or “biological” whatever happens before birth, and “social” whatever happens afterwards. Having clarified our terms, we can now ask why a behavior’s social or natural origin should affect whether we can be held responsible for it. For example, some argue that while justice requires eliminating social inequalities, such as those emerging from discrimination and indoctrination, eliminating natural inequalities is less justified (Nagel 1997). Others deem eliminating natural inequalities equally justified but more difficult, because DOI: 10.4324/9781003282242-47 442

Responsibility and gender

they assume evolutionary forces have a firmer grip on us. This generalization is questionable. Genetic propensities can be extremely weak, while self-extrication from a religion, for example, can be tremendously difficult. This questionable but widespread assumption explains the concern that biological explanations may unduly justify behaviors or inequalities, or voice unscientific, conservative, and offensive views (Lowe 2022). This may be true of specific explanations, but it does not depend on the kind of explanation provided. Social explanations can also unduly justify behaviors or inequalities, or voice unscientific, conservative, and offensive views. Moreover, animals are capable of fast and profound cultural changes (e.g., towards nocturnal, urban, or nonaggressive lifestyles (Sapolsky 2008)), and natural science can be used to assist such changes rather than to deny their feasibility. Gendered behaviors have two types of regrettable consequences: self-harm and harm to others. An example of the first is men’s shortgevity (lesser longevity). An example of the latter is crime. The second question, then, is whether the fact that a behavior sets back our interests or those of others affect our attributions of responsibility. For example, we may think that if we persevere in doing something that harms us, perhaps we just cannot help it. But if we continue to do something that harms others, maybe we just do not care. Combining the question about the origins of gendered behaviors and the question about their consequences yields four possibilities: naturally and socially explicable self-harm, and naturally and socially explicable harm to others. There is, however, a different way to employ the distinction between origins and consequences. Gendered behaviors may originate in respect or neglect of duties (for example to one’s children) and holding people responsible for certain behaviors could have desirable or undesirable consequences. Both considerations may matter more than the kind of scientific explanation offered or the existence of self-regarding or other-regarding effects. The theories of Scanlon, John Rawls, and Ronald Dworkin may have different implications in this context, and retributivists and deterrence theorists may also draw opposite conclusions from similar statistics. Political philosophers who explicitly address this matter are divided. Shlomi Segall (2010: 108) denies men are responsible for their naturally shorter lifespans and argues that the state should invest more in men’s health and less in pregnancy and cervical cancer.2 Philippe Van Parijs (2015) denies male responsibility for either natural or social shortgevity but rejects compensations since shortgevity makes gender inequality merely smaller than assumed. Van Parijs (2015) and Gina Schouten (2015) argue that men’s greater propensity to crime and incarceration are both misfortunes and thus make gender inequality smaller than it would be otherwise. David Benatar (2012: 3, 32, 192) also deems shortgevity, crime, and incarceration major disadvantages of being male, but focuses on discrimination rather than on net outcome inequalities, and argues that sexism creates disadvantages for both sexes that gender egalitarianism should aim to correct. Some authors like John Kekes (1997) take any such discussions to be a reductio ad absurdum of egalitarianism. Such a conclusion is too hasty. Egalitarians may (i) legitimately include both natural and social inequalities among the morally arbitrary circumstances that justice must correct, and (ii) incorporate responsibility concerns in ways that exclude certain inequalities from the range of arbitrary circumstances that justice must correct. Discussing responsibility and gender involves discussing gender equality for various reasons. First, in other areas, such as international or intergenerational justice, many advocate the principle of sufficiency, which requires merely that we grant others enough (Casal 2007: 297). Equality, however, is the sovereign principle when discussing sexual and racial justice, as justice is not served by merely granting sufficient wages or sufficient rights to the 443

Paula Casal

relevant groups. Second, if sufficiency was enough here, discussing responsibility would be less important. For we may ensure individuals are above some minimum threshold, regardless of whether they brought themselves below it through their choices. We may do so for Samaritan reasons, or to maintain individuals at a level at which they can be held liable for their choices. But when the relevant principle is equality, we cannot keep on redistributing resources equally whenever somebody squanders their own (Casal 2007: 321). It is thus important to identify the inequalities we are substantially responsible for. Obviously, this chapter cannot offer the last word on this matter, but it means to stimulate future discussion. Let’s start with socially explicable self-harm.

Gendered self-harm: social explanations of shortgevity The most familiar sociological explanations of male shortgevity include the following: (i) Unhealthy consumption (alcohol, drugs, tobacco, excessive or unhealthy food) (Ritchie 2019; CDCP 2022). (ii) Speeding, drunk driving, and the sort of imprudence recorded by the Darwin Awards (Krisafis 2023).3 (iii) Tendencies to react to problems violently (fights, homicides, suicides . . .). (iv) Occupational hazards. (v) Neglecting childcare, which is associated with longevity. If these are the causes, are men responsible for their shortgevity? Regarding (i), one may note that men can already indulge much more than women in drinking and eating without gaining weight or being criticized or penalized for neglecting their appearance. It is not fair to expect women to exercise self-restraint, and then compensate men for not having exercised it. If some have a carpe diem lifestyle, the prudent should be allowed to keep the benefits of their self-restraint. Moreover, since it is generally women who have to look after men who have led unhealthy lifestyles, burden-sharing is already taking place. Something similar happens with (ii) and (iii), since male violence and risk-taking are already among the main causes of death for young women (CDCP 2017; Goodier 2022).4 Regarding (iv), one may argue that those who wanted those jobs despite their hazards should not be forced to compensate the chosen candidates for risks they would have accepted as part of the job (we return to this later). Caution, moreover, can reduce occupational risks. Finally, regarding (v), in Europe, men have become more involved in childcare, particularly in Scandinavian countries and the Netherlands, which have the smallest longevity gap, while the countries with the most gender inequality, like Turkey, Macedonia, and Romania have the largest gap (Eurostat 2022a). The world’s largest gap (11.6  years) is found in Russia, with its association of vodka with virility, repression of feminists, and large pay gap (Hutt 2016). If these are the social explanations, one way to eliminate lifespan inequalities would be to force men into cautious lifestyles. But men may value having the choice and the freedom to engage in different activities with some risk of self-harm, even at the price of being held liable. Discussing the value of such a choice, Scanlon (1998: 256) uses the example of an area marked as unsafe because of toxic waste. To set issues of corporate responsibility aside, let us imagine that it contains blooming nettles that cause a nasty rash. Scanlon (1998: 257) then asks us to imagine that, despite all the warnings, there’s a woman, Curious, who 444

Responsibility and gender

decides to enter the area. Scanlon denies that Curious can complain about being harmed, since she had adequate opportunities to avoid the harm, and given the warnings, alternative paths, etc., she could not claim society had not done enough for her (Scanlon 1998: 249–294; Williams 2006). Let us now add some gender stereotypes and assume that they are the result of socialization. For example, like the biblical Eve, Curious cannot resist her female botanical curiosity and must inspect the forbidden flowers. Or suppose that Daring, a man socialized into certain gender roles, decides to run naked through the nettles because of his authority-defying, sensation-seeking propensities, or to prove how tough he is. Does the assumption that the behavior is gendered change anything? One might think that if people are responsible for their foolishness, they are responsible for their gendered foolishness. In fact, if individuals make certain choices systematically, identify with them, and even associate them with their gender, the case for compensation seems even weaker than in the case of a one-off walk through some nettles. We may conclude that neither Curious nor Daring have a valid claim for compensation. On the other hand, one may argue that if many men and women make the same unfortunate decisions, perhaps society has not “done enough” to countenance sexist pressures. Suppose, for example, that we find that drugs, crime, and teenage pregnancy are particularly prevalent in a deprived area. Even those who believe we may (i) criticize its inhabitants (attributive responsibility) may deny either (ii) that we should withhold assistance (substantive responsibility) or (iii) that society has “done enough” until it improves the conditions of choice sufficiently to avoid such systematic outcomes (Scanlon 1998: 293; Duff 2010). We’ll return to this after reviewing natural explanations.

Gendered self-harm: natural explanations of shortgevity Among monogamous animals (over 80% of birds), all sex differences, longevity included, are small. Among species in which males contribute the greater parental effort (some birds, fishes, frogs, and insects) large, territorial, aggressive, ornamented females compete for male labor (Trivers 1972). Among polygamous species in which females contribute the greater parental effort (over 80% of mammals), the reverse obtains. Roughly speaking, the more males neglect parenting to chase different females, the younger they die. In very polygynous mammals like elephant seals, males are almost twice the size but almost half as long-lived as females (Anderson 2003). When a species has a history of polygyny (Darwin 1872) and males neglect offspring (Trivers 1972), males are: (i) larger and better armed or ornamented; (ii) more aggressive; (iii) more drawn to competitive interaction and rough play; (iv) likelier to engage in escalating violence; (v) more eager to mate; (vi) less discriminating about mates; (vii) more prone to high-risk behavior (Darwin 1872; Thornhill & Palmer 2000: 37); (viii) likelier to die in accidents, combat, or from disease (Daly & Wilson 1983); (ix) shorter-lived through physiological malfunction (Hamilton 1966); and (x) conceived/born in larger numbers, roughly balancing their premature deaths from violence, disease, malfunction, or imprudence (Alexander et al. 1979). 445

Paula Casal

This suggests a first reason for shortgevity: there is no point in nature investing in the self-repair of those that are likely to die of other causes anyway (Diamond 2002: 110). Another reason is that those that fail to parent also cease to be useful after conception. Male lions, for instance, wait for females to bring home the bacon, and having killed all the cubs they did not sire, leave only scraps for their own offspring to eat. Like feudal lords, they are only useful for deterring their rivals from behaving likewise. And youngsters sometimes do so much better without large, sexually aggressive, resource-guzzling males around, that the best some fathers can do for their offspring is to leave or die, as some do. Zooming in on the kind of mammal we are, like other great apes, elephants, and cetaceans, humans lie at one end of the r-K spectrum of reproductive strategies, i.e., they conceive very few offspring but invest enormously in each. This investment pays off only in long-lived species. This explains why these “slow developing” species are so large: Across species, larger animals, like bowhead whales, live longest, while small insects live the least. Within species, the smaller females, and often smaller individuals live longest (Samaras et al. 2003). “Extreme K” females endure long pregnancies, invest greatly in breastfeeding, protecting, and educating their offspring, and are easy to exploit, as they will not abandon their offspring even if hungry and abandoned themselves. They remain helpful when old, generating the so-called grandmother effect: old orcas, for example, are so useful that they survive menopause and males by several decades (Brault & Caswell 1993). Zooming in further on the kind of ape we are, a combination of a pelvis not initially designed for bipedal walking, or to allow the exit of big-brained babies, makes human pregnancy and childbirth unusually dangerous (Diamond 2002: 117). Mothers must be fit and far from death to survive it and educate their babies. And since mothers pass their genes to their sons, men may live longer as a result (Diamond 2002: 117). From a Rawlsian perspective, inequalities that are collectively beneficial, even to the worst off, are not unjust (Rawls 1999: 64–65). Thus, inequalities in longevity are not unjust, because they benefit male offspring and make them longer-lived than they would otherwise be. For Rawlsians, male shortgevity is not unjust anyway, since it does not unfairly reduce men’s access to income, wealth, the basic liberties, or the social basis of self-respect (Barclay 1999; Segall 2010: 99ff.; Clayton 2001). It is possible, however, that Rawls himself might have considered it unfair that some enjoy lesser capabilities that are not traceable to their behavior (Rawls 2001: 168–176), even if no loss of income or liberty is involved. Following this intuition, one may argue that men are responsible for their shortgevity if it is due to a chosen lifestyle; but neither attributive nor substantive responsibility for the behavior of previous men can fall on current men, as there is nothing they can do about it now. There are at least two answers to this argument. First, we can sometimes do something about inherited propensities, e.g., through modern medicine and prudent lifestyles. Childcare also appears to increase male longevity, perhaps because it lowers testosterone (Gettler et al. 2011), which may also explain why castrated men gain over 11 years of life (Gems 2014). The second answer does not appeal to empirical data, but to Dworkin’s doctrine. For Dworkin, distributive justice is served when individuals endowed with equal bidding power bid for resources in an auction until the markets clear and nobody envies the resource bundle of anyone else (or they would have bidden for it themselves). Then, equally situated individuals purchase insurance against what they consider a misfortune (Dworkin 2000, 2002; Clayton 2000; Williams 2002). This implies, for example, that people who consider their religion a blessing cannot also request funding because its costly pilgrimages 446

Responsibility and gender

leave them with fewer resources, or because its many prohibitions leave them with lower welfare, or because its rituals reduce their capabilities (Dworkin 2002: 119, 138; Williams 2002: 378). If endorsing a condition sufficed to be held responsible for it, we could be spared from settling whether those raised in a religion had an adequate opportunity to avoid it, whether religions are good or bad independently of how believers rate them, or even whether determinism is true. On this view, men who are happy to be short-lived men rather than longerlived women cannot expect any compensation for their shortgevity, since a trait cannot be considered advantageous when it is inextricably linked to other traits like risky pregnancies that are not advantageous. Similarly, since size and longevity are related, tall people who like being tall cannot complain about lacking short people’s life-expectancy, since there is nobody whose circumstances they envy.5 In contrast, women who are paid less for equal work can complain, because being underpaid is separable from being female. Individuals, however, may endorse a condition as a result of brainwashing or discriminatory history, so the self-identification test may not always detect injustice. In addition, Dworkin’s self-identification test must be clarified in relation to, for example, (i) pregnancyrelated disadvantages that, unlike salaries, are connected to being female, (ii) people with genetic disabilities who could not be who they are without the genes that disadvantage them, and (iii) indigenous peoples whose plight is integral to their identity. We do not want to add insult to injury by refusing support to anybody who does not declare their identity a curse. Assuming there is a plausible way to address these concerns, Dworkin’s self-identification test may be combined with Scanlon’s, as follows. Agents are not responsible for conditions that result from bad brute luck. A  condition qualifies as brute luck if agents lack adequate opportunities to avoid it, and as bad luck if agents deem it a misfortune they would have insured against. No opportunity and no endorsement may thus be two necessary and jointly sufficient conditions for being entitled to compensation. Alternatively, we may think both are important considerations, although neither is necessary or sufficient for a verdict of substantive responsibility (Williams 2002).

Gendered harm to others: crime, deterrence, and retribution Let us begin with a recap. A social explanation can be invoked to argue either that (i) men are liable, as they had the opportunity to avoid the relevant gendered behavior, or that (ii) they are not liable because they are the victims of sexist pressure. And a natural explanation can be invoked to argue either that (i) natural inequalities are morally arbitrary and unjust, or that (ii) they cannot be unjust because they are merely facts of life, which do not result from the actions of moral agents. In combination with Rawls’s theory, biological explanations that show that lifespan inequalities are collectively beneficial can be employed to argue that they are therefore not unjust. And in combination with Dworkin’s theory, biological explanations that associate longevity with other costs, like dangerous births, can be used to argue that since shortgevity is part of being male, those glad to be male have no valid claims for compensation. This summary suggests that the choice of explanation is relevant because different scientific explanations permit different philosophical arguments for and against responsibility, but the choice of explanation does not determine if individuals are responsible or not. In this sample, and contrary to expectations, natural explanations allow more arguments 447

Paula Casal

for declaring individuals responsible than social explanations, but either conclusion is possible with either kind of explanation. So does the kind of scientific explanation really matter? When considering the relevance of an explanation, the main difference is not whether it is social or natural, but whether the relevant effects emerge through gendered behaviors or independently of what anybody may do. For example, social explanations of male shortgevity could refer to medical improvements that happen to benefit women more. And one biological explanation concerns the lack of a second copy of the X chromosome that protects against harmful mutations (Carazo et al. 2016). If these are the correct social or natural explanations, then men will die younger, whatever they do. This is the scenario where men are less responsible. By contrast, if men take risks, because of either socialization or evolution, then the case for compensation dwindles. So perhaps the fact that something has natural or social origins makes no difference after all – which is just as well, because both are probably relevant and inextricably intertwined. In the case of crime, however, there are no social or natural factors that have effects independently of our actions. Thus, whether we regard criminals as the social products of a system that failed them, or we think of them as acting under the effects of a potent substance that they did not choose to inject (testosterone), somebody must still pull the trigger. And it is this trigger-pulling, and not its scientific explanation, that makes all the difference. Let us glance at some facts and possible explanations. In Europe and much of the world, 95% of the prison population are men (Benatar 2012: 59; Aebi  & Tiago 2020). Many women are in prison because they are too poor to pay bail or fines (United Nations 2014: 4), or for prostitution, drug, or migration-related offenses. Some studies show that around 86% of female offenders suffered sexual or physical abuse in childhood, 77% had abusive partners, and most ended up with post-traumatic stress disorder, consuming alcohol and drugs, and/or presenting chronic mental and physical health conditions (Dholakia 2021; Trauffer & Widow 2017). Despite such life histories, the more serious the crime, the less likely it is that it was committed by a woman. Women commit smaller crimes and do so less often, with men having higher recidivism rates (Heidensohn 1989: 86). Even shoplifting, which is often seen as the typical female offense, is more often practised by men (ibid.). Girlfriends of criminals may end up being implicated in crimes, but if a crime is violent, or sexual, and particularly if it is both violent and sexual, it is almost certainly a male crime. 92% of the victims of sexual assault are women (Urban Institute 2001), but male victims are also typically attacked by men. Crime was even once thought to derive from the Y chromosome (Witkin et al. 1976). Chivalry theory attributes these statistics partly to male officer’s greater leniency towards female offenders. Later studies found that differences attributed to chivalry could be better explained by the combination of female’s committing much smaller offenses, and showing remorsefulness, rather than belligerence when caught (Scutt 1979). Societies that have become more gender-egalitarian – like Greenland, Norway, Finland, Lichtenstein, Iceland, and Switzerland – have fewer prisoners but a higher percentage of female prisoners; while in countries like Turkey, Bosnia and Albania, the female percentage is very small (Chu et al. 2021; Eurostat 2022b). One may think that this is because less intimidated women commit more crimes. But in fact, it seems that incarceration only increased for some time, then stabilized, and not necessarily because of an increase in female

448

Responsibility and gender

crime, but because as sexism diminished, male crime dropped (Kruttschnitt 2013: 295). In some places like the United States, female incarceration increased because incarceration for very minor offenses increased too. Similar statistics across cultures, and greater male risk-taking and aggressiveness in mammals, suggest a biological explanation. Moreover, men are likeliest to be arrested for violent crime between the ages of 18 and 24 (Urban Institute 2001; Casal 2013). Testosterone peaks between 17 and 21 and the brain areas associated with judgment, sociality, and self-restraint are not fully formed in men until age 25 (Arain et al. 2013). Other data suggests social factors. For example, maternal education, maternal employment, and number of siblings are important predictors of young male crime (Ou & Reynolds 2010). Whatever the explanation, nobody proposes releasing men until the number of male and female prisoners evens out. But other disagreements may emerge. For example, both Van Parijs and John Roemer interpret responsibility in connection to the existence of factors beyond our control, which they detect through the typical behavior of a group member. For example, if all senior steel workers smoke, we should take this into account when discussing funding their lung-cancer treatment (Roemer 1995). If we apply this view in distributive justice to retributive justice, being male becomes an attenuating factor, comparable to being a senior steel worker (Van Parijs 2015: 87). This would suggest penalizing women more for the same crimes because they are so unlikely to commit them. According to “doubledeviance theory” (Heidensohn 1989: 102), this is exactly what already takes place: society is more unsettled by unpredictable crimes and can be harsher on female convicts who break both the law and social expectations (Kennedy 2018). The same statistics, however, could also lead us to conclude that truly exceptional circumstances must have concurred for a woman to attack someone. It seems unfair to punish women more, precisely because they normally put up with much more without reacting. Moreover, if we think that the point of incarceration is deterrence, it seems that men need deterring much more. So, for example, in view of the high number of men who harass, injure, rape, and kill women with whom they are or were in a relationship, in a country with an otherwise low murder rate like Spain, special penalties have been introduced to deter men from crimes classified as “gender violence.” Again, it’s the choice of philosophical outlook rather than the choice of scientific explanation that makes a difference.

Final considerations: duties and incentives Imagine that an island that has achieved equality of resources suffers a great fire. Altruists rush to stop it while others focus on protecting their properties, even seizing the opportunity to profit privately. It would be implausible to say that, since the altruists could have avoided firefighting, and they identify with acting dutifully, they are now liable for their losses. It is not in our interest to live in a society that penalizes those who respect duties and rewards those who ignore them. So, assuming firefighting is a real (rather than an imaginary) duty, individuals should not be liable for the losses they have incurred discharging it, both for incentives and fairness reasons (Casal 1999: 369–370). Thus, when discussing individuals’ responsibility for what they did, we must take into account (i) if they were neglecting or discharging their duty and (ii) if holding people responsible has beneficial consequences.

449

Paula Casal

Those injured by discharging their duties as firefighters, or in the military or police, have attributive but no substantive responsibility, and so should be praised and compensated for injuries incurred on duty, both on incentive and fairness grounds. Let us now reconsider longevity and crime. It is clear that there is no duty to drink vodka, that public policy should disincentivize its overconsumption, and that those who abuse it may be criticized. But this still leaves us with two options. One is to insist that individuals should bear the full costs of their voluntary consumption and hope that this will suffice to disincentivize it. Another is to argue that society had not “done enough” if consumption remains rampant, and so additional measures (taxation, rationing, education) are needed. And the same applies to crime: neither blaming nor punishing people seems enough to reduce crime, and this may suggest that society had not done enough here either. Both being a criminal and being punished for it are serious misfortunes (Van Parijs 2015). And men are not just likelier to kill but represent 82% of homicide victims (Urban Institute 2001; Benatar 2012: 32; Van Parijs 2015). In addition, some aspects of incarceration make it worse for men too (Benatar 2012: 39, 54, 59, 147). This supports Van Parijs’s and Schouten’s claim that men’s higher likelihood of incarceration “is a distinct injustice to men” (Schouten 2015: 138). Their argument rests on a comparison between men who commit crimes and women who sacrifice their careers to engage in care work. Both make choices, Schouten argues, but “in both cases, the inequalities in the background against which the relevant choice is made are not chosen; and in both cases, the relevant choice is harmful to the chooser” (142). There are at least two objections to this view. First, when social scientists tell us that power corrupts and absolute power corrupts absolutely, that the rich voted for Trump, or that convicts are likelier than non-convicts to commit future crimes, we needn’t interpret the relevant statistics as indicating injustice or reduced responsibility, even if being powerful, rich, or a convict was unchosen. Second, women arguably sacrifice their career in response to injustice at home (husbands shirking housework) and at work (being underpaid, under-recognized, patronized, harassed . . .). If traditionally society has been organized as a comprehensive affirmative action plan for men (MacKinnon 1987: 36), a comparable argument is not available to regular men. In defense of Van Parijs and Schouten, one may argue that men are socialized to seek danger to show virility and courage, because having such individuals is beneficial for defense purposes. The Viking berserker cult is an extreme example of this. Berserkers injured themselves and others frequently during peacetime, quite predictably. And precisely because society oriented itself in one direction for so long, a comparable, contrary social investment is arguably necessary to rectify the effects of this history, and for society to have, on balance, “done enough” to divert men from self-harm and crime. Note that the claim that, given their natural or social history, men cannot now be left to their own devices is consistent with claiming that they had adequate opportunities to avoid self-harm and crime, and that they should be both criticized and punished when appropriate (Duff 2010: 137). Having focused mainly on men, let us end by revisiting, with a focus on duties and incentives, the injustice to women that Schouten invokes as paradigmatic. Many women sacrifice paid work to become housewives and mothers, and then end up poor. Let us consider duties first. Parents have a duty to satisfy their children’s needs, rather than a duty to provide them with anything that they may want or that enables them to outcompete others. So only part of the behavior of many mothers who quit paid work constitutes “acting on duty.” Second, we must ask whether holding these mothers responsible for their choices has desirable 450

Responsibility and gender

consequences. These women are generally seen as merely engaging in self-harm. But as in other cases of self-harm, from vodka to suicide, others may be negatively affected. Even in those cases where neither the woman nor her children end up living in poverty, she may harm other women who will be competing with men who have wives who do everything at home. In addition, a woman who has children she cannot support is taking the high-risk, high-stakes gamble of making her future and that of her children depend on her remaining attractive to her husband over time. Is this a behavior we want to encourage? And is it fair to make the women who refrain from having more children than they can support themselves subsidize the losses of those who decided to gamble and specialize in homemaking and reproduction? If the answer is no, then rather than focus on whether there is a biological or sociological explanation for such behaviors, we should focus on the kind of behavior we want to incentivize. For example, measures that facilitate the lives of working mothers may be preferable to those that focus on picking up the tab when the high-risk, high-stake gamble of choosing homemaking as a profession does not work out. And society has certainly not yet done enough for working mothers and gender equality.

Conclusion Philosophers disagree on whether people are responsible for their gendered behaviors. They may do so because they disagree on their scientific explanation, as well as on what normative consequences follow from the choice of explanation. They may also disagree about what follows from the fact some behaviors involve self-harm or harm to others. However, one may invoke different scientific explanations either to deny or to affirm responsibility. And all behavior, and not just gendered behavior, probably has deeply intertwined social and biological roots and involves both harm and self-harm. Perhaps, rather than continuing to focus on these two distinctions (social versus biological explanations, and harm versus self-harm), we should pay more attention to whether, when performing the relevant actions, individuals neglect or discharge their duties, and whether or not holding individuals responsible has desirable consequences. And we should not focus only on whether individuals are blameworthy or liable for their gendered behaviors, but also on whether society has done enough to school them out of them.

Notes I thank Tom Parr, Areti Theofilopoulou, and Andrew Williams for excellent detailed comments, and Philippe Van Parijs and David Benatar for discussion. All links were working in February 2023. 1 I cannot discuss here whether we have good definitions of “male” and “female.” In biology, females are those that contribute the larger gamete. 2 It is unfair to deem pregnancy and cervical cancer the exclusive responsibility of the infected or pregnant patient, when both conditions are caused by semen, and often because of men’s insistence, since statistically men are keener on sex, likelier to take risks and more assertive. Pregnancy, moreover, is also vital for male fetuses and required for the continuation of humanity, and so not just a women’s health issue. But Segall could just modify his examples. 3 Darwin Awards winners risk death unnecessarily: https://darwinawards.com/darwin/darwin 2022-01.html. 4 Women are also more likely to be killed by environmental problems (UN 2022) and men more likely to cause them (Carlsson Kanyama et al. 2021), even if damage caused by wars is excluded from the calculations. 5 Sometimes there will be a person they can envy, as some tall people are longevous, even if most are not.

451

Paula Casal

Further reading Whether we are more (or less) responsible for our gendered behavior than for our nongendered behavior and whether the scientific origin of gendered behaviors makes a normative difference are underexplored issues. Van Parijs 2015 touches on both when he explains four dimensions in which he deems men worse off than women (longevity, education, incarceration, and sexual frustration). He receives responses from Ana de Miguel, Gina Schouten, Jesús Mora and Paula Casal, and responds to all of them in the same issue of Law, Ethics and Philosophy 2015. The whole symposium is succinct, engaging, informative, and philosophically interesting. Scanlon 1998 and Dworkin 2000 and 2002 do not discuss gender, but one can always read them profitably and bear gender in mind. The empirical data of Susan Moller Okin’s book (1989) Justice, Gender and the Family New York: Basic Books is now dated. For example, women no longer engage in anticipatory surrender and quit education, but instead attempt surpassing men’s qualifications to protect themselves from subordination at work or at home. But Okin’s influential ideas remain worth studying.

References Aebi, M. F.,  & Tiago, M. M. (2020). Prisons and Prisoners in Europe 2020: Key Findings of the SPACE I  Report. Council of Europe and University of Lausanne. https://wp.unil.ch/space/ files/2021/06/210329_Key_Findings_SPACE_I_2020.pdf. Alexander, R. et al. (1979). “Sexual Dimorphism and Breeding Systems of Pinnipeds, Ungulates, Primates and Humans,” in N. A. Chagnon & W. Irons (ed.), Evolutionary Biology and Human Social Behavior (pp. 402–443). North Scituate: Duxbury. Anderson, G. (2003). “Elephant Seal Body Form,” Marine Science, 5.3.2 www.marinebio.net/ marinescience/05nekton/esbody.htm. Arain, M. et al. (2013). “Maturation of the Adolescent Brain,” Neuropsychiatric Disease and Treatment, 9, 449–461. Barclay, L. (1999). “The Answer to Kekes’ Question,” Ethics, 110(1), 84–92. Benatar, D. (2012). The Second Sexism. Oxford: Blackwell. Brault, S.,  & Caswell, H. (1993). “Pod-Specific Demography of Killer Whales,” Ecology, 74(5), 1444–1454. Carazo, P. et al. (2016). “Inbreeding Removes Sex Differences in Lifespan in a Population of Drosophila Melanogaster,” Biology Letters, 12, 20160337. Carlsson Kanyama, A. et al. (2021). “Shifting Expenditure on Food, Holidays and Furnishing Could Lower Greenhouse Emissions by 40%,” Journal of Industrial Ecology, 25(6), 106–116. Casal, P. (1999). “Environmentalism, Procreation, and the Principle of Fairness,” Public Affairs Quarterly, 13(4), 363–376. ———. (2007). “Why Sufficiency Is Not Enough,” Ethics, 117(2), 296–326. ———. (2013). “Sexual Dimorphism and Human Enhancement,” Journal of Medical Ethics, 39(12), 722–728. Centers for Disease Control and Prevention, CDPD. (2017). “Leading Causes of Death.” www.cdc. gov/women/lcod/2017/all-races-origins/index.htm. ———. (2022). “Excessive Alcohol Use Is a Risk to Men’s Health.” www.cdc.gov/alcohol/fact-sheets/ menshealth.htm#:~:text=Adult%20Men%20Drink%20More%20than,with%2049%25%20 of%20adult%20women.&text=Men%20are%20more%20likely%20to%20binge%20 drink%20than%20women. Chu, D. C. et al. (2021). “Gender Equality and Female Offending,” International Journal of Offender Therapy and Comparative Criminology, 65(13–14), 1496–519. Clayton, M. (2000). “The Resources of Liberal Equality,” Imprints, 5(1), 63–84. ———. (2001). “Rawls and Natural Aristocracy,” Croatian Journal of Philosophy, 1(3), 239–259. Daly, M., & Wilson, M. (1983). Sex, Evolution and Behavior. North Scituate: Duxbury. Darwin, C. (1872). The Origin of the Species. London: Penguin. de Waal, F. (2022). Different. New York: Norton. Dholakia, N. (2021). “Women’s Incarceration Rates Are Skyrocketing,” Vera, May 17. www.vera. org/news/womens-voices/womens-incarceration-rates-are-skyrocketing. Diamond, J. (2002). The Rise and Fall of the Third Chimpanzee. London: Vintage.

452

Responsibility and gender Duff, R. A. (2010). “Blame, Moral Standing and the Legitimacy of the Criminal Trial,” Ratio, 23(2), 123–140. Dworkin, R. (2000). Sovereign Virtue. Cambridge: Harvard University Press. ———. (2002). “Sovereign Virtue Revisited,” Ethics, 113(1), 106–143. Eurostat. (2022a). “Mortality and Life Expectancy Statistics.” https://ec.europa.eu/eurostat/statisticsexplained/index.php?title=Mortality_and_life_expectancy_statistics. ———. (2022b). “Prison Statistics.” https://ec.europa.eu/eurostat/statistics-explained/index.php?title= Prison_statistics. Gems, D. (2014). “Evolution of Sexually Dimorphic Longevity in Humans,” Aging, 6(2), 84–91. Gettler, L. T. et al. (2011). “Longitudinal Evidence that Fatherhood Decreases Testosterone in Human Males,” PNAS, 108(39), 16194–16199. Goodier, Michael. (2022). “Male Drivers Three Times more Likely to be in Road Collision with Pedestrians,” The Guardian, October 9. Hamilton, W. (1966). “The Moulding of Senescence by Natural Selection,” Journal of Theoretical Biology, 12(1), 12–45. Heidensohn, F. (1989). Crime and Society. London: Macmillan. Hutt, R. (2016). “Which Countries Have the Largest Gender Gap in Life Expectancy?” World Economic Forum, May  20. www.weforum.org/agenda/2016/05/countries-where-women-outlive-men-bydecade/. Kekes, J. (1997). Against Liberalism. Ithaca: Cornell University Press. Kennedy, H. (2018). “The Myth of the She-Devil: Why We Judge Female Criminals More Harshly,” The Guardian, October 2. www.theguardian.com/uk-news/2018/oct/02/the-myth-of-the-she-devilwhy-we-judge-female-criminals-more-harshly. Krisafis, A. (2023). “Toxic Masculinity Fuels Dangerous Driving like Drink French Advert says,” The Guardian, February 8. www.theguardian.com/world/2023/feb/08/toxic-masculinity-fuels-dangerous-driving-like-drink-french-advert-says?CMP=Share_AndroidApp_Other. Kruttschnitt, C. (2013). “Gender and Crime,” Annual Review of Sociology, 39, 291–308. Lowe, D. (2022). “Biological Explanations of Social Inequalities,” Pacific Philosophical Quarterly, 103(4), 694–719. MacKinnon, C. (1987). Feminism Unmodified. Cambridge: Harvard University Press. Nagel, T. (1997). “Justice and Nature,” Oxford Journal of Legal Studies, 17(2), 303–321. Ou, S. R., & Reynolds, A. J. (2010). “Childhood Predictors of Young Adult Male Crime,” Children and Youth Services Review, 32(8), 1097–1107. Rawls, J. (1999). A Theory of Justice. Cambridge: Harvard University Press. ———. (2001). Justice as Fairness. Cambridge: Harvard University Press. Ritchie, H. (2019). “Who Smokes More, Men or Women?” Our World in Data. https://ourworldin data.org/who-smokes-more-men-or-women. Roemer, J. (1995). “Equality and Responsibility,” Boston Review, 20(2), 3–7. Samaras, T. T. et al. (2003). “Is Height Related to Longevity?” Life Sciences, 72(16), 1781–1802. Sapolsky, R. (2008). Stress: Portrait of a Killer. National Geographic Documentary. www.youtube. com/watch?v=AYFZAYenR20&ab_channel=GeorgeKalarritis%2CClinicalPsychologist. Scanlon, T. M. (1998). What We Owe to Each Other. Cambridge: Harvard University Press. Schouten, G. (2015). “Are Unequal Incarceration Rates Unjust to Men?” Law, Ethics and Philosophy, 3, 136–150. Scutt, J. A. (1979). “Myth of the Chivalry Factor in Female Crime,” Australian Journal of Social Issues, 14(1), 3–20. Segall, S. (2010). Health, Luck, and Justice. Princeton: Princeton University Press. Thornhill, R., & Palmer, C. (2000). A Natural History of Rape. Cambridge: MIT Press. Trauffer, N., & Widow, C. S. (2017). “Child Abuse and Neglect, and Psychiatric Disorders in Nonviolent and Violent Female Offenders,” Violence and Gender, 4(4), 137–143. Trivers, R. (1972). “Parental Investment and Social Selection,” in B. Campbell (ed.), Sexual Selection and the Descent of Man (pp. 136–179). Chicago: Aldine. United Nations. (2014). Handbook of Women and Imprisonment (2nd ed.). New York: UN. www. unodc.org/documents/justice-and-prison-reform/women_and_imprisonment_-_2nd_edition.pdf. ———. (2022). “How Gender Inequality and Climate Change Are Related.” www.unwomen.org/ en/news-stories/explainer/2022/02/explainer-how-gender-inequality-and-climate-change-are-

453

Paula Casal interconnected#:~:text=Women%20and%20girls%20experience%20the,less%20access%20 to%2C%20natural%20resources. Urban Institute. (2001). “Risk of Violence Highest for 18–24-Year-Olds,” March 8. https://webarchive. urban.org/publications/900403.html. Van Parijs, P. (2015). “Four Puzzles on Gender Inequality,” and “Real Freedom for All Women (and Men): A Reply,” Law, Ethics and Philosophy, 3, 79–89 and 161–177. Williams, A. (2002). “Equality for the Ambitious,” Philosophical Quarterly, 52(208), 377–389. ———. (2006). “Liberty, Liability, and Contractualism,” in N. Holtug  & K. Lippert-Rasmussen (eds.), Egalitarianism (pp. 241–261). Oxford: Clarendon Press. Witkin, H. A. et al. (1976). “Criminality in XYY and XXY Men,” Science, 193(4253), 547–555.

454

Section 9: Responsibility in the law

36 LEGAL AND MORAL RESPONSIBILITY Peter Cane*

According to a widespread opinion in contemporary moral philosophy, law is (at best) a distorted image of ‘morality’, the latter being understood as pure theory or objective fact, uncontaminated by practice. As such, morality stands to law as critical standard to social custom. The basic argument of this chapter is that for philosophers of a naturalistic turn of mind, an understanding of responsibility in law is (or should be) integral to understanding responsibility in morality, and responsibility period. The first section provides a broadbrush account of ‘responsibility in law’. The second section discusses modern, mainstream philosophical approaches to responsibility. The third section offers a reassessment of the relationship between concepts of responsibility in law and moral philosophy.

Responsibility in law The origins of the use, in British English, of the word ‘responsibility’ have been traced to the late eighteenth century (McKeon 1957); but the word rarely appears in the lexicon of English law.1 However (perhaps under the influence of late-twentieth century moral philosophy),2 ‘responsibility’ is now widely used in theoretical legal literature3 to refer to various aspects of the conceptual design of what lawyers call ‘areas of law’ – such as criminal law and contract law. It is useful to distinguish between three different ‘models of legal responsibility’: a criminal law model, a civil law model and a public law model.4 The criminal law model of responsibility focuses on agents, conduct and sanctions. Victims of crime make only cameo appearances in the legal drama. Indeed, certain crimes are ‘victimless’ (offenses of possessing drugs, for instance), and unsuccessful attempts to commit a ‘victim crime’ may themselves be punishable offenses. The impact of victim crimes on victims is part of the definition of the proscribed conduct. Most criminal prosecutions are initiated and conducted by state authorities, not by victims. Occasionally, a convicted person may be ordered to pay compensation to the victim for harm suffered as a result of the crime, and this may indirectly affect the penalty imposed on the criminal; but compensation is a minor adjunct to, not an integral part of, the system of criminal law responsibility.

457

DOI: 10.4324/9781003282242-49

Peter Cane

By contrast, the civil law model of responsibility is concerned as much with those adversely affected by conduct as with as with its agents. Claims in civil law are made by (or on behalf of) the person(s) ‘harmed’ by specified conduct. There is no such thing as a ‘victimless’ civil wrong, and attempting a civil wrong is not, itself, a civil wrong. Although we sometimes say that a person is ‘guilty of’ a civil wrong such as a tort or a breach of contract, the term ‘liable for’ is more common. The typical legal response to a civil law wrong is an ‘award’ of monetary recompense to be made by the wrongdoer to the adversely affected person. Most commonly, such recompense takes the form of ‘compensation’ for ‘harm’ suffered. Occasionally, a court may order a person (by the ‘award of an injunction’) not to commit an anticipated civil wrong or to desist from a course of civil wrongdoing. Very occasionally and quite exceptionally, a civil claimant may be awarded damages that provide more than recompense, in the name of punishing the wrongdoer. But as, conversely, in the case of compensation orders made against criminals, penal awards in favor of civil claimants are a minor adjunct to, not an integral part of, the system of civil law responsibility. The public law model of responsibility rests on a distinction between governors and the governed in the context of ‘the (nation) state’. Governors  – let us call them ‘public officials’ – may be subject to the criminal law responsibility regime; and there are some crimes that, by definition, can only be committed by a public official. Public officials may also be subject to civil law responsibility; and there is at least one civil law wrong (misfeasance in public office) that, by definition, can only be committed by a public official. However, structurally, public law responsibility is a sort of hybrid. Public law claims are typically initiated by individuals; but actionable public law wrongs may or may not adversely affect identifiable individuals. In certain cases where the wrong does not adversely affect any identified individual, an (unaffected) individual may make a public law claim as a sort of ‘private prosecutor’ would in the criminal law context. Public law wrongs are defined in terms of conduct of public officials: public officials must not exceed their official powers (to make decisions and rules) or neglect their official duties; they must observe fair procedures in carrying out their functions; and they must not act ‘unreasonably’. Where a public official exceeds their powers or fails to observe fair procedure or acts unreasonably, their decision may be ‘quashed’ – deprived of legal effect. A public official who neglects their duty may be ordered to perform it; and a public official may be ‘prohibited’ from exceeding their powers. However, such orders are understood to be prophylactic, not penal. The definition of a legal wrong – whether in criminal law, civil law or public law – typically has two components, one concerned with action (or inaction) (known in criminal law as ‘actus reus’ and here called ‘the action element’) and another concerned with mental state (known in criminal law as ‘mens rea’ and here called ‘the fault element’). Basic fault elements found in the law include ‘deliberation’ (as in ‘she did that deliberately’), intention (understood as relating to consequences of conduct that are ‘aimed at’), recklessness (understood, in its core sense, as deliberate action that carries known risks of adverse consequences) and ‘malice’ (understood in terms of an agent’s motives or reasons for action). Additionally, however, the fault element of many legal wrongs is ‘negligence’ defined not as ‘inadvertence’ but as failure to comply with a specified standard of conduct (the so-called ‘standard of reasonable care’). Although inadvertent conduct may be negligent in this sense, so may deliberate, intentional or reckless conduct. In this sense, legal negligence is not a mental state. Finally, many legal wrongs have no fault element but attract ‘strict’ liability. Strict liability is liability regardless of the presence or absence of fault, not liability in the 458

Legal and moral responsibility

absence of fault. Deliberate, intentional, reckless or negligent conduct may all attract strict liability, as may faultless conduct. Although concepts of responsibility play a central role in the law, not all legal liability is responsibility-based. The classic example is the liability, of the recipient of a payment made to them as the result of a mistake, to return the mistaken payment. However, such liability is exceptional, rare and confined to civil law; and even in that context, the label is contested by some. A distinction of great significance for understanding legal responsibility is that between what we might, respectively, call ‘prospective’ responsibility and ‘retrospective’ (or ‘historic’) responsibility. In a well-known discussion (Hart 1968: 211–230), HLA Hart referred to prospective responsibility as ‘role responsibility’ (see also Cane 2016). However, to the extent that the language of ‘roles’ suggests responsibility attaching to institutional affiliation, it is too narrow. Individuals, as members of society, have responsibilities to others merely by virtue of their sociality and independently of affiliation to any group, organization or institution within society. The more basic distinction is between having responsibilities, present and future, and being responsible for the past. Law is as much concerned with prospective responsibility as with its retrospective counterpart. Law commonly uses terms such as ‘duty’ and ‘obligation’ to refer to prospective responsibilities. An obvious connection between the two modes of responsibility is that retrospective responsibility assumes prospective responsibility. Less obviously, the appropriateness of particular conditions of retrospective responsibility may depend on the nature and content of the prospective responsibilities to which retrospective responsibility relates. For instance, criminal penalties are sensitive to the seriousness of the crime; and in civil law, the choice between intention, recklessness and negligence (for instance) as conditions of responsibility may depend on the nature and content of the duty or obligation that the wrongdoer has breached. Several other features of legal responsibility practices deserve notice. One is the distinction between an agent’s personal responsibility for their own conduct and its consequences, and a person’s vicarious liability for the agency of another. Like much else about legal responsibility, vicarious liability is concealed by a focus on the criminal law model, in which vicarious liability (as opposed to accessory liability)5 plays no part. There may also be a case for arguing that vicarious liability is, in some sense, not responsibility-based.6 Another major feature of legal responsibility practices is the imputation to metaphysical entities, such as corporations and states, of responsibility independent of the responsibility of any of the human agents involved in the entity’s activities (see further Cane 2002: ch 5). Finally, an integral aspect of legal practices of responsibility is the requirement to prove satisfaction of the various conditions of responsibility. Although evidence and proof are often treated as ‘adjectival’ to responsibility practices, of purely practical rather than theoretical import, they are central to the significance and acceptability of those practices. Suffice it to note the difference between the standard of proof under the criminal law model (beyond reasonable doubt) and that under the civil law model (on the balance of probabilities); or the complexities of ‘proving’ the existence of mental states such as intention.7

The (moral) philosophy of responsibility Modern, mainstream philosophy of responsibility is concerned primarily with retrospective (moral) responsibility, and that will be the focus of what follows. The foregoing account of (retrospective) responsibility in law is based on the practices of authoritative law-making, 459

Peter Cane

social institutions – primarily legislatures and courts. Law is a complex set of social and institutional practices. To say that law is a social practice is to say that it is a function of social facts. So much is typically common ground between legal ‘positivists’8 and modern ‘natural lawyers’;9 they disagree not over whether law and morality are distinctively different from one another but about the relationship between them. Positivism is typically associated with a ‘separation thesis’ to the effect that what the law is and what it ought to be are independent of one another. Non-positivists, by contrast, deny the independence of law from morality, although there are different views about the precise nature of the mutual dependence of law and morality and, indeed, about the nature of morality. Many moral philosophers seem to subscribe to a separation thesis: law is one thing, morality is categorially another; and in any conflict between law and morality, morality will prevail. In this view, morality stands to law as critical standard to social practice. It does not follow, of course, that morality may not itself be a function of facts. Indeed, modern empirical philosophers apparently assume that morality, like law, is a form of social practice amongst the ‘folk’, whose moral concepts they seek to investigate (Connolly 2021). However, historically, many philosophers have begun with an assumption that retrospective responsibility depends not on social facts but on facts about the way the ‘natural’ world works and, particularly, on whether or not the universe is causally deterministic. Being ‘morally’ responsible, they say, depends on being ‘in control’ of all the factors on which an attribution of responsibility is based; and if the universe is causally deterministic (or if, in theological terms, there is no such thing as free will) such control may be lacking. In that case, our responsibility practices (whether legal or extra-legal) would, root and branch, be inconsistent with ‘natural’ physical facts and, for that reason, unjustifiable. Such ‘responsibility-sceptics’ (Caruso 2018), however, do not purport to know whether the universe is causally deterministic.10 Others argue that although responsibility requires control, the sort of control it requires – cashed out in terms of human capacities and opportunities – is consistent with causal determinacy (Talbert 2019: §1). Such ‘compatibilism’ is essentially consistent with legal practices of responsibility. The terms of debates about retrospective responsibility were significantly changed by Peter Strawson in 1962 (Strawson 1962). He argued that responsibility is best understood in terms of ‘reactive emotions’ such as guilt and blame, not in terms of determinism and control. Strawson can be read as developing the argument in two different ways. On the one hand, he accepts responsibility-skepticism but claims that it is simply ‘irrelevant’ to his emotions-based account. On the other hand, he presents himself as a believer in responsibility but rejects utilitarian justifications for responsibility practices and argues, instead, that the reactive emotions are an inherent part of our human nature and, hence, that our responsibility practices are an inevitable part of human social life (Watson 2014). He draws an analogy between responsibility judgments and inductive reasoning (Strawson 1962: 23, fn.1). Even though, on reflection, we may accept the flaws of inductive reasoning, nevertheless we go on engaging in inductive reasoning because, by nature, we cannot do otherwise (see also Shoemaker 2017). Similarly, we may acknowledge the impact of determinism on responsibility, yet still go on making responsibility judgments simply because we cannot do otherwise. As a result, our reflective and non-reflective judgments (about induction and responsibility, for instance) may conflict – implying that such conflict is, itself, a feature of the human condition. Many found Strawson’s approach unsatisfying because he purported not only to explain our responsibility practices but also to justify them. How could unreflective practices inconsistent with our considered judgments be justified? 460

Legal and moral responsibility

Inspired by Strawson, others have attempted to confront what Victoria McGeer and Philip Pettit call the ‘hard problem of responsibility’ generated by human, social practices of blaming people for their conduct even in the face of the possibility that their conduct may be explicable by ‘brute facts’ of physics over which they have no control (McGeer & Pettit 2015: 162). Jay Wallace (Wallace 2014) argues that the control required for moral responsibility is not causal and does not reside in the capacity to act in a reasonably expected way, but rather in the capacity ‘to recognise and act on moral reasons’ (Cullity 1997: 805) for acting in that way – ‘reasons-responsiveness’ as it is commonly called. As this capacity is cashed out, most people possess it to a sufficient degree most of the time that they are appropriate subjects for attributions of retrospective responsibility. McGeer and Pettit (McGeer & Pettit 2015) add a temporal element to such an approach by arguing that the capacity to be responsive to reasons can change or grow or diminish over time, and that engagement in responsibility practices may strengthen and improve reasons–responsiveness. Manuel Vargas argues that our responsibility practices (as opposed to particular instances of those practices) are justified because they ‘foster a distinctive form of agency in us . . . sensitive to and governed by moral considerations  .  .  . intrinsically valuable  .  .  . or  .  .  . important for other things we care about’ (Vargas 2013; see also Sie 2018). T M Scanlon’s Strawson-inspired argument is that we value choice and control ‘instrumentally’ to the extent that it allows us to predict and shape the future; ‘representatively’, in that choice expresses our individuality; and “symbolically,” in that choice allows us to operate as full ‘adult’ members of society (Scanlon 1998). The value of choice (Scanlon argues) would not be undermined by deterministic or indeterministic outside causes. As long as the causes ‘affect our responses [to situations of choice] only by affecting what we are like’, our choices may perform their instrumental, representative and symbolic functions as well as if they were not subject to causal processes beyond our control (Scanlon 1998: 255–256). These various Strawson-inspired approaches rest on some distinction between ‘brute facts’ of nature on the one hand, and human, social, normative practices on the other. They seek not only to explain social practices of responsibility but also to justify them. This they do typically by finding some forward-looking, consequentialist objective, such as improving reasons–responsiveness, that is intrinsic to the logic of the practices themselves in the sense that it reflects the value we place on those practices. Ironically, however, it is precisely such a strategy that Strawson himself sought to avoid. Instead of finding the justification for responsibility practices in their utility he found it in facts about human nature and psychology.11 In appealing to the psychological impossibility, as he saw it, of abandoning our responsibility practices, Strawson may be understood as espousing a rigorous naturalism that understands human behavior generally, human social behavior more particularly, and normative, social behavior most particularly, as themselves the result of processes that conform to the same brute facts of physics as underlie the non-human, ‘natural’ world. In other words, in seeking to defend social responsibility practices against naturalistic skepticism based on facts of physics, Strawson can be understood as explaining responsibility in naturalistic terms of human biology and psychology that need not be inconsistent with the very facts of physics that the sceptics think threaten to undermine our social responsibility practices. Put differently, Strawson may be read as challenging the assumption of discontinuity between the ‘natural’ world of physics and chemistry on the one hand, and human social and cultural (including legal) life on the other, which seems implicit in much traditional naturalism. 461

Peter Cane

Taking this approach, the challenge is to explain how normative, social (or ‘institutional’) behavior may emerge out of ‘brute (‘pre-institutional’) facts’ about human biology and psychology. The idea that human beings and their social practices, including their normative practices, are part of the natural world, not separate from it, was influentially developed by J.Q. Wilson (Wilson 1993) who, in the words of Scott Soames (Soames 2021: 352), argues that ‘we have a moral sense consisting of a complex set of social and biological dispositions . . . which . . . can . . . provide a factual basis relevant to the moral assessment of agents . . . in widely different circumstances’.12 More recently, Philip Pettit (Pettit & Hoekstra 2018; Pettit 2019; see also Delacroix 2017) has taken inspiration from H.L.A. Hart’s ‘rational reconstruction’ (Ullmann-Margalit 1977: 2–3) of a transition from a ‘pre-legal’ to a ‘legal’ society produced by human behavioral adaptations to changing environmental and social conditions (Hart 2012: 91–99, critically assessed in Gardner 2013). Pettit goes further back in time to offer a speculative account of a transition from a pre-normative to a normative society: in Pettit’s elegant terminology, from Pre-Normitania to Normitania to Lexitania or, more colloquially, from a world of behavioral regularities (‘habits’, ‘customs’) to a world of social, behavioral norms (‘morality’ and so on) and thence to a world of highly institutionalized behavioral norms (‘law’). Michael Tomasello offers a more anthropological (but philosophically informed) account of the evolution of human morality (Tomasello 2016, 2018). Both Pettit and Tomasello understand normativity as a product of the operation of human capacities and propensities in specific circumstances of time, place, climate and so on. By reason of the the capacity and propensity (amongst others) to understand the world in terms of inter-personal values and norms, humans have been able to develop unique forms of social interaction. On such an understanding, normativity is a distinctively human achievement. Moreover, law and morality are both products of human normative behavior. Such evolutionary approaches rest on a fundamentally different understanding of the relationship between fact and value than is assumed by the distinction between ‘brute facts of nature’ and human, social practices. This distinction imagines nature and social practice as separate and discontinuous such that human practices may be judged ‘irrational’ and, hence, ‘morally’ unjustifiable in the light of apparently inconsistent ‘brute facts’. By contrast, the evolutionary approach assumes that human nature and human practices are ultimately the product of (and in an important sense, continuous with) the very same physical processes and regularities as so-called ‘brute facts’. On this assumption, the human capacity and propensity to frame and understand the world in terms of values and norms, and practices that express and embody that capacity and propensity must, in principle and ultimately, be as explicable by ‘brute facts’ as any other feature of the material world. If the assumption of ultimate continuity between brute facts of nature and human social practices, including normative practices, is accepted, brute facts about physical, causal processes do not present a challenge to such practices but are an essential part of the explanation of such practices. Of course, we may never fully understand the physics, chemistry, biology and psychology of the human capacity and propensity (or ‘inclination’) to frame the world in terms of value and to assess human behavior in accordance with such values. However, once we accept that this must be possible in principle, the foundation of skepticism about moral responsibility based on ‘brute facts of nature’ is undermined.

462

Legal and moral responsibility

Reassessing the relationship between law and morality and between responsibility in law and morality According to the evolutionary approach, law, morality and other normative systems are products of the human capacity and propensity to understand the world in terms of value. As far as we know, this is a distinctively human, species-universal characteristic. Amongst humans, inability to understand the world in terms of value is considered a mental disability. Because the characteristic is species-universal, one might expect that it would, in some respects, manifest itself universally in all human societies. On the other hand, in the light of current understandings of natural and cultural evolution, there are reasons to expect that its manifestations will vary from one society to another, or even within societies, as basic human capacities and dispositions interact with specific ‘environmental’ conditions. So, there is no reason to object, in principle, to the analytical philosophers’ search for responsibility universals, present in human societies at all times and in all places. However, the reason to allow the possibility of such universals is very different under the evolutionary approach than under the analytic–philosophical approach. Under the latter approach, the existence of ‘moral’ universals depends on the assumption that such universals are independent of human practices; and the diversity amongst laws in different places and at different times is explicable precisely because it is a social practice. By contrast, under the evolutionary approach, the existence of universals depends on uniformity of normative practice across all societies and at all times. Under the philosophical approach, morality and law differ not only in their respective natures (‘ontologically’) but also in terms of how we acquire knowledge of them (‘epistemologically’). Under the evolutionary approach, morality and law are, ontologically and epistemologically, essentially similar. How, within an evolutionary framework, might the relationship between law and morality be best construed? We must start with the observation that widespread social practice does, indeed, distinguish between law and morality. Moreover, morality is often used as a benchmark against which to assess either the law’s validity (in the natural law tradition) or the existence and strength of obligation of obedience to the law, which the law itself demands (in the positivist tradition).13 In the positivist tradition, Hart encapsulated this point in a distinction between ‘positive’ morality and ‘critical morality’. In his famous debate with Lord Patrick Devlin sparked by reform, in the late 1950s and the 1960s, of the criminal law relating to homosexual acts between consenting males in private, Devlin argued that society had a right to use the criminal law to enforce moral values (Cane 2006). Hart countered by accusing Devlin of confusing a society’s actual, ‘positive’ morality with ‘critical morality’: certainly, society is entitled to enforce the right moral principles through law; but it is always an open question whether society’s positive morality coincides with critical morality. Importantly, Hart was non-committal about the metaphysics of morality: he refused to pronounce on whether morality was practice-dependent or practice-independent  – and, if the latter, as to whether it was mind-dependent or mind-independent (Hart 2012: 168, 253–254). Unfortunately, this agnosticism undercuts the distinction between positive and critical morality because genuine, reasonable and persistent disagreement about critical morality disqualifies it as a benchmark against which to judge the law. Whereas a society’s positive morality may display diversity, critical morality cannot if it is to perform the role Hart assigned it. For that, critical morality must be univocal. Given what we know about human psychology, the only way to guarantee univocality of value judgments is to reduce the number of judges to one. Whereas members of society

463

Peter Cane

may, and typically will, recognize some institution akin to Hart’s idea of positive morality, only an individual can espouse a critical morality in the way that Hart conceived it. From his perspective, we might say, the function of (critical) morality in the lives of individuals is to provide them with a shield against what they consider to be unacceptable, value-based demands from other individuals or from society and to allow the individual to assert his or her values against all others (Cane 2012; for discussion see Tasioulas 2013: 172–173). In this way of thinking, the human inclination and propensity to understand the world in terms of values is a characteristic of individuals. That characteristic, coupled with the ability to visualize other minds, plays a central role in facilitating human cooperation and social life by the creation of social norms. Successful social life may require individuals to sacrifice their own values to those of others in order to realize the benefits of sociality. Human beings are, at one and the same time, unique individuals and social animals. Social life may demand some sacrifice of individuality. However, there may come a point where the sacrifice required is so great that it would, effectively, destroy the individual’s sense of their own uniqueness. Critical morality, on this view, provides individuals with means to prevent their complete absorption into the social entity. Such an approach can explain why ‘morality’ is treated as the ultimate arbiter of value, immune to the demands of others and of society. From this point of view, (critical) morality and law are ontologically similar and epistemologically related. Ontologically, both are expressions of the human inclination and propensity to frame the world in terms of values. Epistemologically, they are related because individuality and sociality are both basic human values that are expressed and discernible in human behavior. Personal values and norms express our individuality, and social values and norms express our sociality. It is of the very essence of social norms that they may not coincide with an individual’s personal norms. However, for the sake of sociality, individuals may and often do comply with norms inconsistent or in tension with their own values. Put differently, as human beings we are inclined to incorporate social norms and values into our personal norms and values. For most people most of the time, their personal norms do not discourage or prevent them from complying with social norms because their personal values include sociality, and many of our personal values are learned by social interaction. At the margin, however, compliance with social norms may be seen as so threatening to individuality that personal norms trump social norms. Since sociality is amongst the values of human beings as a species, social norms, including legal norms, have personal value for individuals. From this perspective, the common idea that law is, at best, a more-or-less (un)successful attempt to track morality, unworthy of careful attention in its own right, seems wrongheaded. Law is a species of human normative behavior that is significantly continuous with other species of human normative behavior, such as ‘morality’. Understanding human normative behavior depends as much on taking law seriously as on taking morality seriously.

Notes * Emeritus Distinguished Professor, The Australian National University; Honorary Professor of Law, University of Sydney. Many thanks to Max Kiener for penetrating questions and comments, and extremely helpful suggestions. I owe a huge debt to Tony Connolly for more than 20 years of philosophical instruction, discussion and companionship. This essay is for him. 1 I focus on English law and leave readers to decide the strength of the argument in relation to other systems of law.

464

Legal and moral responsibility 2 The foil for my argument is mainstream, Anglophone moral philosophy of responsibility, which may be considered to have been kicked off by Strawson 1962. 3 As opposed to literature written primarily for legal practitioners and students studying for a first law degree. 4 For more detailed treatment of this and other issues discussed in this section see Cane 2002. 5 Vicarious liability is civil liability for conduct of another based merely on the existence of a relationship between the two parties. The employer–employee relationship is the one to which vicarious liability (of the employer) most commonly attaches. By contrast, accessory liability is liability for the agent’s own conduct in ‘aiding and abetting’ or encouraging or facilitating conduct of another. Accessory liability is rare but not unknown under the civil law model. But, for instance, aiding and abetting a tort is not itself a civil wrong. 6 See previous note. 7 Cane 2002: 44–49. 8 The classic text of modern, Anglophone legal positivism is Hart 2012. The Introduction and Postscript to this edition explore more recent debates, which have become highly technical. 9 Finnis 1980 is a classic statement of a modern natural law position. 10 Hard sceptics argue that even if, as modern physics suggests, the world is, at some level, causally indeterministic or ‘random’, still humans would lack the control necessary for moral responsibility. 11 Around the same time, Hart appealed to facts about human nature in his account of ‘the minimum content of natural law’ (Hart 2012: 193–200). 12 There is, of course, a huge literature. I  have found the following items especially useful: Opp 1982; Oldenquist 1990; Tooby & Cosmides 1992; Laland et al. 2000; Ostrom 2000; Runciman 2001; Bradie 2007; Searle 2010; Laland & Brown 2018; Krebs 2020; Grüneisen & Wyman 2020; Reader & Hughes 2020. 13 On the positivist and natural law traditions see text at notes 7 and 8.

Further reading Strawson 1962 and Hart 1968 (211–230) are highly generative starting points for modern sociolegal theorising about responsibility. My approach in Cane 2002 (on which the first part of this essay is based) was much influenced by both. In The Concept of Law (Oxford: Clarendon Press, 1961) Hart also provided material for constructing a social theory of morality and its relationship to law. My approach to morality, elaborated in Cane 2012, builds on Hart’s discussion. In the 1961 book, too, Hart told a speculative story about how and why ‘law’ might have arisen in a human society regulated by ‘customary’ norms. Philip Pettit adds philosophical rigour to Hart’s story in Pettit 2019. The now-enormous literature in evolutionary psychology on the emergence of religion, ‘morality’ and normative behaviour more generally, complements (and, to my mind, grounds) the philosophical approach. Tomasello 2016 is an excellent starting point. The development of law in many human societies in the past 4,000 years and more is explored in Fernanda Pirie’s extraordinary book, The Rule of Laws (London: Profile Books, 2021).

References Bradie, M. (2007). “Evolution and Normativity,” in M. Matthen & C. Stephens (eds.), Philosophy of Biology. Amsterdam: North-Holland/Elsevier. Cane, P. (2002). Responsibility in Law and Morality. Oxford: Hart Publishing. ———. (2006). “Taking Law Seriously: Starting Points of the Hart/Devlin Debate,” The Journal of Ethics, 10, 21. ———. (2012). “Morality, Law and Conflicting Reasons for Action,” Cambridge Law Journal, 71, 59–85. ———. (2016). “Role Responsibility,” The Journal of Ethics, 20, 279–298. Caruso, G. (2018). “Skepticism About Moral Responsibility,” Stanford Encyclopedia of Philosophy. https://plato.stanford.edu/entries/skepticism-moral-responsibility/.

465

Peter Cane Connolly, A. J. (2021). “Philosophical and Judicial Thinking About Moral Concepts: Cane’s Critique of Philosophical Method 20  Years On,” in J Goudkamp, M. Lunney,  & L. McDonald (eds.), Taking Law Seriously: Essays on Honour of Peter Cane (pp. 305–331). Oxford: Hart Publishing. Cullity, G. (1997). “Book Review: Responsibility and the Moral Sentiments,” Mind, 106, 803–807. Delacroix, S. (2017). “Law and Habits,” Oxford Journal of Legal Studies, 37, 660–686. Finnis, J. (1980). Natural Law and Natural Rights. Oxford: Oxford University Press. Gardner, J. (2013). “Why Law Might Emerge: Hart’s Problematic Fable,” in L. D. D’Almeida, J. Edwards, & A. Dolcetti (eds.), Reading HLA Hart’s the Concept of Law. Oxford: Hart Publishing. Grüneisen, S., & Wyman, E. (2020). “The Ontogeny and Evolution of Cooperation,” in L. Workman, W. Reader,  & J. H. Barkow (eds.), The Cambridge Handbook of Evolutionary Perspectives on Human Behaviour (pp. 265–275). Cambridge: Cambridge University Press. Hart, H. L. A. (1968). Punishment and Responsibility: Essays in the Philosophy of Law. Oxford: Clarendon Press. ———. (2012). The Concept of Law (3rd edn with an Introduction by L. Green). Oxford: Oxford University Press. Krebs, D. L. (2020). “Can Evolutionary Processes Explain the Origins of Morality?” in L. Workman, W. Reader,  & J. H. Barkow (eds.), The Cambridge Handbook of Evolutionary Perspectives on Human Behaviour (pp. 139–149). Cambridge: Cambridge University Press. Laland, K. N., & Brown, G. R. (2018). “The Social Construction of Human Nature,” in E. Hannon  & T. Lewens (ed.), Why We Disagree About Human Nature. Oxford: Oxford University Press. Laland, K. N., Odling-Smee, J., & Feldman, M. W. (2000). “Niche Construction, Biological Evolution, and Cultural Change,” Behavioral and Brain Sciences, 23, 131–175. McGeer, V., & Pettit, P. (2015). “The Hard Problem of Responsibility,” in D. Shoemaker (ed.), Oxford Studies in Agency and Responsibility, Volume 3 (pp. 161–187). Oxford: Oxford University Press. McKeon, R. (1957). “The Development and the Significance of the Concept of Responsibility,” Revue Internationale de Philosophie, 11, 3–32. Oldenquist, A. (1990). “The Origins of Morality: An Essay in Philosophical Anthropology,” Social Philosophy and Policy, 8, 121–140. Opp, K-D. (1982). “The Evolutionary Emergence of Norms,” British Journal of Social Psychology, 21, 139–149. Ostrom, E. (2000). “Collective Action and the Evolution of Social Norms,” Journal of Economic Perspectives, 14, 137–158. Pettit, P. (2019). “Social Norms and the Internal Point of View: An Elaboration of Hart’s Genealogy of Law,” Oxford Journal of Legal Studies, 39, 1–30. Pettit, P., & Hoekstra, K. (2018). The Birth of Ethics: Reconstructing the Role and Nature of Morality. Oxford: Oxford University Press. Reader, W., & Hughes, S. (2020). “The Evolution and Function of Third-Party Moral Judgment,” in L. Workman, W. Reader, & J. H. Barkow (eds.), The Cambridge Handbook of Evolutionary Perspectives on Human Behaviour (pp. 150–157). Cambridge: Cambridge University Press. Runciman, W. G. (2001). “From Nature to Culture, from Culture to Society,” in W. G. Runciman (ed.), Human Social Institutions. Oxford: Oxford University Press. Scanlon, T. M. (1998). What We Owe to Each Other. Cambridge, MA: Belknap Press. Searle, J. R. (2010). Making the Social World: The Structure of Human Civilization. Oxford: Oxford University Press. ———. (2017). “Response-Dependent Responsibility: Or, A Funny Thing Happened on the Way to Blame,” Philosophical Review, 126, 481–527. Sie, M. (2018). “Sharing Responsibility,” in K. Hutchison, C. Mackenzie, & M. Oshana (eds.), Social Dimensions of Moral Responsibility. Oxford: Oxford University Press. Soames, S. (2021). The World Philosophy Made: From Plato to the Digital Age. Princeton and Oxford: Princeton University Press. Strawson, P. F. (1962). “Freedom and Resentment,” in P. F. Strawson (ed.), Freedom and Resentment and other essays (pp. 1–25). London: Methuen & Co. Talbert, M. (2019). “Moral Responsibility,” Stanford Encyclopedia of Philosophy. https://plato.stanford.edu/entries/moral-responsibility.

466

Legal and moral responsibility Tasioulas, J. (2013). “Hart on Justice and Morality,” in L. D. D’Almeida, J. Edwards, & A. Dolcetti (eds.), Reading H.L.A. Hart’s the Concept of Law. Oxford: Hart Publishing. Tomasello, M. (2016). A Natural History of Human Morality. Cambridge, MA: Harvard University Press. ———. (2018). “Précis of a Natural History of Human Morality,” Philosophical Psychology, 661–668. Tooby, J.,  & Cosmides, L. (1992). “The Psychological Foundations of Culture,” in J. H. Barkow, L. Cosmides, & J. Tooby (eds.), The Adapted Mind: Evolutionary Psychology and the Generation of Culture. New York: Oxford University Press. Ullmann-Margalit, E. (1977). The Emergence of Norms. Oxford: Clarendon Press. Vargas, M. (2013). Building Better Beings: A Theory of Moral Responsibility. Oxford: Oxford University Press. Wallace, R. J. (2014). “Emotions and Relationships: On a Theme from Strawson,” in D. Shoemaker & N. Tognazzini (eds.), Oxford Studies in Agency and Responsibility, Volume 2 (pp. 119– 142). Oxford: Oxford University Press. ———. (2014). “Peter Strawson on Responsibility and Sociality,” in D. Shoemaker & N. Tognazzini (eds.), Oxford Studies in Agency and Responsibility, Volume 2. Oxford: Oxford University Press. Wilson, J. Q. (1993). The Moral Sense. New York: Free Press.

467

37 THE VOLUNTARY ACT REQUIREMENT John Hyman

A defendant’s criminal liability generally depends on two related factors: their conduct and their concurrent state of mind. Both of these elements of a criminal offense raise interesting philosophical questions, because both are defined by means of concepts that are difficult to explain precisely, such as intention, knowledge, action and causation. This chapter is about the conduct element, the actus reus, and in particular the Voluntary Act Requirement (VAR), the controversial principle that a person can only be criminally liable for the performance of a voluntary act.

The empiricist background of the VAR The attraction of the VAR is that it reduces several plausible constraints on criminalization to a single principle, for a person surely cannot, or at least should not, be criminally liable for acts they do when they are under hypnosis, unconscious, or asleep, for a seizure or reflex movement, or for their most wicked and dangerous desires or intentions; and perhaps it is plausible that the reason is the same in every case, viz., that none of these phenomena qualifies as a voluntary act.1 But this should only be considered an attraction if the best explanation of all of these constraints is the unitary explanation advocates of the VAR propose. The objection to the VAR this chapter will be mainly concerned with is that this is not in fact the case. There are three main objections to the VAR. (i) There are kinds of conduct that are legitimately criminalized despite not involving a voluntary act. (ii) There is no convincing justification for treating automatism as a denial of actus reus while insanity is treated as a defense.2 (iii) The reasons for restricting criminal liability to activity are different from the reasons for restricting criminal liability to voluntary conduct. The first objection is that there are kinds of conduct that are legitimately criminalized despite not involving a voluntary act, including crimes of possession (e.g., of obscene DOI: 10.4324/9781003282242-50 468

The voluntary act requirement

materials, weapons, or drugs) and crimes of omission (e.g., a parent’s failing to feed or care for their child). This line of argument has been thoroughly explored, and I shall not consider it further here, except to say that the criminal law does not in fact treat the performance of an act as a necessary condition for the commission of a crime, unless the term “act” is interpreted – as Bentham (1996: 75ff.) interpreted it – in such a way as to include omissions, that is, failures to act.3 The second objection is that in the case of automatism, for example, there is no convincing justification for treating the claim that a defendant’s conduct was not voluntary as a denial of actus reus while insanity is treated as a defense, or potentially as a denial of mens rea.4 Furthermore, automatism is generally subject to a prior fault exception. For example, Scots law requires that automatism is neither “self-induced” nor caused in a way that the defendant was “bound to foresee.”5 But this is incompatible with treating automatism as a denial of actus reus. Compare intoxication. A defendant’s culpability certainly can depend on whether they became intoxicated voluntarily or whether their drink was spiked without their knowledge. But the question of whether their behavior qualifies as, say, driving a vehicle, cannot so depend. For if they were driving a vehicle, that must remain true even if it transpires that they were involuntarily intoxicated at the time. Hence, involuntary intoxication must be regarded as a defense rather than a denial of actus reus, and the same applies to automatism. The third objection to the VAR is that the reasons for restricting criminal liability to activity are different from the reasons for restricting criminal liability to voluntary conduct; and even the reasons for restricting criminal liability to activity seem to differ depending on whether activity is contrasted with inactivity or with thought. Thus, the reason for not criminalizing involuntary behavior might be that it is not normally under an individual’s control; the reason for not criminalizing inactivity or omission, or at least doing so with particular caution, might be that “responsibilities to prevent harm [are not] as universal and stringent as [. . .] responsibilities not to do harm” (Duff 2007: 161); and the reason for not criminalizing thought might be that it is not sufficiently harmful to deserve punishment, or that criminalizing it would represent an unwarranted intrusion on an individual’s privacy. The VAR therefore seems to bundle together a variety of limits or restrictions on legitimate criminalization rather than providing the unitary theoretical basis upon which they all stand. As we shall see, the main cause of this bundling together is the empiricist model of a voluntary act as a bodily movement caused by a “volition” or “act of will,” which was introduced into jurisprudence by Bentham and Austin, and has not yet been completely purged from it, despite the compelling criticisms the model was subjected to in the midtwentieth century, especially by Wittgenstein and Ryle.6 Empiricist philosophers from Locke to Mill maintained that a voluntary act is a movement of the agent’s body caused by an act of will. Locke called the act of will a “volition,” whereas Bentham and Mill preferred the term “intention,” but their conception of it was the same. The volition or intention was a kind of conscious choosing or deciding, not merely an appetite or aversion, but a sui generis operation of the will. Here is Ryle’s presentation of the empiricist theory in The Concept of Mind. Ryle did not accept the theory; on the contrary, he was one of its most trenchant critics. But he describes it well: I think of some state of affairs which I wish to come into existence in the physical world, but, as my thinking and wishing are unexecutive, they require the mediation of 469

John Hyman

a further executive mental process. So I perform a volition which somehow puts my muscles into action. Only when a bodily movement has issued from such a volition can I merit praise or blame for what my hand or tongue has done. (Ryle 1949: 50) Ryle’s description of the empiricist theory brings out two points. First, the theory implies that every act is voluntary. It is generally agreed that a muscular spasm is not an act – at any rate, it is not an act that is attributable to a person as the agent, any more than the contraction of the heart or the peristaltic movement of the gut. But if the reason for this is that a spasm is not caused by a volition, and if being caused by a volition makes a movement voluntary, then no movement that is not voluntary is an act. Hence, if the empiricist theory is accepted, an “act requirement” and a “voluntary act requirement” come to the same thing. Second, volitions simultaneously play two roles in the empiricist theory: they explain both how active movements of the body differ from passive ones, such as spasms, and how the movements for which a person is morally responsible (sc. eligible for praise or blame) differ from the movements for which they are not morally responsible. Metaphorically, they explain the difference between the movements in respect of which a person is a perpetrator and the movements in respect of which they are a victim. Hence, if the empiricist theory is accepted, and if acts for which the agent is not morally responsible should not be criminalized, then a voluntary act requirement makes good sense. The empiricist theory of voluntary action and the equation it implies between action and voluntary action were taken for granted by jurists from John Austin to Glanville Williams and are still widely (although not universally7) accepted.8 But as we shall see, voluntary action is in fact not a single phenomenon with a single definition, but a combination of two phenomena that need to be defined separately: voluntariness and action. This explains the disunity of the justifications for the restrictions on legitimate criminalization bundled together in the VAR and explains also why the list of acts or movements the VAR is designed to exclude includes both movements that are not acts at all (spasms and convulsions) and acts that are not voluntary (sleepwalking and hypnosis).

Neuroscience and the VAR The relationship between the VAR and the empiricist theory of voluntary action is brought out nicely by a recent debate concerning the bearing of neuroscience on the VAR. The debate was specifically concerned with experiments conducted by Benjamin Libet in the 1980s, which appeared to show that activity in the motor cortex initiates the movement of a limb before the agent is consciously aware of an intention or volition to make the movement (Libet 1999). Libet’s result is controversial, but if we assume that his claim about the priority of activity in the motor cortex is true, what are the implications for the criminal law? On the one hand, Denno argues that the VAR in general and the formulation of it in the American Model Penal Code (MPC) in particular will need to be substantially revised. “Libet’s research,” she claims, “confirms that there appears to be no sound scientific basis for the MPC’s dichotomy between voluntary and involuntary behavior” (Denno 2002: 328). On the other hand, Pardo and Patterson claim that “the empirical results do nothing to

470

The voluntary act requirement

undermine the existing legal framework” (Pardo & Patterson 2015: 125). As we shall see, these opposing views about the bearing of Libet’s research on the VAR are based on different definitions of voluntary action. Article 2 of the MPC states that “A person is not guilty of an offense unless his liability is based on conduct which includes a voluntary act or the omission to perform an act of which he is physically capable” (§ 2.01). However, the MPC does not define a voluntary act. The article explicitly excludes reflexes and convulsions; acts that are done while the agent is asleep or unconscious; acts that are done under hypnosis or post-hypnotic suggestion; and “bodily movement[s] that [are] not a product of the effort or determination of the actor.” And the commentary on the article says that voluntary conduct is “within the control of the actor.” But a definition of a voluntary act is not attempted in either place. Both Pardo and Patterson and Denno fill this gap, but they do so in different ways. Pardo and Patterson write as follows: A minimal level of control over one’s actions will make them voluntary for purposes of the criminal law. The type of control necessary to make an act voluntary in this minimal sense depends on a type of two-way power; the agent not only has the power to act but also the power to refrain from so acting. (Pardo & Patterson 2015: 123) By contrast, Denno writes: “voluntary actions or acts have three key elements: (1) an internal event, or volition; (2) an external physical demonstration of that volition; and (3) causal connection between the internal and external elements” (Denno 2002: 275–276). She adds the following comment in a note: The enigma of voluntariness is captured in Wittgenstein’s famous question: “[W]hat is left over if I subtract the fact that my arm goes up from the fact that I raise my arm?” . . . What is “left over,” then, must be voluntariness, which implies the actor’s control.  .  .  . Harry Frankfurt argues that voluntariness and responsibility do not require that the agent be able to act otherwise. (Denno 2002: 276n26; cf. Moore 1993: 134) Thus, according to both accounts voluntary acts are under the agent’s control. To this extent they agree, and to this extent they both use language that appears in the MPC. But Denno includes the empiricist theory of voluntary action in her account of the existing legal framework, and excludes the ability to act otherwise, whereas Pardo and Patterson do the reverse. They include the ability to act otherwise and exclude the empiricist theory. But neither of these terms – neither “volition” nor “ability to act otherwise” – is used in the MPC. This explains why they disagree about Libet’s research. Pardo and Patterson are right in thinking that it does not undermine the existing legal framework if voluntary action is defined in terms of the ability to act otherwise, whereas Denno is right that it does undermine it if we accept the empiricist theory of voluntary action. I shall enlarge on this now, commenting on Pardo and Patterson first. As we have seen, what Libet’s research showed, or is thought to have shown, is that activity in the motor cortex initiates the movement of a limb before the agent is consciously

471

John Hyman

aware of an intention or volition to make the movement. But if we assume that this is true – even if we assume that it is also true outside Libet’s experimental situation – it does not follow that an agent who makes a movement does not have the ability to act otherwise; in other words, it does not follow that he is incapable of keeping still. For if a person who makes a certain kind of movement was able at the time to refrain from making it, it does not follow that the movement was caused by a volition or intention, regardless of whether the empiricist theory of voluntary action is correct. For example, sneezes are not caused by volitions or intentions, but we are sometimes able to inhibit a sneeze. As a matter of fact, as Pardo and Patterson point out, our daily lives are filled with purposeful movements that are not preceded by a conscious awareness of the intention to make them, either before or after the activity in the motor cortex initiates them. They mention the movements that occur when we answer the phone or type a sentence (Pardo & Patterson 2015: 128). But it does not follow that we cannot help making these movements. In fact, we know quite well which kinds of movements we cannot help making, at least in most cases. They include the movements mentioned in the MPC: reflexes, convulsions, and so on. But they do not include the movements we make in ordinary circumstances when we answer the phone. Hence, Libet’s research leaves the MPC provision unscathed if voluntary action is defined in terms of the ability to act otherwise. However, it does undermine the existing legal framework if the empiricist theory of voluntary action is accepted. For if a voluntary act is a bodily movement caused by a volition, then Libet’s research does prove that at least in the experimental situation he devised, a subject’s movement of a limb is not a voluntary act, because the volition cannot initiate the movement if it happens after the movement has been initiated by the motor cortex. Can we defend the existing legal framework by distinguishing between initiation and causation? Again, not if we accept the empiricist theory of the will. Remember, the whole point of postulating volitions was that “as my thinking and wishing are unexecutive, they require the mediation of a further executive mental process.” Libet himself suggested that there is sufficient time for the agent to prevent a movement from occurring after their conscious awareness of the intention occurs (Libet 1999: 51–52). But he estimated that this window of opportunity is less than two-tenths of a second. Considering that the mean reaction times for sprinters at the Beijing Olympics were seventeen-hundredths of a second for men and nineteen-hundredths of a second for women, this is evidently a challenge. It is true that these figures include the time it takes to process an auditory signal, but even so, once we have reached middle age we do not stand a chance. Hence, if the empiricist theory of voluntary action is right, then it is plausible, as Denno says, that the VAR is not acceptable as it stands, because it is based on outdated science.

Two definitions of voluntary action Which definition of voluntary action should we accept, Denno’s or Pardo and Patterson’s? The answer, I submit, is neither. To recall, Denno defines a voluntary act as a bodily movement caused by a volition, and she concludes that Libet’s research undermines the VAR, because it proves that bodily movements are not caused by volitions. But we should only accept this conclusion if we regard the empiricist theory of voluntary action as irrefutable. Otherwise, we should regard Libet’s research as a refutation of the theory, or at least as evidence that it is false. 472

The voluntary act requirement

This is the conclusion reached by some of Libet’s followers. For example, Daniel Wegner accepts that Libet’s research shows that volitions do not cause movements: It seems that conscious wanting is not the beginning of the process of making a voluntary movement but rather one of the events in a cascade that that eventually yields such movement . . . in fact it might not even be that. It might just be a loose end – one of those things, like the action, that is caused by prior brain and mental events. (Wegner 2018: 52)9 But Wegner does not conclude that there is “no sound scientific basis for the dichotomy between voluntary and involuntary behavior,” because he does not define voluntary behavior in terms of conscious wanting or volition. He says that the simplest definition of voluntary action may be this: “A voluntary action is something a person can do when asked” (Wegner 2018: 30). This is not a satisfactory definition of voluntary action, because of a combination of two reasons. First, something a person can do when asked – such as smile or blink or make a fist – is a kind of act, not a particular act; hence the definition implies that kinds of acts – e.g., sneezing, blinking, or making an omelette – are either voluntary or not voluntary. But, second, there are many kinds of act some instances of which are voluntary and others not. For example, we can blink voluntarily, and can also normally blink when asked, but blinking is also commonly an involuntary reaction. Hence, while it is true that if someone does an act because they were asked to do it, they do it voluntarily in most cases, this does not provide the basis for a definition of a voluntary act.10 Nevertheless, Wegner draws the right conclusion from Libet’s research, namely, that volitions do not cause bodily movements. What about Pardo and Patterson’s definition of voluntary action? Can voluntariness be defined in terms of the ability to do otherwise? The answer is that it cannot, because if a person does something voluntarily, it does not follow that it is possible for them to avoid doing it. Frankfurt invented several cases that prove this, as Denno points out, but the simplest and clearest case is due to Locke: Suppose a man be carried, whilst fast asleep, into a room, where is a person he longs to see and speak with; and be there locked fast in, beyond his power to get out: he awakes, and is glad to find himself in so desirable company, which he stays willingly in, i.e. prefers his stay to going away. I ask, is not this stay voluntary? I think, nobody will doubt it; and yet being locked fast in, ’tis evident he is not at liberty not to stay, he has not freedom to be gone. (Locke 1997: 2.21.10) Locke is surely right in saying that the man stays in the room voluntarily: as he puts it, “voluntary is not opposed to necessary; but to involuntary” (Locke 1997: 2.21.11). But we need to be clear about the reason why. The reason is not that he stays willingly, or that he does not regret staying; because if he stays reluctantly, or out of politeness, and does regret staying, it remains true that he stays voluntarily. The true reason is that he is not actually prevented from leaving the room by the locked door, even though he would be prevented if he tried. If we replace the lock in the story with a guard stationed outside the door, this becomes more obvious. The guard is there to prevent the man from leaving, in other words, to compel him to stay, should that be necessary. 473

John Hyman

But only the guard’s actual intervention, or the lock’s actual resistance to the pressure the man applies to it, prevents him from leaving, and of course this only happens if he attempts to leave. Similarly, if an adult of sound mind freely consents to sex without being aware that her partner would have raped her if she had refused, she has sex voluntarily, since she is not actually compelled to do so by force or threats. Again, this remains true whether she consents willingly or reluctantly and whether she regrets doing so or not. It follows that we should not accept either the empiricist theory of voluntary action or the ability to act otherwise theory.

A different approach What then is the correct approach to understanding voluntary action? One approach is to modify the empiricist theory of volitions by discarding the requirement that a volition is conscious. Doing so would take the sting out of Libet’s research – i.e., it would no longer prove that either voluntary action is impossible or the empiricist theory of voluntary action is false. Another approach is to interpret “voluntary” movements – at least for the purposes of the VAR  – as ones the agent is physically able to control, and involuntary ones, such as “the involuntary running of urine” or “the involuntary closing of the eyelids when the surface of the eye is touched,” as ones the agent is not physically able to control (both examples are taken from the Oxford English Dictionary). But neither approach provides a basis for a VAR, for on either proposal, an uncontrollable reflex or a convulsion would fail to qualify as an actus reus, but an act done under hypnosis or while asleep would not fail to qualify. For example, consider the Canadian case R v. Parks ([1992] 2 S.C.R. 871). Kenneth Parks drove fourteen miles to his parents-in-law’s home, entered the house using a key they had given him and bludgeoned his mother-in-law to death. His defense team argued that he was in a state of non-insane automatism (i.e., he was sleepwalking), and he was acquitted. But if this was the right verdict, the reason is not that he was not in control of the movements of his body, or that they were not caused by volitions. We do not lose all physical control of our bodies when we fall asleep, otherwise the involuntary running of urine would be a more common problem than it is. In fact, Parks’s movements clearly were controlled and purposive. He could not have driven fourteen miles, let himself into the house, etc., had they not been. If the court was right to acquit Parks, the reason is not that he was not able to control the movements of his body, or that they were not caused by volitions; it is that he was unable to control his will – in other words, the expression in action of his desire to kill his in-laws – and was unable to be guided by the moral and legal reasons to desist. This is consistent with Hart’s dictum that “unless a man has the capacity and a fair opportunity or chance to adjust his behavior to the law its penalties ought not to be applied to him” (Hart 2008: 181), but it suggests that Hart’s claim that “the controlling agency is ‘out of action’ when the agent is unconscious or asleep” (106) does not provide the right basis for an acquittal in cases of automatism such as Parks. Weinreb (1970: 112) and Husak (2007: 2454 and passim) among others adopted and developed Hart’s suggestion that the actus reus requirement should be couched in terms of control rather than voluntariness. But none of these authors distinguishes between the absence of control of bodily movements (e.g., spasms or reflexes) and the absence of control of the “will,” or impulse control, i.e., control of the expression of emotion and desire in action (e.g., Parks, victims of 474

The voluntary act requirement

frontal lobe injury such as Phineas Gage, instances of killing as a result of provocation, and of course many violent crimes and sex crimes of which the defendant is justly convicted). The former can plausibly be regarded as negating actus reus whereas the latter cannot.11 The right approach to understanding voluntary action involves a more radical departure from the empiricist theory than we have considered so far, and it begins with the recognition that voluntary action is not a single phenomenon, with a single definition, but a combination of two phenomena, which need to be defined separately: action and voluntariness.12 Consider action first. The concept of action is a highly abstract causal concept: to act is to make something happen, to cause some kind of change. So, a theory of action purports to explain the difference between movements in a person’s body they cause personally themselves, such as the movements of a person’s lips when they speak, or the contraction of their fist when they clench it, and movements in their body they do not cause personally, such as spasms and convulsions, or the contraction of the heart. Donald Davidson reinvigorated one element of the empiricist theory when he argued in the 1970s that intention is the mark of human agency. “A person does as agent whatever he does intentionally under some description” (Davidson 2001: 45–46). But the idea was well established by the nineteenth century, as we have seen. Thus, Reid (2010: 31) claims that if an act was done without the agent’s “will and intention,” then “it ought not to be imputed to him as the agent”; and Bentham (1996: 85) declares that “if the act be not intentional in the first stage [i.e., the bodily movement], it is no act of yours.” In fact there are several kinds of exceptions to the generalization that all human action is intentional under some description or stems from an intentional movement, including the following: automatic reactions such as ducking or drawing back one’s head to avoid a blow, or making an unconscious adjustment to one’s posture to maintain balance; some kinds of habitual action, including verbal tics such as echolalia; some kinds of uncontrolled action done in abnormal or pathological states of mind; unconscious action such as talking in one’s sleep; and the spontaneous expression of emotion in facial expressions, vocalizations, and gestures, such as smiling, scowling, pouting, shrugging, and laughing, or crying out with pleasure or pain. The truth is that most human acts that matter sufficiently to be reported or recorded stem from an intentional movement, with the important exception of spontaneous expressions of emotion. But intention is not definitive of individual human agency. Individual human agency, like the agency of any kind of organism, is defined by functional integration. For human beings, this means the integrated operation of cognitive and motor systems, but it does not necessarily involve intentions.13 Voluntariness is an entirely different concept. The defining characteristic of voluntary conduct is that it is not due to ignorance or compulsion. This is not an original or controversial claim, nor is it a stipulation. It is Aristotle’s definition of voluntariness in the Nicomachean Ethics (2009: 1109b35–1110a4), which is in effect the source (via Latin) of the word “voluntary.” But what is the connection between ignorance and compulsion? Why are these two very different phenomena, one of them cognitive and the other causal, bound together in the idea of voluntariness? The empiricist theory of voluntary action explains that this is because voluntary acts are caused by volitions, which are instances of consciously choosing or deciding or willing the act one does. Consciousness excludes ignorance and choosing or deciding or willing excludes compulsion. But if we reject the empiricist theory, we need a different answer. 475

John Hyman

The right answer does not refer to the causes of voluntary conduct. Indeed, there is no reason to think that voluntary conduct has a distinctive kind of cause, since perception, emotion, and desire can cause voluntary and involuntary behavior alike.14 The true reason why ignorance and compulsion are bound together in the idea of voluntariness is that they are normally exculpations. To be precise, they are exculpations unless they are themselves culpable, or result from culpable conduct by the individual concerned. So far as moral culpability is concerned, these exculpations are generally regarded as excuses; but in the criminal law, they include factors that negate the mens rea element of an offense, such as a mistake concerning part of the actus reus (under the heading of ignorance),15 as well as defenses including duress (under the heading of compulsion).16 In sum, to describe conduct as “voluntary” is to say that it was not due to factors that normally exculpate, factors that normally exclude guilt or free someone from blame. Thus, whereas action is a causal concept, voluntariness is an ethical concept. Its basic function is to inform the appraisal of individual conduct, and in particular the assessment of innocence and guilt.17 But the word “conduct” here does not just refer to action, and the idea of voluntariness applies just as much to omissions as it does to acts. In fact, it applies just as much to the reception of an act as it does to the performance of an act.18 For example, it can be just as important to know whether a person received a sex act or underwent a medical procedure voluntarily as it is to know whether someone did an act voluntarily. Action and voluntariness are profoundly different concepts, serving different intellectual needs. The basic error in the empiricist theory of voluntary action is that it tries to explain both phenomena at once. The modern idea of a volition was introduced into philosophy by Descartes, who postulated volitions to explain action and not voluntariness in his treatise Les Passions de L’Âme. But beginning with Locke, volitions (or “intentions”, as they were subsequently termed) were regarded by empiricist philosophers both as the source of individual agency and as the source of voluntariness. And the error was imported into jurisprudence together with the theory of volitions. Hence Austin defines an act as “a voluntary movement of my body.” Holmes writes: “An act is always a voluntary muscular contraction and nothing else. . . . An act implies a choice.” Williams defines an act as “a willed bodily movement,” adding, “since every act is by definition willed, there is no need to call it voluntary.” Moore defines an act in the same way as Williams, as a “willed bodily movement,” and states (now following Mill) that “human acts are themselves a compound of other events, consisting of a volition causing a bodily movement.”19 Here again is Ryle’s description of the empiricist theory of voluntary action: I think of some state of affairs which I wish to come into existence in the physical world, but, as my thinking and wishing are unexecutive, they require the mediation of a further executive mental process. So I perform a volition which somehow puts my muscles into action. Only when a bodily movement has issued from such a volition can I merit praise or blame for what my hand or tongue has done. (Ryle 1949: 50; cf. Reid 2010: 31) Ryle’s passage shows how the empiricist theory provides a rationale for the VAR, but without this rationale, there is no justification for the claim that an actus reus must be a voluntary act, and there is no reason to think that the various restrictions on legitimate criminalization that are bundled up in the MPC’s formulation and other canonical

476

The voluntary act requirement

formulations of the VAR have a unitary explanation. This is not to say that the concepts of action and voluntariness have no role to play in the criminal law. On the contrary, they are both pertinent, as we have seen, but independently. For the concept of action pertains to the actus reus, whereas the concept of voluntariness can pertain to offenses or defenses, and either to the actus reus or the mens rea element of the offense.

Conclusion As I noted at the outset, there are several plausible restrictions on legitimate criminalization, which are jointly thought to provide a rationale for a voluntary act requirement: thoughtcrimes such as heresy; behavior that occurs while a person is unconscious or asleep; movements of the body that are not attributable to the agency of the individual at all, such as reflex movements and convulsions; and (in a qualified sense) omissions. But the idea that the justification for each of these restrictions is the same – viz., that they are not voluntary acts – is based on an imaginary idea of volition as the root of both agency and responsibility, of both an individual’s active engagement with the world, and of the responsibility they bear for the events and circumstances in it that they cause. The MPC formulation of the VAR and other similar formulations purport to provide a unitary explanation of several plausible restrictions on legitimate criminalization, but the conception of voluntary action on which the VAR depends is a myth.

Notes 1 The American Model Penal Code (MPC; American Law Institute 1985), which includes a formulation of the VAR, is not designed as a constraint on criminalization, i.e., a principle excluding legislation that would create criminal liability in the absence of a voluntary act, but as a constraint on criminal liability itself, excluding conviction or punishment in the absence of a voluntary act. The MPC does not refer to thought-crimes, but jurists commonly justify the VAR in this way. See for example Denno (2002: 276) and Duff (2007: 95–96). 2 I describe insanity as a defense because it provides a reason why a defendant ought not to be convicted of a crime despite meeting both its conduct (actus reus) and mental state (mens rea) conditions. A defense may avert conviction by providing either a justification or an excuse. 3 Fletcher (1978: 421) argues in favor of this kind of definition whereas Moore (1993: 22–24) opposes it, and acknowledges that there are exceptions to the VAR for this reason. Duff (2007: 107–115) argues for an act, “presumption,” i.e., the principle that the criminalization of failures to prevent harm ought to be rarer than the criminalization of actively causing harm. 4 See Hogg 2022. The Law Commission’s 2013 discussion paper on insanity and automatism states that “the correct classification of the defense [of automatism] is as a denial of actus reus” (para 5.8). But as Hogg points out, this formulation is confused: “Either automatism amounts to a denial of actus reus, or it is an excusing defense. To describe it as a defense the basis of which is act negation is to obfuscate the conceptual distinction between actus reus/mens rea and defenses” (2022: 7). 5 Ross v. HM Advocate [1991] JC 210. 6 For criticism of the empiricist theory of voluntary action see Ryle (1949: ch. 3); Wittgenstein (1958: 151ff.); Hyman (2015: ch. 1). The most influential recent defense of the VAR explicitly incorporates this model (Moore 1993, 1997), but its influence is also evident in jurisprudence where it is not explicitly endorsed, and even where it is superficially opposed (e.g., Fletcher 1978: 420–439). 7 See for example Simester (2021: ch. 8). 8 See below, p. 15.

477

John Hyman 9 Husak (2007: 2453) appears to adopt a similar position. 10 But note that if the request is accompanied by a threat of violence, then they would not normally be said to have done the act voluntarily. See Hyman (2015: ch. 4). 11 Denno adopts Weinreb’s principle that “a person does not engage in conduct voluntarily if the conduct is not subject to [that person’s] control” (Denno 2002: 358). But she is evidently thinking about impulse control as opposed to the control of bodily movements when she adds: “However, there must be some limits on the reach of the meaning of the word ‘control’ so that it cannot include such out of control behavior as addiction or dangerous sexual offenses” (359). 12 Compare Moore: “It is the beginning of wisdom here to see that the unconnectedness of status [e.g., being a vagrant], omission, mental state, and involuntary bodily movements is not much of a reason for thinking that the opposite of each of these four things is not in reality one thing with a unified nature” (Moore 1993: 38). 13 See Hyman (2015: ch. 2). 14 Davidson’s famous example of the climber who loses his grip on a rope (Davidson 2001: 79) is a complex illustration of this point. 15 Lord Hailsham emphasizes the fact that a mistake concerning part of the actus reus is not properly regarded as a defense in Morgan ([1976] AC 182, 214), and the point is repeated in the Heilbron Report (Advisory Group on the Law of Rape (1975, Cmnd 6352, 4)). It is not affected by the subsequent change in the law concerning the mens rea element of rape. 16 Lord Widgery summarized the English legal position regarding duress as follows: “It is clearly established that duress provides a defense . . . if the will of the accused has been overborne by threats of death or serious personal injury so that the commission of the alleged offence was no longer the voluntary act of the accused” (Hudson and Taylor [1971] 2 All Er 244). 17 This theory of voluntary conduct raises the interesting question of why voluntariness, in addition to being a condition for culpability, should also be a condition for the validity of contracts, wills, consent to medical procedures, etc. Hart addresses this question in “Legal Responsibility and Excuses” (Hart 2008: ch. 2). 18 See Aquinas (2006: 1a2æ, q. 6, a. 5); Anscombe (1963: 89–90). The empiricist literature has always ignored the voluntary reception of an act, as opposed to its voluntary performance, because it is incompatible with the volition–motion model of voluntary conduct. 19 Austin (1885: vol. 1, 415); Holmes (2009: 83, 51); Williams (1983: 148); Moore (1993: 39; 1997: 251). See also Dias (1985: 309–311); Fletcher (1978: 421). Denno’s text contains the same mistake: “The enigma of voluntariness is captured in Wittgenstein’s famous question: ‘What is left over if I subtract the fact that my arm goes up from the fact that I raise my arm?’ . . . What is ‘left over’, then, must be voluntariness” (Denno 2002: 276n26). But it should be obvious by now that this is muddled. First, what is left over must be a fact since Wittgenstein is talking here about an arithmetic of facts. Second, what is left over is not the fact that the movement of my arm is voluntary, it is the fact that I made it happen, the fact that I caused the motion of my arm.

Further reading Duff, R. A. (2007). Answering for Crime: Responsibility and Liability in the Criminal Law. Oxford: Hart. (An important recent study of criminal responsibility, which includes a nuanced examination of the voluntary act requirement.) Hart, H. L. A. (2008). “Acts of Will and Responsibility,” in Punishment and Responsibility: Essays in the Philosophy of Law (2nd ed.). Oxford: Oxford University Press. (A classic sceptical assessment of the voluntary act requirement, influenced by Wittgenstein and Ryle.) Husak, D. (2007). “Rethinking the Act Requirement,” Cardozo Law Review, 28, 2437–2459. (An influential defence of the idea, originally mooted by Hart, that a control requirement is preferable to an act requirement.) Hyman, J. (2015). Action, Knowledge, and Will. Oxford: Oxford University Press. (A treatment of the main problems in the philosophy of action, including a detailed study of voluntariness.) Moore, M. S. (1993). Act and Crime: The Philosophy of Action and Its Implications for Criminal Law. Oxford: Oxford University Press. (The most influential defence of the voluntary act requirement since Hart’s sceptical assessment of it.)

478

The voluntary act requirement

References Advisory Group on the Law of Rape. (1975). Report of the Advisory Group on the Law of Rape. London: HMSO. American Law Institute. (1985). Model Penal Code: Official Draft and Explanatory Notes: Complete Text of Model Penal Code as Adopted at the 1962 Annual Meeting of the American Law Institute at Washington, D.C., May 24, 1962. Philadelphia, PA: The Institute. Anscombe, G. E. M. (1963). Intention. Cambridge, MA: Harvard University Press. Aquinas, T. (2006). Summa Theologiæ. Volume 17, Psychology of Human Acts (1a2æ. 6–17), trans. with intr., notes, appendices and glossary by Thomas Gilby. Cambridge: Cambridge University Press. Aristotle. (2009). The Nicomachean Ethics, trans. David Ross. Revised with an intr. and notes by Lesley Brown. Oxford: Oxford University Press. Austin, J. (1885). Lectures on Jurisprudence, or The Philosophy of Positive Law. 2 vols (5th ed.). London: John Murray. Bentham, J. (1996). An Introduction to the Principles of Morals and Legislation, ed J. H. Burns & H. L. A. Hart. Oxford: Clarendon. Davidson, D. (2001). Essays on Actions and Events (2nd ed.). Oxford: Clarendon. Denno, D. W. (2002). “Crime and Consciousness: Science and Involuntary Acts,” Minnesota Law Review, 87, 269–400. Dias, R. W. M. (1985). Jurisprudence (5th ed.). London: Butterworths. Duff, R. A. (2007). Answering for Crime: Responsibility and Liability in the Criminal Law. Oxford: Hart. Fletcher, G. P. (1978). Rethinking Criminal Law. Reprint. Oxford: Oxford University Press, 2000. Hart, H. L. A. (2008). Punishment and Responsibility: Essays in the Philosophy of Law (2nd ed.). Oxford: Oxford University Press. Hogg, C. (2022). “Automatism: Law and Theory,” unpublished MS. Holmes, O. W. (2009). The Common Law. Cambridge, MA: Belknap. Husak, D. (2007). “Rethinking the Act Requirement,” Cardozo Law Review, 28, 2437–2459. Hyman, J. (2015). Action, Knowledge, and Will. Oxford: Oxford University Press. Law Commission. (2013). “Criminal Liability: Insanity and Automatism. A  Discussion Paper,” accessed April 24, 2023, from http://lawcommission.justice.gov.uk/areas/insanity.htm. Libet, B. (1999). “Do We Have Free Will?” Journal of Consciousness Studies, 6, 47–57. Locke, J. (1997). An Essay Concerning Human Understanding, ed Roger Woolhouse. London: Penguin UK. Moore, M. S. (1993). Act and Crime: The Philosophy of Action and Its Implications for Criminal Law. Oxford: Oxford University Press. ———. (1997). Placing Blame: A Theory of the Criminal Law. Oxford: Oxford University Press. Pardo, M. S., & Patterson, D. (2015). Minds, Brains, and Law: The Conceptual Foundations of Law and Neuroscience. Oxford: Oxford University Press. Reid, T. (2010). Essays on the Active Powers of Man. Edinburgh: Edinburgh University Press. Ryle, G. (1949). The Concept of Mind. Reprint. London: Penguin, 2000. Simester, A. P. (2021). Fundamentals of Criminal Law: Responsibility, Culpability, and Wrongdoing. Oxford: Oxford University Press. Wegner, D. M. (2018). The Illusion of Conscious Will (New ed.). Cambridge, MA: MIT Press. Weinreb, L. L. (1970). “Comment on Basis of Criminal Liability; Culpability; Causation: Chapter 3; Section 610,” in National Commission on Reform of Federal Criminal Laws (ed.), Working Papers of the National Commission on Reform of Federal Criminal Laws. Volume 1, Relating to Chapters 1–13 of the Study Draft of a new Federal Criminal Code (pp. 105–151). Washington, DC: U.S. Government Printing Office. Williams, G. (1983). Textbook of Criminal Law (2nd ed.). London: Stevens. Wittgenstein, L. (1958). Philosophical Investigations (2nd ed.), trans G. E. M. Anscombe. Oxford: Blackwell.

479

38 STRICT LIABILITY AND STRICT RESPONSIBILITY R A Duff

1  Strict liability and strict responsibility “Actus non facit reum,” the slogan runs, “nisi mens sit rea”: an act does not make the actor guilty unless the mind is also guilty. The act is conduct that satisfies the definition of a crime, but the slogan declares that that definition should include not only conduct that has a specified impact on the world (the “actus reus”) but a specified agential relationship to every element of that actus – a relationship of intention, knowledge, recklessness (typically defined in English and American criminal law as consciously taking an unreasonable risk), or perhaps negligence; that agential relationship is orthodoxly called “mens rea.” Further, given the presumption of innocence, the prosecution must prove all the elements of the offense: both “actus reus” and “mens rea.”1 Thus the offense of criminal damage (Criminal Damage Act 1971, s. 1) involves “destroy[ing] or damag[ing] any property belonging to another intending to destroy or damage any such property or being reckless as to whether any such property would be destroyed or damaged”: to secure a conviction, the prosecution must therefore prove not only that the defendant (D) acted in a way that damaged another’s property (the actus reus), but also that he intended to cause, or was reckless as to the risk of causing, such damage, and realized or was reckless as to the risk that what he damaged was another’s property (the mens rea). Given such proof, D is then liable to conviction for criminal damage – unless he can offer a defense: he can still avert conviction by offering a legally recognized defense, for instance that he was acting under duress, or out of necessity to avert a greater evil. We need not discuss such general defenses here, but must remember that proof of the commission of an offense entails only presumptive liability  – liability unless D offers a successful defense. Sometimes (increasingly often, critics say) the law’s offense definitions do not fit this orthodox paradigm: they do not require the prosecution to prove both actus reus and mens rea as to every element of that actus, nor do they allow defendants to avoid conviction by offering evidence of lack of mens rea. Thus under s. 14 of the Policing and Crime Act 2009, a person commits an offense if he agrees to pay for sexual services from a prostitute who has been the victim of “exploitative conduct”; the prosecution must prove the intent to purchase sexual services, but it is “irrelevant . . . whether [the buyer] is, or ought to be, aware” DOI: 10.4324/9781003282242-51 480

Strict liability and strict responsibility

that the prostitute has been subjected to such exploitation. Once the prosecution proves the intentional purchase of sexual services, and the fact of subjection to exploitation, D is liable: it will not avail him to offer evidence that he had no reason to believe that the person he paid had been subjected to exploitation; criminal liability is “strict” as to that aspect of the offense. So too, a pharmacist who supplies drugs on a forged prescription commits an offense, even if she had no reason to believe it was forged: the prosecution must prove that she intentionally filled the prescription, but only the fact of its forgery; nor will it avail her to offer evidence that she had no reason to suspect that it was forged; criminal liability is “strict” as to that element of the offense.2 One type of strict liability is “constructive liability.” Criminal liability for an offense X is constructive when it is “constructed” out of liability for a lesser, related, offense Y. Thus the mens rea of murder in English law is an intention to cause death or “grievous bodily harm”: someone who intends to cause grievous bodily harm, the mens rea required for the offense of causing grievous bodily harm with intent (Offenses Against the Person Act 1861, s. 18), is guilty of the more serious offense of murder if his attack causes death; the prosecution need not prove that he intended to cause, or realized that he would or might cause, death; nor will it avail him to deny that he acted with any such intent or realization; his liability for murder is constructed out of his liability for the lesser offense (see Ormerod & Laird 2021: 532–535). Sometimes, the law uses a strategy that falls between the orthodox paradigm of nonstrict liability and strict liability as defined previously: it initially defines the offense in strict terms that do not require explicit proof of mens rea as to an essential element of the offense; but it then allows defendants to avert conviction by offering a “defense” of lack of fault as to that element. Thus under s. 57 of the Terrorism Act 2000, a person commits the offense of “[p]ossession for terrorist purposes” if “he possesses an article in circumstances which give rise to a reasonable suspicion that his possession is for a purpose connected with” terrorism; but it “is a defence” for him “to prove that his possession of the article was not for [such] a purpose.” The offense for which a convicted defendant is held liable is not merely possessing articles that might be useful to terrorists (a separate offense under s. 58 of the Act), but possessing those articles “for terrorist purposes”; but the prosecution need not prove such a purpose – whereas under the orthodox paradigm of non-strict liability it should have to prove purpose as well as possession. However, liability for this offense is not absolutely strict, since although the prosecution need not prove the relevant fault, D can avoid conviction by “disproving” it (and by s. 118, “proof” that he had no terrorist purpose need only involve “adduc[ing] evidence which is sufficient to raise an issue with respect to the matter”). There is still something “strict” about such offenses: the prosecution need not explicitly prove mens rea as to every element of the offense. However, liability is not strict, since D can still avoid conviction by adducing evidence of lack of fault as to the relevant elements of the offense. We can instead talk of strict responsibility. I am criminally liable if I am liable to conviction and punishment for an offense. I am, by contrast, criminally responsible if I can be called to answer for the commission of an offense, and am liable to conviction if I cannot offer an exculpatory answer. This distinction is evident in the orthodox distinction between offenses and defenses. The prosecution must prove every element of the offense, including (in the orthodox paradigm) both actus reus and mens rea, but that might not suffice to prove D’s liability: for D might now offer a defense that admits the commission of the offense, but adduces further factors that show 481

R A Duff

its commission to be justified or excused. I admit that I intentionally damaged another’s property, thus committing the offense of criminal damage, but then offer evidence that I acted under exculpatory duress (under a threat of immediate and serious harm), or out of justificatory necessity (this was the only way to avert a greater evil). I admit criminal responsibility for the commission of the offense, and am therefore held responsible, answerable, for that commission. The evidential burden now shifts to me: I can avoid liability only by offering evidence of a defense, which it would then be for the prosecution to disprove. Something similar happens in our extra-legal moral lives. I might admit that I am (or be held) responsible for a presumptively wrongful action – intentionally breaking your window, for instance: but I am not yet liable to moral blame for so doing, since I might offer a defense, a justification or an excuse, that blocks the inference from responsibility to liability; I broke it, I claim, only because someone plausibly threatened to maim me if I refused, or because that was the only way to get urgent access to someone having a heart attack in the room (see Duff 2007: 19–23). Criminal liability presupposes criminal responsibility: I  can be criminally liable for Φ only if I am criminally responsible for Φ. For criminal liability is liability to conviction and punishment, which censure the person convicted as a wrongdoer; but I  can be censured only for that for which I am held responsible. But criminal responsibility does not entail criminal liability: I can be held criminally responsible or admit criminal responsibility for Φ, but avoid criminal liability by offering a defence. Given this distinction between criminal liability and criminal responsibility, we can also distinguish strict criminal liability from strict criminal responsibility. Under the orthodox paradigm, neither liability nor responsibility is strict. If I am charged with criminal damage, I  must answer to that charge: I  am expected to make a formal answer by pleading “Guilty” or “Not Guilty”; if I plead “Not Guilty,” I might then have to offer a substantive answer to the prosecution case, to show that it does not prove my commission of the offense. But I  do not have to answer for that commission unless and until the prosecution proves, or I admit, that I damaged another’s property intentionally or recklessly; until then there is nothing for which I must answer in criminal court on pain of conviction if I do not provide an exculpatory answer. By contrast, in offenses of strict liability, both liability and responsibility are strict. The pharmacist who fills a forged prescription is strictly criminally responsible for doing so: she must answer for doing so, in the criminal court, even if the prosecution has not proved mens rea as to the prescription’s forgery. But she is also strictly liable for doing so: although she could offer some legally recognized general defense, such as duress or necessity, she cannot avert liability by offering evidence that she had no reason to suspect that the prescription was forged (though this might mitigate her sentence). Suppose, however, that the law was amended, so that it would be a defense for someone charged with this offense to “prove” that she did not know, and had no reason to suspect, that the prescription was forged; that she exercised “due diligence” as to the authenticity of prescriptions she was asked to fill (see Simester et al. 2022: 219–22l; Law Commission 2010: Part 6). Criminal responsibility would still be strict, since the prosecution would not initially bear the burden of proving mens rea as to the prescription being forged – the pharmacist would have to answer for filling a forged prescription even absent proof that she had reason to suspect that it was forged. But criminal liability would not now be strict, since she could avert it by offering an exculpatory answer for filling the forged prescription, which showed that she was not at fault in filling it. 482

Strict liability and strict responsibility

Now we have seen what strict liability and strict responsibility amount to, we must turn to the normative question of whether they can ever be justified; we can also look at our moral practices of attributing moral responsibility and liability (liability to moral blame), to see how far they are analogous to such practices in criminal law.3

2  Justifying (formally) strict criminal liability? Strict criminal liability seems on its face manifestly unjust. It permits the conviction and punishment of someone who acted without proved intention, knowledge, recklessness or even negligence as to an essential element of the offense with which he is charged, who thus acted without proved culpability, and without a fair opportunity to conform his conduct to the law’s demands: given the condemnatory nature of criminal law (it censures those whom it convicts) and the burdens that conviction and punishment impose (see Hart 2008), this seems unjust. A conscientious chemist deceived by a cleverly forged prescription could have avoided committing the offense: she could have insisted on telephoning the doctor who supposedly issued the prescription whenever she was asked to fill one; she could have refrained from becoming a pharmacist. But those are not measures that we could reasonably expect her to take, or blame her for not taking: they would be unreasonably burdensome for her, and would undermine the social benefits of having pharmacists who will dispense drugs to those who need them. To say that strict criminal liability is unjust is not to say that it cannot be justified: if the benefits of making liability strict, or the harms this would avert, are great enough, we might justify the admitted injustice that it involves. This is a familiar justification for strict liability offenses (see Simester et al. 2022: 214–219): given the difficulty of proving mens rea, it would be too hard to secure convictions; the law would be ineffective as a deterrent against harmful kinds of conduct; to secure adequate deterrence, we must make liability strict. One answer to this argument (s. 3) is that we might make the law adequately effective by imposing strict responsibility rather than strict liability. The argument against strict criminal liability sketched above is persuasive, but only when liability is substantively strict – only when it sanctions the conviction of those who lack a fair opportunity to avoid committing the offense, or whose conduct is blameless. However, criminal liability that is formally strict, in that it permits conviction without explicit proof of legally recognized mens rea as to some element of the offense, might not be substantively strict: it might still permit the conviction only of those who had a fair opportunity to avoid committing the offense, and who can justly be censured for committing it. For instance, constructive liability in murder (see s. 1) is formally strict: the prosecution need not prove that D intended, or foresaw a real risk of, death. However, we could argue that one who intends to cause grievous bodily harm is reckless of his victim’s life or death, and is properly convicted of murder if the victim dies. For either he realized that his attack would endanger the victim’s life or he did not. If he did realize it, he was certainly reckless, as consciously taking an unreasonable risk. If he did not realize it, we must ask how he could have failed to attend to something so closely connected to his intended action. The only possible answer is that he did not care about his victim’s life: he acted with the kind of practical indifference as to whether the victim lived or died that properly counts as reckless. That is why the formally strict liability involved here is not substantively strict: what must be proved (an intended serious attack) also proves the necessary fault as to death.4 By contrast, the formally strict liability imposed on a conscientious chemist who unknowingly 483

R A Duff

fills a forged prescription is also substantively strict (and thus unjust), since proof that she filled the prescription does not prove any kind of fault as to its being forged. This argument applies a limited version of the doctrine “Versanti in re illicita imputantur omnia, quae sequuntur ex delicto”: if someone engages in an “illicit” activity, we impute all its consequences to him. This is a version of the doctrine, because it takes engagement in this “illicit,” i.e., criminal, conduct to entail liability for some of its consequences; it is a limited version, because the “illicit” conduct is closely connected to the fatal effect for which the agent is then held liable. Can we extend formally strict liability wider than this, to other cases in which engagement in some “illicit” activity (with whatever mens rea that requires) suffices to prove fault in relation to a further, perhaps unforeseen, consequence of the activity? First, this argument does not justify imposing formally strict liability as to every criminal result of any criminal activity: there must be a suitable relationship between the res illicita (the criminal activity) and the criminal sequel. Contrast, for instance, the offenses of causing death by dangerous driving and of causing death by driving when one is uninsured. Under the Road Traffic Act 1988 s. 1, someone who is guilty of dangerous driving – driving in a way that “falls far below what would be expected of a competent and careful driver,” when “it would be obvious to a competent and careful driver that driving in that way would be dangerous” – is guilty of the more serious offense of causing death by dangerous driving if he in fact causes death. Now what makes driving “dangerous” is, according to s. 2 of the Act, that it creates a danger either of personal injury or of damage to property, and there is room for argument both about the closeness of the connection between “danger” as thus defined and death, and about the fault element of the basic offense; but we can see the basis here for a kind of constructive liability analogous to that imposed for murder: if I culpably drive dangerously, I culpably create a risk of serious harm; I can therefore properly be held responsible, and liable, for the harm that results if that risk is actualised. By contrast, a person commits the offense of causing death by driving while uninsured, under s. 3ZB of the Act, if he causes death by driving, at a time when he is uninsured: but there is no intrinsic connection between being uninsured and causing death; a driver does not endanger others’ lives by driving whilst uninsured. In this case, then, the res illicita is not illicit in a way that connects it to the further effect for which D is then held formally strictly liable; the commission of the base offense, even if intentional or knowing, does not display fault in relation to that further effect. In such a case, in the absence of an appropriate kind of link between the culpable res illicita and the further effect for which D is held criminally liable, that liability is substantively as well as formally strict, and is therefore unjust. Must the res illicita be criminal for formally strict liability to be just? Consider again the offense of purchasing sexual services from a prostitute who has been subject to “exploitative conduct” (Policing and Crime Act 2009, s. 14): liability as to that exploitation is formally strict, since it is “irrelevant . . . whether [the buyer] is, or ought to be, aware” of it; but is it also substantively strict; or could it be argued that paying for sexual services is a “res illicita” engagement in which entitles the law to hold the agent liable for (in effect) complicity in the exploitation, even though paying for sexual services is not in itself criminally illicit? The argument would need to be that it is nonetheless a kind of wrong that we could legitimately criminalize, since exploitation is, as everyone should know, a feature of such commercial sexual practice; and that although we might have good reason not to criminalize the activity itself, we can say that anyone who engages in it takes a risk that the person he pays is a victim of exploitation, and therefore a risk that he will be held liable if she is in fact such a victim. He is walking on morally very thin ice, and cannot complain 484

Strict liability and strict responsibility

if he falls in.5 A  similar argument might justify imposing formally strict liability as to a person’s age in various sexual offenses: if a person engages in sexual activity with someone who is in fact under a specified age, he is guilty of an offense even if he does not know and has no reason to suspect that the person is under the specified age (see, e.g., Sexual Offences Act 2003, ss 5–11). To make the argument stronger, we could require that the defendant is considerably older than the alleged victim (V) (see, e.g., Washington Criminal Code 9A, 44.073, 076, 079), and suggest that an older person who engages in sexual activity with a much younger person is engaged in an activity that is, even if not criminal, at least morally suspect, because likely to create a serious risk of sexual exploitation: one who engages in such activity thus takes (and should realize that he is taking) a serious moral risk that he is exploiting V; he cannot complain if he is held criminally liable when the risk is actualized. However, this line of argument is problematic in this context: it is hard to specify the “res illicita” with any precision; apparent age can often be very deceptive; and the risk that V is under the specified age is not the same as the risk of sexual exploitation, since age is a very imperfect proxy for vulnerability. It is therefore probably unjust to make criminal liability for offenses of “underage sex” strict as to V’s age. We might instead so define the offense that the prosecution must prove not just that V was in fact underage, but also that D did not reasonably believe that V was over the specified age (compare Sexual Offences Act 2003, ss 9–12), thus grounding liability in negligence. Or we could instead make responsibility, rather than liability, strict: the law could define the offense initially in strict terms, as it does now; but provide for a defense of “reasonable belief” that V was above the legally specified age (compare Sexual Offences Act 2003, ss 16–19). Similarly, it is unjust to hold the conscientious chemist who is deceived by a cleverly forged prescription criminally liable, and if the prosecution had to prove negligence (that she should have realized it might be forged), that might make it too difficult to convict genuinely careless or corrupt pharmacists. A solution would be to make responsibility rather than liability strict: to define the offense initially in strict terms, but to allow a defense to a pharmacist who could offer evidence that she had exercised “due diligence.” This suggests that formally strict criminal liability will rarely be just. It is just when it can be argued that although the prosecution is not required explicitly to prove mens rea or fault as to some element of the offense, proof of what the prosecution is required to prove amounts in effect to proof of an appropriate kind of fault as to that element: but this condition is likely to be satisfied only in a limited range of “constructive” liability offenses; and perhaps in cases, like that under s. 14 of the Policing and Crime Act 2009, in which the prosecution must prove conduct that could justifiably be criminalized, and that involves an appropriate kind of fault (compare Michaels 1999). It also suggests, however, that once we distinguish strict liability from strict responsibility, we might see more reason to impose strict responsibility: if one aim is to make it less difficult to obtain convictions, and thus to enforce the law more effectively, we might be able to achieve this without the injustice involved in strict liability, by making responsibility rather than liability strict. It is that possibility, and the question whether strict criminal responsibility can be just, that we must now consider.

3  Justifying strict criminal responsibility? On its face, strict criminal responsibility seems at odds with the presumption of innocence (that “golden thread” that runs “[t]hroughout the web of the English criminal law”; 485

R A Duff

Woolmington v DPP [1935] AC 462, at 481): for whereas on the orthodox paradigm the prosecutor must prove both actus reus and mens rea, conduct and fault, when responsibility is strict she need not prove mens rea as to some aspect of the offense – the burden instead falls on the defendant to disprove it, or at least to offer evidence of its absence. But this means that the prosecutor need not prove an essential dimension of guilt; the defendant must disprove it – which seems inconsistent with the principle that it must be for the prosecutor to prove the defendant’s guilt (see Ashworth & Blake 1996). A hint as to how we might nonetheless show some kinds of strict criminal responsibility to be just is provided by the way that defenses figure in a criminal trial. A defendant is guilty of the crime charged only if he both committed the offense as defined by the law, and had no legally recognized defense for doing so, but the prosecution need not initially prove the absence of any defense  – prove that D was not legally insane, for instance, or acting under exculpatory duress or out of justificatory necessity: it is for D to offer a defense, and evidence to support it; only then does the burden shift back to the prosecution to disprove it (see Duff 2007: ch. 9). This is thought to be consistent with the presumption of innocence (at least if D does not have to prove a defense, but need only offer evidence that suffices to create a reasonable doubt), even though the prosecutor is not required to prove D’s guilt from the start; why is that? A first answer is that the court is entitled to presume the absence of such abnormal factors as insanity, duress and necessity, and thus to presume D’s guilt given proof of his commission of the offense; that answer might be supported by saying that proof of the commission of the offense is proof of a wrong for which the perpetrator can be called to answer in a criminal court on pain of being convicted if he does not offer an exculpatory answer. If, for instance, I intentionally wound you, or destroy your property, it is reasonable to call me to answer in a criminal court for such conduct – to call me either to admit my guilt, or to offer an explanation that will exculpate me. It might now be said that moral responsibility is typically strict: why should we not also make criminal responsibility always strict? If I damage your property, you can reasonably hold me responsible, answerable, for it: I should be ready to answer to you for the harm that I caused, even when my answer will exculpate me – that I caused it through non-culpable inadvertence or accident. So why should I not be held criminally responsible, in a criminal court, if I am proved to have committed the actus reus of a criminal offense? I should not of course be held liable if I did not act with the relevant kind of fault, nor should I have to prove that I was not at fault; but why should I not at least have to offer evidence that I was not? In that case, in a criminal trial the prosecution would initially have to prove only the actus reus; the evidential burden would then shift to D to offer evidence of lack of mens rea. The answer to this question appeals to the difference between the context of interpersonal moral relationships and that of a criminal court. If criminal responsibility is strict, one who does not offer a successful defense of lack of fault faces conviction of and punishment for the offense charged, whereas I face no such consequences in the moral context. You can criticize me if I refuse to answer to you, but you are not then entitled to presume my moral guilt – to presume that I caused the harm intentionally, recklessly or negligently: you are entitled, at worst, to suspect that I was morally at fault; even if you decide that I was at fault, the consequences are typically much less onerous than those of a criminal conviction. Further, the burden of going to trial (apart from the risk of being mistakenly convicted) is very much greater, materially and psychologically, than that of answering for my conduct in an informal moral setting. 486

Strict liability and strict responsibility

That is why it seems reasonable to demand that the prosecution prove not merely that I committed the actus reus of a crime, that I caused some criminal harm, but also that I did so intentionally, recklessly, or at least negligently, before I must answer for causing that harm in a criminal court, on pain of conviction if I  fail to offer an exculpatory answer. On the other hand, it does not seem unreasonable to impose strict criminal responsibility on the pharmacist who fills a forged prescription, or supplies the wrong drug by mistake: that is, to define the offense initially in strict terms, so that it is committed by a pharmacist who did not realize that the prescription might have been forged; but to allow a defense of reasonable ignorance, or due diligence, which lays on D the onus of adducing evidence of such a defense.6 Why might it be reasonable to impose strict criminal responsibility on the pharmacist? A few features of her situation are relevant here. First, she chose to engage in this activity. This feature is not that significant in her case, since the activity is socially valuable: she could have avoided the risk of being prosecuted by not becoming a pharmacist, but we should not discourage competent and careful people from taking up this profession. The fact of choice is rather more significant in activities that lack such social value: for instance, if someone chooses to drive when decent public transport is readily available, or chooses to sell recreational drugs on the black market, it is relevant to note that they could have avoided the risk of causing the kind of harm inherent to the activity, and the risk of criminal prosecution, by not engaging in it. Second, the activity in which she chooses to engage creates distinctive risks of significant harm – the risk, for instance, of providing the wrong drug, or of facilitating dangerous drug use or trade by filling a forged prescription. The law can therefore reasonably impose on her a stringent prospective responsibility to guard against such risks: to take precautions, exercise “due diligence,” to make sure that they are not actualized. If one of those risks is actualized, this raises a question about whether she properly fulfilled that prospective responsibility; it is reasonable to call her to answer that question in court, by showing that she did fulfil it (on prospective as against retrospective responsibility (or answerability), and the ways in which retrospective responsibility depends on prospective responsibility, see Cane 2002: ch. 2; Duff 2007: ch. 1). Third, it is possible to specify, to some degree, what taking due care requires. This could be done by a (legal or professional) code of good practice: pharmacists can then know what they must do to discharge their responsibility. Analogously, drivers have a responsibility to ensure that their cars are in a safe condition; in requiring owners to submit their vehicles to regular MOT tests, the law defines part of the content of “due care” in this context. Fourth, and accordingly, it will be relatively easy for a conscientious pharmacist to offer evidence that she had exercised due diligence: she can report the procedures she put in place, and will have kept records of them. Analogously, a driver can report that and when she put her car to a MOT test, and offer a MOT certificate as formal evidence of this. This kind of justification of strict criminal responsibility is most plausible in relation to activities that are already regulated, in which regulations and guides make clear what kinds of care those engaged in the activity must take, and how they can show that they have taken such care. But here is a different kind of example that is worth considering. The English law of rape used to require the prosecution to prove that D knew that V was not consenting, or was reckless as to whether V consented: this led to the notorious decision (based on the orthodox definition of recklessness as conscious risk-taking) that a man who believed, however unreasonably, that the actually non-consenting victim consented was 487

R A Duff

not guilty of rape (DPP v Morgan [1976] AC 182). In 2003 the law was changed: under s. 1(1) of the Sexual Offenses Act 2003, all that is required by way of fault or mens rea (apart from the intention to sexually penetrate) is that D did “not reasonably believe that” V consented. The absence of such reasonable belief is, under the current law, something that the prosecution must prove; but suppose we suggested that criminal responsibility for this offense should be strict? The prosecution would still have to prove intentional sexual penetration and actual lack of consent, but D would then be liable to conviction unless he offered a defense of “reasonable mistake.”7 How might such a suggestion be supported? It would not be enough to argue that this would make it easier to secure convictions: it would also need to be argued that given the risk of non-consent and the harms involved in suffering non-consensual sexual penetration, anyone intending such penetration incurs a stringent prospective responsibility to ensure that it is consensual; and that it would be easy for a conscientious person to take steps to establish consent, and to offer evidence that he had done so. We cannot explore this issue further here: the point is simply to note the structure of the argument that would be needed to show strict criminal responsibility to be just. We would need to argue that, given the nature and value of the activity, and its inherent risks, it would be reasonable (and not unreasonably burdensome) for the law to hold that those who choose to engage in it incur a stringent responsibility to take steps to guard against such risks: a responsibility so stringent that they will be held criminally responsible if those risks are actualized, and criminally liable if they cannot offer evidence that they exercised due diligence. We might then ask whether this line of argument could justify imposing strict criminal responsibility on those who pay for sexual services, in relation to the risk that those whom they pay have been subjected to exploitation.8 Finally, to bring the discussion back to strict liability, we can now suggest one potential indicator of cases in which it might be appropriate to impose not just strict responsibility, but formally strict liability. Imposing strict responsibility leaves it open for D to offer a defense of due diligence: I engaged in this admittedly dangerous activity, but took due care to guard against its risks. Suppose now that someone who killed a person by attacking them in a way intended to cause serious bodily injury argues that he took due care, exercised due diligence, to avoid their death – perhaps he took care in the nature of his attack to try to ensure that it seriously injured without killing; he intended to beat V to within a foot, rather than an inch, of his life. The obvious answer to his claim is that whatever steps or precautions he took did not amount to due care, or due diligence: the only care or diligence that would have counted as “due” would have been not to attack V in the first place. If there is no room for a defense of due diligence, to block the inference from responsibility to liability, we may impose formally strict liability, in the form of constructive liability. We can then ask whether someone who pays for sexual services should be heard to argue, as a defense, that he took due care to check that the person from whom he purchased them had not been exploited; or should we rather say that the only kind of care or diligence that would be “due” would be to refrain from purchasing such services in the first place?

Notes 1 This is a radically simplified sketch of the structure of Anglo-American criminal liability; the qualifications necessary for a complete account are not important here. 2 Medicines Act 1968, s. 58(2); Pharmaceutical Society of Great Britain v Storkwain [1986] 2 All ER 635. 3 We cannot here look at strict liability and responsibility in other areas of law, such as tort law.

488

Strict liability and strict responsibility 4 See further Duff 2007: 252–260. Perhaps this justification is plausible only if “grievous bodily harm” must involve an obvious risk of death (see Law Commission 2006: paras 2.88–2.94); that would still involve formally strict liability, if the prosecution need not prove that D noticed that obvious risk, and it would not avail him to offer evidence that he did not notice it. 5 For the classic formulation of the “thin ice” principle, see Knuller [1973] AC 435, at 463 (Lord Morris). 6 For a partial analogy, compare the defences recognised by ss 67A-D of the Medicines Act 1968 – though they concern ignorance rather than due diligence. 7 Or the law could create an “evidential presumption” (compare s. 75 of the Act): given proof of nonconsensual sexual penetration, D “is to be taken not to have reasonably believed that [V] consented unless sufficient evidence is adduced to raise an issue as to whether he reasonably believed it.” 8 We could adopt a relative of the “versanti in re illicita” slogan (see text following n. 4): “Versanti in re periculosa imputantur omnia, quae sequuntur ex actio”; to one who embarks on a dangerous activity will be imputed all that follows from it – but what is imputed is responsibility, not liability.

Further reading Cane, P. (2002). Responsibility in Law and Morality. Oxford: Hart Publishing (a deep and complex discussion of varieties of responsibility in different areas of law, and of the social and relational character of responsibility). Gardner, J. (2007). Offences and Defences. Oxford: Oxford University Press (a collection of essays on criminal and moral responsibility by one of the subtlest and most insightful contemporary theorists). Hart, H. L. A. (2008). Punishment and Responsibility (2nd ed.). Oxford: Oxford University Press (a classic collection, including several papers on the nature and grounds of criminal responsibility). Honoré, T. (1999). Responsibility and Fault. Oxford: Hart Publishing (a distinctive account of “outcome responsibility” as a type of strict responsibility, and of the role of such responsibility in morality and in law). Simester, A. P. (ed.). (2005). Appraising Strict Liability. Oxford: Oxford University Press (a collection of essays by leading criminal law theorists, analyzing different types of strict liability and the possibility of justifying them).

References Ashworth, A. J.,  & Blake, M. (1996). “The Presumption of Innocence in English Criminal Law,” Criminal Law Review, 306–317. Cane, P. (2002). Responsibility in Law and Morality. Oxford: Hart Publishing. Duff, R. A. (2007). Answering for Crime: Responsibility and Liability in the Criminal Law. Oxford: Hart Publishing. Hart, H. L. A. (2008). “Legal Responsibility and Excuses,” in Hart, Punishment and Responsibility (2nd ed., pp. 28–53). Oxford: Oxford University Press. Law Commission. (2006). Murder, Manslaughter and Infanticide (No. 304). London: HMSO. ———. (2010). Criminal Liability in Regulatory Contexts (Consultation Paper No 195). London: Law Commission. Michaels, A. (1999). “Constitutional Innocence,” Harvard Law Review, 112, 828–902. Ormerod, D., & Laird, K. (2021). Smith and Hohgan’s Criminal Law (16th ed.). Oxford: Oxford University Press. Simester, A. P., Spencer, J. R., Stark, F., Sullivan, G. R., & Virgo, G. J. (2022). Simester and Sullivan’s Criminal Law: Theory and Doctrine (8th ed.). Oxford: Hart Publishing.

489

39 RESPONSIBILITY AND PRETRIAL DETENTION Kimberly Kessler Ferzan

The practice of holding someone responsible for criminal wrongdoing requires a guilty plea or verdict. When defendants do not immediately plead, the crucial inquiry often becomes whether the defendant should be released or detained, and if released, whether it should be on her own recognizance or with conditions attached, such as the relinquishment of a passport or the payment of a financial security. Making these decisions typically involves examining the defendant’s dangerousness, risk of flight, and risk of obstruction of justice. At first blush, these all look to simply be inquiries into the future conduct of the defendant. Any predictive enterprise raises a host of problems, including how good our predictive instruments are and whether our calculations rely on distributively unjust inputs, as pithily captured by Sandy Mayson’s title, Bias In, Bias Out (Mayson 2019; see also Starr 2014). There is a robust literature on these issues (Mayson 2018). However, viewing these reasons as largely predictive risk assessments obscures crucial questions about the role of the defendant’s wrongdoing in justifying the detention. Often it is the defendant’s responsibility for an action that goes significant distance in making the detention permissible. And, for detention inquiries, the focus is not even on one kind of action. Instead, justificatory heavy lifting may potentially be done by different claims of responsibility. The defendant’s detention may be predicated on his responsibility for the crime charged, his responsibility for posing a current threat, or even his responsibility for the creation of legitimate distrust. Ultimately, we are not looking at just one predictive question that might be fed into an algorithm; instead, we are looking at a complex interweaving of risks, rights, and responsibility. This chapter begins by explaining the different relationships that detention claims may have with respect to a defendant’s rights. From there, it disentangles dangerousness as an exemplar of distinct responsibility claims that potentially render predictive detentions permissible. After flight and obstruction of justice are considered, the chapter asks whether this disaggregation masks the possibility that different rationales may be combined to justify detention. A final path to detention – inability to pay bail – is also examined. Although the reader will likely conclude that much of our current detention practices are unjust (at least in the United States), this chapter aims to provide a clear normative framework for DOI: 10.4324/9781003282242-52 490

Responsibility and pretrial detention

evaluating detention rationales and the kinds of arguments about the defendant’s responsibility that are necessary to render detentions permissible.

Rights: duties, forfeitures, lesser-evils justifications, and eliminative actions To assess the complexity of detention decisions, let us begin with rights. The reason the state wants to detain the defendant is because it fears she may do something in the future: harm another, obstruct the investigation, or abscond. However, when the state offers its reason for detention, these reasons are not sufficient to justify detention. This is because the defendant has a right against being harmed, including having her liberty restricted. If someone were to take you and put you in a cage, that would be kidnapping. Something more is needed. For example, what transforms kidnapping to criminal punishment is that (1) the defendant has in some sense “forfeited” the right against punishment and (2) then, the state’s good reasons to impose punishment – be they desert, deterrence, incapacitation, or rehabilitation – can kick in. But looking to the state’s reasons is only half the story; the other half is the defendant’s rights. Hence, when we look to pretrial detention, here, too, we should ask exactly what may be at work with respect to the defendant’s rights when the state aims to detain her. In some cases, the defendant may have forfeited, waived, or otherwise lost her rights against the detention. We see these sorts of arguments in other instances in which a person or the state harms another. A defendant who has been found to be guilty lacks a right against punishment.1 A culpable aggressor, by virtue of her conduct, lacks a right against proportionate and necessary defensive force. Or, one may actually have a duty to perform the act; one cannot complain about being forced to pay taxes if one has a duty to do so. When the person lacks a right against a certain treatment, it is easier to the justify the action. The fact that a culpable aggressor lacks a right means that the defender does not balance her life against the aggressor’s; instead, the defender may simply point to her interest in living. So, too, if a defendant lacks a right against detention, then the state may more easily justify the liberty deprivation. These are duty, liability, or forfeiture accounts, where no right stands in the way. I will use “forfeiture” as the shorthand for this category, though nothing turns on my nomenclature. If one thinks we simply don’t have rights in these circumstances – that is, that our rights are specified in particular ways so that we don’t have right against deserved punishment, for example – the same analysis applies here. Ultimately, when there is not a right, the state’s reason to detain will often be sufficient to justify detention. When the defendant does have a right against the adverse treatment, then harming her is far more difficult to justify. Consider the Trolley Problem. There, the question is whether one may justify turning a runaway trolley from five people to one. Many theorists believe it is permissible, and perhaps required,2 to turn the trolley. This is an overriding account, where the strength of the reasons to turn the trolley are far more substantial than the reasons not to turn the trolley. Notably, the Trolley Problem allows harming when the benefit to harm ratio is five to one. At this point, it is thought that an innocent person may be harmed as a side effect of a net positive action. Despite their superficial similarity, detention cases are not trolley problems. The question is not how much harm one may impose on an innocent bystander to prevent injury to five others. Instead, the question is how much harm one may impose to eliminate a threat. It is as if you are turning the trolley to the one to prevent what he will otherwise do. To be clear, this is not a forfeiture case, but rather a case in which we predict a future 491

Kimberly Kessler Ferzan

act that would constitute a forfeiture. Accordingly, the question is whether, with respect to responsible agents, the state accords them the trust and respect to which they are entitled if the state aims to eliminate the “threat” the defendant poses based merely on a prediction of future wrongdoing. This sort of eliminative account is one that presupposes that the defendant has not forfeited a right but is still a threat (in an amorphous and inchoate sense of the word). Harming the defendant in order to minimize the threat he presents is the goal. Each of these accounts has different upshots. If the state puts forth a reason to detain the individual, and the individual has no right against the detention, we can see why the detention would be justified. (There could be reasons why such a detention would still not be justified all things considered, for example, because of harm to third parties. But we can put those contingencies to the side for our purposes.) In contrast, if the defendant’s rights are overridden for the greater good, then typically, we think that we need to compensate the defendant for the right’s infringement. This is seen in other areas of the law and morality, such as when we take someone’s property for public use or when a boat owner ties his boat to a dock to prevent significant economic losses (Vincent v. Lake Erie 1990). In each instance, payment is required for the infringement of the right. Finally, eliminative cases should strike one as problematic in that they involve predicting future conduct and a standpoint of distrust. There are thus two potential rights at issue: the right not to be harmed because there has not been any forfeiture and the right to a degree of civic trust between the state and its citizens. There is much imprecision in the judicial and scholarly discourse on detention. Courts detain individuals to stop amorphous future threats. Adding to the lack of rigor in this area, theorists have flattened the scholarly discourse into cost-benefit analysis instead of attending to the way in which different rationales have different impacts on rights (Baradaran Baughman 2017; Yang 2017). There is also little precision in terms of what the agent is potentially responsible for that would justify detaining her. To see the conflation of various issues, let us begin with dangerousness.

Disentangling dangerousness Consider a defendant who is charged with committing a violent felony, such a bank robbery. Assume that it will take two months for his case to come to trial. May the state permissibly detain the defendant until that trial because the defendant is “dangerous”? To say that a defendant is dangerous is to claim that there is a moment in the future when the defendant is likely to harm someone. Ordinarily, we would not think that we could simply take someone off the street, predict they are dangerous, and detain them. (For the purpose of our inquiry, we are excluding nonresponsible agents, such as individuals with severe mental illness.) One reason not to do this would be that we might not be very good at predicting who the future dangerous actors are, but the other worry would be that it could violate the person’s rights to detain them until they had done something (Ferzan 2011; Walen 2011a). In other words, what does it mean to say we get to detain someone because he is dangerous? Detaining individuals pretrial on dangerousness grounds masks three distinct potential pathways to permissibility premised on the defendant’s responsibility (Ferzan forthcoming). First, we might ask whether the defendant’s commission of the charged offense is such that he forfeits his right against detention. (To be sure, this seems to run afoul of the presumption of innocence, a problem to which we will return.) Second, we might consider whether 492

Responsibility and pretrial detention

the defendant has engaged in actions that would justify the state intervening along the lines of self-defense. Third, we might ask whether, even if the defendant has not done anything, we are entitled to act based on the prediction that they will engage in a future wrong. Notice that the first dangerousness account premises detention on a forfeiture claim – that the defendant is responsible for having committed a crime. The second approach is also forfeiture, but here, it is premised on the defendant’s responsibility for posing a threat. The third approach is eliminative, and it is not predicated on the defendant’s responsibility for an action that grounds the loss of a right.

Dangerousness and responsibility for the crime charged Let us consider the crime charged first. The way we typically incapacitate someone who is dangerous and has committed an offense is to find him guilty, and then, as part of his punishment, to incapacitate him. This approach has two aspects. First, the reason no right stands in the way is because the person committed the offense. Second, a potential reason to incarcerate the defendant is to stop him from committing future offenses.3 There is a builtin proportionality to this forfeiture. For both retributivists and side-constrained consequentialists, the former (the amount of the right loss) will set the upper limit for the latter (the extent to which it is permissible to then pursue the goal of incapacitation) (Ferzan 2018). In other words, for a serious offense, because the defendant deserves punishment or has otherwise forfeited her right against hard treatment, we are entitled to pursue goals including potentially the intrinsic goodness of giving someone what they deserve, the value of deterring others or the defendant, and the value of incapacitating the defendant to protect others from harm, but only up to the amount of the forfeiture of the right. In contrast, we balk at the idea that we would be entitled to incapacitate someone for life for a minor offense because the commission of a minor offense does not entail the forfeiture of a right against punishment for a significant period of time. The prior criminal act that the defendant committed determines the breadth of the state’s ability to act on its prediction of future harm. The emphasis of this discussion should strike you as odd. Why in the midst of a discussion of pretrial detention are we discussing punishment? Moreover, don’t we need proof beyond a reasonable doubt before we can invoke this kind of forfeiture? The puzzle, then, is how we could possibly be justified in using the defendant’s guilt for this offense as a justification to hold him for the trial on the very same offense. What seems unseemly here is that this approach runs afoul of our political morality. Although, in the United States at least, this is not actually a violation of the presumption of innocence, it does seem to undermine the value that undergirds it. If we owe the defendant a trial before we are permitted to say he forfeited his right, how is it that we could possibly get a head start on punishing him? To some theorists, the answer is that this is not possible (Baradaran 2011; Duff forthcoming; Mayson 2018). We are simply not permitted to access punishment as a reason until the defendant has actually been found guilty. But let us consider the strongest case for permissibility. One could argue that the defendant is not wronged if he truly committed the offense. From a fact-relative perspective (the omniscient, objective perspective), perhaps no right stands in the way.4 Moreover, given that this incarceration “counts” for purposes of the total sentence, he is not being punished twice. If the defendant has actually committed the offense, then detaining him prior to trial looks quite a bit like shooting someone who was about to shoot you but you didn’t know 493

Kimberly Kessler Ferzan

it, or harming someone without knowledge that they deserve it.5 For some theorists at least, even though the actor does not know the fact that justifies her action (that the action will be defensive or deserved), this is irrelevant so long as there is both a forfeiture (guilt) and a reason to impose the harm. That would mean that in the case of pretrial detention, the most the defendant can complain about is the violation of her right to a trial, not the right against detention. The defendant is detained, minimizing dangerousness, and this counts as punishment; hence, for so long as the defendant actually is guilty, the defendant’s rights are only violated with respect to their trial right, and that violation is not particularly stringent because the trial will be shortly forthcoming. Notably, whatever theoretical purchase the minimization of the rights violation might have, in the real world, this approach is highly problematic. Pretrial detention has a negative effect on trial outcomes, thus leading to innocent defendants being found guilty (Heaton et al. 2017). And, individuals who are detained but then found not guilty (or have their cases dismissed) are not compensated in any way, despite the significant negative impacts that detention may have had on their lives. Finally, sending the expressive signal and taking the practical position that the trial does not matter if the defendant is actually guilty undermines rule of law values. Accordingly, whatever the thin philosophical bona fides of this position, in practice, it is incredibly destructive of people’s lives.

Dangerousness and responsibility for a culpable threat In contrast to a dangerousness justification that looks to the crime charged, detention might instead be predicated on the defendant’s responsibility for a specific threat the defendant currently poses. To understand this sort of rationale, we should look to self and other defense. If A culpably threatens to harm B, B is entitled to use proportionate and necessary force. And, C is permitted to aid B in so doing. Hence, we might think that, if the defendant (A) is charged with attempting to kill his estranged wife (B), that the state (C) should step in and detain the defendant prior to trial. The crucial inquiry in these sorts of cases is when the defendant loses the right against being detained  – that is, the conditions for forfeiture (Ferzan 2011). One question that plagues self-defense theorists is whether an aggressor must be culpable or whether even innocent or nonresponsible theateners may lose rights. But, that puzzle is unlikely to be the relevant inquiry for pretrial detention. We aren’t predicting that someone is going to be thrown down a well or to mistake a hologram for a real gun, popular cases for philosophical hypotheticals. And, to the extent that we are worried about nonresponsible actors, those problems will be dealt with within the context of civil confinement measures. So, we are focusing on culpable aggressors. The question is at what point do they lose rights such that the state may intervene? Consider each pole. Do we have to wait until the harm is dangerously close to occurrence? For individual citizens, they must wait for the threat to be imminent before they act in self-defense. But many scholars believe the reason that citizens must wait is because the state should intervene earlier (Fletcher 1996). Given that we are talking about state intervention, surely the state is not required to wait. Moreover, even theorists who contend that in the criminal context someone has not committed an attempt until she is very close to completing the crime have not endorsed this sort of requirement for self-defense. The whole point is that the state ought to be able to stop people who have embarked on criminal paths before the threat is imminent. 494

Responsibility and pretrial detention

On the other end of the spectrum, may we intervene as early as possible? Without relying on science fiction-like hypotheticals, we know from our everyday experiences that we can sometimes predict that someone will come to form a culpable intention, even if she does not have one now. But this kind of reasoning is the sort that leads to prediction of future dangerousness, not the kind of choice that leads to forfeiture conditions now. Even if we feel the pull of anticipatory defense, this is distinct from claiming the defendant herself has done something such that she no longer has a right against defensive force. Larry Alexander calls these cases “anticipated culpable aggressors” or “ACAs” (Alexander 2016). Again, thinking an ACA might be a future threat and harming her to stop its fruition is an eliminative action, whereas, we are asking whether she has done something sufficient to forfeit the right. Though I do not intend to defend this view here, elsewhere I have argued that it is at the time of the aggressor’s formation of the culpable intention that she forfeits her rights against harm (Ferzan 2012). She can no longer complain that the defender is stopping her once she has chosen to harm him. Even if one disagrees with me that this is where the line should be drawn, we must still engage in an argument about the point at which it is fair to say the defendant has engaged in an action that forfeits her rights. Notice that at this point we have a distinct way of thinking about dangerousness. The earlier punishment view grounded the right to incapacitate on the commission of the charged offense. But here, the responsible action that grounds the right to incapacitate is akin to self-defense – there is a future harm that the defendant has embarked upon in such a way that she loses rights against the state stopping her.

Dangerousness and eliminative actions Let us now return to our accused bank robber. If our commitments to the presumption of innocence prevent us from detaining him prior to trial because of his (alleged) guilt for the crime charged, and if we lack sufficient evidence to believe that at this moment he has an intention to commit another crime, then our intuitive urge to detain him on dangerousness grounds is based on the anticipation that he might cause harm in the future. State action based on prediction is troubling. One concern is the kinds of factors that lead to predictions of future wrongdoing. To be sure, they may include actions and choices made by defendants, but they might also include mere statistical categories into which the defendant falls. That is, prediction does not require prior responsibility for any sort of choice. We would thus be eliminating a threat based on factors over which the defendant had no agential control, and it seems unfair to make predictions about people based upon what reference class they may happen to fall within (Bolinger 2021). But what would it mean for the state to act on a good prediction? May the state harm you now because of what it thinks you will do in the future? A cursory objection would be that this sort of detention by the state is autonomy denying. But that isn’t quite right. The state is not denying that you will autonomously choose whether to act. Instead, the state is distrusting you. It is not giving you the space to act rightly. Imagine that the state had people take a dangerousness test every time they bought milk. If you failed the test, you could be detained. If you passed, you’d be free for now. Your objection would be that no matter how good the state’s predictive powers, you have not given the state reason to distrust you. In other words, you have not acted in a way that forfeit rights. Hence, the argument against detention for dangerousness is that the defendant 495

Kimberly Kessler Ferzan

is not being treated as a free and equal citizen entitled to the same kind of civic trust as the rest of us (Duff forthcoming; Mayson 2018 (suggesting a similar hypothetical)). Now, perhaps there is a way to draw a distinction between the milk purchaser and the charged defendant. Let us return to forfeiture. To this point, I have suggested that the defendant might forfeit a right against punishment and that the defendant might forfeit a right against defensive force. But can a defendant forfeit a right to trust? If the evil of pretrial or preventive detention is that the state is distrusting its citizen, then that evil is abated if one can forfeit a right to trust. If one can so forfeit, the analysis may come closer to an overriding account. To some extent, this kind of forfeiture should be familiar to all of us. There’s the saying, “Fool me once . . .” We do think that when people lie and mislead us, we are suckers if we believe them the second time around. That is, whatever baseline trust we owe them initially, we certainly don’t owe it to them after they have abused that trust. And, indeed, part of our justification for bail and other requirements is that we may be entitled to withhold trust from someone for whom there is probable cause to believe she has committed a crime. What we need though is to fully examine the bounds of this sort of forfeiture (Ferzan forthcoming; Mayson 2018). It seems difficult to say that just because we have probable cause to believe someone has committed an offense that we are entitled to withhold trust to such an extent that we are entitled to put them in a cage as dangerous. But let’s examine the most compelling case for this distrust – someone with a series of prior offenses who is now being charged with committing the offense again. Say someone has committed five bank robberies, and for each robbery, she has been punished. Before we ask about pretrial detention for a charged offense, we should ask: Should we be entitled to continue to keep her locked up simply based on these priors? Has she forfeited her right to be trusted going forward? It is difficult to see what the precise contours are of this forfeiture if it is not part of the punishment (compare Walen 2011b). Now, perhaps we can distinguish the case of the person pending trial because there is probable cause (at least if not more) to believe the person has committed an offense. But how far can that get us? If we balk at just detaining a defendant who has committed five bank robberies, then why would we be more comfortable detaining someone who has committed four and is charged with her fifth? After all, in both instances, the justification for detaining the person is to prevent the sixth offense. A potential response from the state is to point to the limited duration of pretrial detention that will simply fill the gap until the trial (and then continued detention on incapacitation grounds). So, the idea is that although we do not robustly forfeit rights not to be trusted, we can forfeit some degree of trust. Notably, the upshot of this position is that our fellow citizens, who have chosen to do wrong in the past, are not returned to full citizenship after they have been punished. They walk among us as second-class citizens, who are more easily caged again. This is certainly an unpalatable position to reach. It is also hard to dismiss out of hand, given that we often distance ourselves from others and distrust them after they have done wrong. If this argument is as undeniable as it is unseemly then two moves are available to alleviate the discomfort. First, we simply need to fully understand the contours and limits of this sort of forfeiture. Certainly, a person who commits one offense does not lose her right to be trusted forever and ever. So, what are the boundaries? The second approach admits that perhaps individuals have forfeited their rights to be trusted, but takes the position that

496

Responsibility and pretrial detention

we must nevertheless return them to a trusting baseline as a part of what we owe each other as citizens (Duff forthcoming).

Beyond dangerousness: flight and the duty to appear Dangerousness is not the only reason why courts detain defendants pretrial. Sometimes, defendants are thought to be a risk of flight. It is important to distinguish between those who simply “fail to appear” and those who “flee” (Gouldin 2018). As Lauryn Gouldin notes, many criminal defendants may fail to appear for reasons unconnected with trying to avoid a trial. Instead, they lack childcare, can’t get off work, or simply forget. For these defendants, more parsimonious and compassionate approaches can be taken, including simply calling or texting defendants to remind them of their court appearances. We are then left with those who truly plan to abscond. Here, we may think that these defendants will fail to do their duty to appear (and thus support just institutions), and the failure to do this duty will also result in our failing to achieve whatever positive aims we have for punishment (assuming the defendant is found guilty). One concern here is that this is not a particularly good reason to deprive someone of her liberty. Would we think we could detain someone for trial on a jaywalking charge even if she was found about to board a flight to a non-extradition country? To be sure, we don’t seek to detain (or extradite!) on this basis, but this shows that the kind of crime charged matters quite a bit for this rationale. The value we lose in calling the person to account, and if guilty, punishing them, is a function of how serious the crime charged is. Even with serious offenses, offenses of such magnitude that we can justify putting someone in a cage to try and sentence the offenders, there is again the question of how this prediction relates to rights. Three potential justifications resurface. First, we are permitted to stop a defendant who, as Antony Duff puts it, has already embarked on a plan to flee (Duff 2013). Notice that like attempts and self-defense, we would need to determine what counts here. One might think that at the point in time at which the defendant forms the intention to flee, she has forfeited her right not to be stopped from executing it. Second, one might argue that we are entitled to override her right. Again, however, this requires not trusting her. Finally, there is the question of whether she has forfeited her right to be trusted to appear, perhaps because she has failed to appear in previous cases (Duff forthcoming). In practice, however, this account would need to be extremely cabined. After all, recall that most cases aren’t important enough to justify the liberty deprivation. And, in practice, most failures to appear are not individuals who abscond to avoid trial. Failures to appear that are justified or excused do not forfeit rights to trust and cannot be used to more easily justify detaining individuals.

Obstruction The final oft-cited reason for pretrial detention is obstruction. This rationale can be collapsed into the prior two. Either the defendant seeks to interfere with the truth-seeking process of the trial, and the evaluation is akin to flight, or the defendant seeks to harm a witness and should be evaluated akin to dangerousness. Because those who obstruct pose a particular threat to the trial process, they are instances in which both the idea of dangerousness and the undermining of the trial (akin to flight) are at play, and we may take these sorts

497

Kimberly Kessler Ferzan

of threats to be particularly troubling. That said, there are no additional interests at work. Although we could specify precisely the way that obstruction assimilates into one or both of the other reasons, the point for our purposes is that there is no additional justificatory lens for detention that this perspective offers.

Whole greater than the sum of its parts To this point, I have suggested that the defendant’s responsibility for different acts potentially results in his lacking a right against pretrial detention. However, one might object that these analyses should not be siloed. Instead, it could be that these forfeitures aggregate to justify detaining someone pretrial. Although I have argued elsewhere that desert and defense may be aggregated to justify imposing more harm than either would allow individually (Ferzan 2018), I am skeptical that aggregation is possible here. The worries both about punishment and about distrust are premised, not on questions about the limits of the upper limits of proportionality, but instead on concerns about our political morality and what we owe each other. If our preventive practices are side-constrained in these ways, it is difficult to see how aggregating the reasons circumvents the objection. Additionally, in instances in which there truly is a defensive forfeiture, because, say the defendant intends to kill a witness, no more is needed. The only space for aggregation, I believe, is that perhaps defendants can give us multiple reasons for distrusting them, such that we are entitled to more distrust than any individual rationale provides. That is, a defendant who had fled several times in the past and who has committed many dangerous acts while out on bail, may, at some point, create legitimate distrust about their future actions.

Bail Despite pretrial detention’s pretense of rigor, the vast majority of jails are filled with defendants who cannot meet their bail. Bail hearings are often only a couple minutes long, where the charged offense and criminal history loom largest (Scott-Hayward and Fradella 2019). What is most puzzling about bail is what I will call “the justificatory gap.” Let’s say your friend owes you ten dollars. When you see him next, he only has a fifty-dollar bill on him. If you desperately needed the money, perhaps you could take the fifty and pay him back forty dollars later. But you wouldn’t get to keep the fifty. Instead, you need some sort of second reason to make up the difference. This inquiry is similar that in the self-defense literature where some theorists argue that minimally responsible threats only forfeit a small amount of their defensive rights, but then a lesser evils calculation justifies imposing further harm (Bazargan 2014). When it comes to defendants who cannot make bail, however, the failure to make bail is somehow seen as also justifying the detention. But the reasons to ask for bail may be insufficient to fill the justificatory gap. The practice of detaining individuals who cannot make bail is ultimately suspect for two reasons. First, if a defendant cannot make bail, then the state should not be asking for that amount. Arguably, the state is entitled to distrust the defendant somewhat because of the crime charged. And, as Antony Duff argues, it can ask for some assurances (Duff forthcoming). Surely, it can ask you to turn over your passport. Perhaps it can also ask you to put some money up to make sure you are fully committed to returning to your trial date. That 498

Responsibility and pretrial detention

said, given that most bail determinations do not take into account ability to pay (ScottHayward & Fradella 2019), most requested financial securities are morally suspect. Even if you ask someone to provide an assurance, you should not ask them to provide an assurance that they cannot provide. If bail was typically set at an amount defendants could meet, they would meet it and be released. Most detentions are therefore simply unjust. However, even if, for some reason the request is legitimate and yet the defendant cannot pay the amount, the question is why, when the defendant cannot pay, the next step is that the state is entitled to detain the defendant. Courts have simply assumed that the defendant’s appearance at trial would justify such a drastic step. Yet, the lesser evils calculation must be done here. What risks are at stake, and how likely are they, such that we are entitled to put someone in a cage when (1) they cannot assure us that we ought to trust them but (2) we lack the kind of forfeiture that would justify detaining them outright? If someone was going to break your pencil, you would not be entitled to cage them to stop them, so why are we entitled to detain those who cannot provide the minimum amount of assurance required? Ultimately, we should be deeply skeptical that our bail practices are remotely legitimate.

Conclusion The view of pretrial detention as an exercise in risk prediction has obscured the complex ways in which the defendant’s responsibility for particular acts may justify pretrial detention. And, in failing to see what is necessary to actually justify our current practices, we instead rely only on predictions, thereby detaining many defendants unjustly.

Notes 1 See Alexander & Ferzan 2018: 200–204 for an exploration of whether defendants who are merely factually guilty may lack a right against punishment. 2 For this view, see Frowe 2018. 3 A retributivist will take the defendant’s negative desert as both supplying the forfeiture of the right and the reason to impose to punishment; moreover, this desert will effectively set both the ceiling and the floor. 4 Wellman (2012) explored such a view, but he later rejected it (2017). 5 For analysis of various approaches to whether justificatory intent is required, see Dillof 2002.

Further reading Mayson, S. (2018). “Dangerous Defendants,” Yale Law Journal, 127, 490–568 provides a probing analysis of the legal and theoretical questions at the heart of detaining individuals on predictive/ eliminative grounds. Scott-Hayward, C. S. & Fradella, H. F. (2019). Punishing Poverty: How Bail and Pretrial Detention Fuel Inequalities in the Criminal Justice System. Oakland: University of California Press, gives a useful overview to pretrial detention practices in the United States. Walen, A. (2011). “A Punitive Precondition for Preventive Detention: Lost Status as a Foundation of Lost Immunity,” San Diego Law Review, 48, 1229–1272, is an exemplar of the sort of forfeiture theory necessary to overcome a presumption of trust.

References Alexander, L. (2016). “Recipe for a Theory of Self-Defense: The Ingredients, and Some Cooking Suggestions,” in C. Coons  & M. Weber (eds.), The Ethics of Self-Defense. New York: Oxford University Press.

499

Kimberly Kessler Ferzan Alexander, L. & Ferzan, K. K. (2018). Reflections on Crime and Culpability: Problems and Puzzles. Cambridge: Cambridge University Press. Baradaran Baughman, S. (2011). “Restoring the Presumption of Innocence,” Ohio State Law Journal, 72, 723–776. ———. (2017). “Costs of Pretrial Detention,” Boston University Law Review, 97, 1–30. Bazargan, S. (2014). “Killing Minimally Responsible Threats,” Ethics, 125, 114–136. Bolinger, R. J. (2021). “Demographic Statistics in Defensive Decisions,” Synthese, 198, 4833–4850. Dillof, A. M. (2002). “Unraveling Unknowing Justification,” Notre Dame Law Review, 77, 1547–1600. Duff, R. A. (forthcoming). “Risk, Responsibility, and Pre-Trial Detention,” in A. Placani & S. Broadhead (eds.), Risk and Responsibility in Context. London: Routledge. ———. (2013). “Pre-Trial Detention and the Presumption of Innocence,” in A. Ashworth, L. Zedner, & P. Tomlin (eds.), Prevention and the Limits of the Criminal Law. Oxford: Oxford University Press. Ferzan, K. K. (2011). “Beyond Crime and Commitment: Justifying Liberty Deprivations of the Dangerous and Responsible,” Minnesota Law Review, 96, 141–193. ———. (2012). “Culpable Aggression: The Basis for Moral Liability to Defensive Killing,” Ohio State Journal of Criminal Law, 9, 669–698. ———. (2018). “Defense and Desert: When Reasons Don’t Share,” San Diego Law Review, 55, 265–289. ———. (forthcoming). “The Trouble with Time Served,” Brigham Young University Law Review. Fletcher, G. P. (1996). “Domination in the Theory of Justification and Excuse,” University of Pittsburgh Law Review, 57, 553–578. Frowe, H. (2018). “Lesser-Evils Justifications for Harming: Why We’re Required to Turn the Trolley,” The Philosophical Quarterly, 68, 460–480. Gouldin, L. P. (2018). “Defining Flight Risk,” University of Chicago Law Review, 85, 677–742. Heaton, P., Mayson, S. & Stevenson, M. (2017). “The Downstream Consequences of Misdemeanor Pretrial Detention,” Stanford Law Review, 69, 711–794. Mayson, S. (2018). “Dangerous Defendants,” Yale Law Journal, 127, 490–568. ———. (2019). “Bias In, Bias Out,” Yale Law Journal, 128, 2281–2300. Scott-Hayward, C. S. & Fradella, H. F. (2019). Punishing Poverty: How Bail and Pretrial Detention Fuel Inequalities in the Criminal Justice System. Oakland: University of California Press. Starr, S. (2014). “Evidence-Based Sentencing and the Scientific Rationalization of Discrimination,” Stanford Law Review, 66, 803–872. Vincent v. Lake Erie, 124 N.W. 221 (Minn. 1910). Walen, A. (2011a). “A  Unified Theory of Detention, with Application to Preventive Detention for Suspected Terrorists,” Maryland Law Review, 70, 871–938. ———. (2011b) “A Punitive Precondition for Preventive Detention: Lost Status as a Foundation of Lost Immunity,” San Diego Law Review, 48, 1229–1272. Wellman, C. H. (2012). “The Rights Forfeiture Theory of Punishment,” Ethics, 122, 371–393. ———. (2017). Rights Forfeiture and Punishment. New York: Oxford University Press. Yang, C. (2017). “Toward an Optimal Bail System,” New York University Law Review, 92, 1399–1493.

500

40 RESPONSIBILITY FOR OTHERS Jenny Steele

Introduction This chapter asks whether legal doctrines of liability for others, and philosophical debates concerning vicarious responsibility, can be usefully compared in a way that is of reciprocal interest. The answers aim to contribute to a better understanding of the general relationship between accounts of legal and moral responsibility in relation to the acts of others. My focus is on private law, where there is recognition both of vicarious liability, and (more exceptionally) of liability for wrongs done by a ‘third party’ (outside the reach of vicarious liability). Both of these will be considered. So far as vicarious liability is concerned, lawyers commonly assume that vicarious liability, since it does not depend on any fault on the part of the vicariously liable party,1 departs from the ambit of moral responsibility (and hence, that it perhaps needs special justification). But some philosophers have argued that vicarious responsibilities of various sorts are not morally dubious or exceptional, but widely prevalent in our lives, acknowledging that this creates some puzzles for responsibility theory. In this context, both law and philosophy are in search of justifications for responsibilities that are acknowledged to exist. To what extent can the search for justification in each of them assist the other? I do not intend to cover old ground by working through the set of possible legal justifications for vicarious liability. This could take us further from the points where accounts of legal and moral responsibility connect. Instead I will consider whether the structures of liability for others recognized in legal reasoning show any similarity to philosophers’ answers to the puzzle of responsibility for others. Where there are differences, these too may be used to explore the different facets of these responsibility questions.

Background: complexity and the least exceptional exception Vicarious liability in law is possibly one of the least exceptional exceptions of all, i.e., it is very common indeed. Though sometimes considered ‘exceptional’ in the sense that it attaches responsibility (in law) without blame, and for the wrongdoing of other parties, many tort liabilities are vicarious. Organizations in particular are mostly (though not

501

DOI: 10.4324/9781003282242-53

Jenny Steele

completely) confined to committing torts vicariously. Although organizations, including companies, are treated as having legal ‘personality’ (a point to which we will refer again), so that in law they are generally treated as agents in their own right just like human persons, in reality their activities are conducted by human agents who are separate from the company. In other words, there is recognized in law to be separate personality, or separate agency, between an organization, and the people who conduct its activities. It is true that there are some instances of ‘primary liability’ on the part of an organization, for example where an organization (such a local authority) has a statutory duty. Here, the duty is that of the organization itself. There are also some very confined circumstances where the wrongful acts which amount to a tort are those of an individual who can be identified with the organization, perhaps as its ‘guiding mind’. Here, there is ‘attribution’ to the company of the individual’s acts.2 But generally, the organization’s liability in tort is vicarious liability, for the acts of another. Since the majority of tort defendants are not individual agents, but organizations, vicarious liabilities are the mainstay of tort liability. May (1983) rightly observed – in tune with the explanation previously – that corporate responsibility is generally vicarious, but reached a conclusion at odds with the present discussion (and with the law), namely that a corporation’s responsibility should be altered to be more in line with (the author’s understanding of) moral responsibility, so that it would be dependent on the organization (or its members) having opportunities to avoid the harm. This is, like the present exercise, an interpretation of the relationship between legal and moral responsibility for the wrongs of others, though more narrowly focused on circumstances where the vicarious party is a corporation. The conclusion reached here about the relationship between legal and moral responsibility is however different. I suggest that where legal development is consistently more demanding than this, it is wise to have regard to what might underlie the position it has reached, even if this is reached rather intuitively. It is possible that law is more sensitive to responsibility concerns than its surface reasoning suggests. Philosophers too work with intuitions and search for an underlying explanation, adopting this more overtly as a method of reasoning. As a methodological point therefore, it would appear that working from a supposed solution to a position that suggests much of legal development must change is not necessarily the most useful approach. It is fruitful to ask whether we can find a basis in responsibility for the general structure of vicarious liabilities in law? Is responsibility in law always consistent with moral responsibility? Arguably, legal liability necessarily embodies a sort of responsibility judgment. To hold someone liable is to find them legally ‘responsible’, in the sense that it is for them to make a response. This is not to say that the liable party is necessarily responsible for the ‘events that triggered the liability’ (Cane 2002), but that there is sufficient reason why a response should be required of them.3 This responsibility frequently is based in ‘historic’ responsibility, as Cane puts it. Indeed, he argues that liability without such historic responsibility is ‘liability without responsibility’, although it is suggested that this does not exhaust the reasons for responsibility. But even here, as Cane also argues, the law’s judgment need not perfectly reflect moral responsibility. May’s analysis therefore jumps too quickly to a conclusion that legal responsibility should always be modelled upon (his perception of) moral responsibility, because we should not cease looking for lessons that either may hold for the other. Equally, it may be argued that where the two are different, there ought to be some sort of reason for this. In identifying these reasons, we may find that in fact, moral and legal responsibility (the type that underpins liability) have much stronger reciprocal lessons than first assumed. 502

Responsibility for others

Reasons why legal responsibility is acceptable may be instructive for thinking about moral responsibility. Equally, thinking about moral responsibility may help to intensify the need to rationalize legal responsibilities. While legal liability and moral responsibility will not necessarily be identified in entirely the same circumstances, this does not mean that they have no lessons for one another, and a consideration of differences can, if anything, lead to more shared rationalization. Divergence between legal and moral responsibility is to be expected partly because of the different implications of liability compared with moral responsibility judgments. It is fair to say that in some respects, and particularly in its range of available responses, law is destined to be pretty heavy handed. An action in tort for example can only be met by a small range of responses compared to all the possible implications of bearing moral responsibility, and there is relatively little scope for these to reflect the basis of the responsibility (or perhaps, agency) involved. Liability responses are generally financial in nature (damages of a compensatory type including nominal or aggravated, exemplary, or in some jurisdictions vindicatory; for some wrongs, in limited circumstances, an apology4). Philosophers, on the other hand, have considered a wide range of possible responses which are to some degree sensitive to the type of engagement with the wrong that is done. These may include self-blame, a sense of responsibility or regret, possible sense of obligation to visit a victim or even to demonstrate regret through flowers as well as apologies (for example Cohen 2017). These are internally focused; as we will see shortly, it has been suggested that there may be an expectation on the part of others that these internal responses exist, and (for some) even a separate responsibility to act accordingly. The responses are also more nuanced, reflecting the relationship between the responsible party, and the harm that is done. It is important therefore to point out that one possible reason for law’s willingness to attach responsibilities for harm done by others is that financial liabilities though onerous may be shifted. Moral responses are more personal, so that obligations are not susceptible to being shifted in the way that a financial liability may be shifted. This property of legal liability is of some significance in the way that responsibility is understood in law. It might otherwise be thought that such an onerous ‘response’ cannot be explained in terms consistent with the nuanced approach to moral responsibility for others. I suggest that if we look more closely to legal liabilities in the light of this, we may find important elements of a response to our opening question, namely whether there can be a basis in moral responsibility for vicarious (legal) liabilities? Considering institutional factors that shift the burden of liability in this way does not necessarily mean leaving questions of responsibility behind. Instead, it may help to explain the particular responses developed intuitively by the law, and thus shed different light on the general question of our responsibility for acts of others. At the same time, considering the complexity of the philosophical debates over responsibility for others should offer lawyers some reassurance against the suspicion that liability floats free of any ‘moral’ underpinnings in this context. Attending to the complexity of institutional contexts and moral responsibility potentially offers some assistance to both law and philosophy.

Vicarious liability and organizational responsibility Vicarious liability has expanded without real agreement over how to conceptualize it, so that it reflects both the practical and the intuitive nature of legal development: it is indispensable, but in responsibility terms it is awkward and disputed. In order to explain it, do 503

Jenny Steele

we simply fall back on practicality, on the idea of incentives, or on some rather general idea of ‘fairness all things considered’ (you take the benefit, and therefore should take the risk)? All of these ideas have played a part in the law, but they do not address our initial question. Alternatively, do we accept that organizational responsibility is in fact part of our sense of how social norms of responsibility operate? This introduces a different perspective on our initial question, by asking about expectations of a community in relation to responsibility. Perhaps, a society in which organizations escaped responsibility would not only offend in economic terms (for lack of incentives), but because we do in fact attach agency to organizations (and maybe groups).5 The legal personality of organizations is far reaching. As one illustration of this, it is common for organizations to make use of human rights.6 Absence of vicarious liability would be consistent with separate personality, but inconsistent with other elements of the way that the agency of organizations is generally understood and indeed given effect. Here then is a further question. It was argued above that we do attach a sense of agency to organizations, to the extent that we consider them to be responsible. How then might we connect this view to the ways in which certain philosophers have approached the puzzle of what would be lacking if we as individuals did not accept or somehow take on some responsibility for harms done by those connected to us in various ways? In philosophy, the sense that this is necessary is not so widely accepted. But the efforts of those philosophers who have identified a sense of responsibility for others may well be of assistance. If so, the link is unlikely to be completely straightforward. To what extent can we, or should we, seek to retask ideas connected to taking, or having, responsibility for the acts of others, in order to make sense of the (more widespread, but still disputed in terms of its rationale) idea of organizational or individual liability for the acts of others? We should recognize that vicarious liability in law began its journey as a doctrine which did indeed link individuals to one another. In the early days of vicarious liability, before the rise of larger corporate structures, employment was more literally a question of ‘master’ and ‘servant’, and the conceptualization of vicarious liability as ‘attributing’ the tort to the master (the master’s tort theory) was a serious candidate. Modern vicarious liability has however plainly developed beyond these origins in response to changing organizational structures. ‘Attribution’ of the action to the vicariously liable party is not a technique that is now used. Nevertheless, while extended to fit new circumstances, vicarious liability is expressed in terms which could also in theory be applied to the responsibility of an individual employer.

Philosophy and vicarious responsibility Recent philosophical discussion has sought new bases for vicarious responsibility, against the background fact that liability for others appears to be beyond the recognized ambit of responsibility, which is typically identified with the agency of the individual. Different accounts have been offered, couched in terms of ‘penumbral’ agency (Enoch 2012) or ‘moral entanglement’ (Goetze 2021). Although these are rival accounts, with significant differences, both of these approaches deal with questions which work in the opposite direction from the contemporary legal questions identified previously. Philosophers have debated the ways in which an individual might have or acquire responsibility either for the actions of another individual, or for a collective with which they have some connection, such as their country. This reflects the fact that questions of responsibility are generally associated with 504

Responsibility for others

human agents. Legal responsibility, as we have seen, is equally applicable to organizations having separate ‘personality’, so that the question is often when such an organization can bear responsibility for the acts of human others. How have philosophers approached the question of responsibility for others? Some have approached the question by beginning with Williams’ ‘moral luck’ idea (Williams 1981). Enoch, while not accepting the argument for moral luck, nevertheless begins with Wolf’s elaboration of a particular virtue – to take responsibility generously even beyond the boundaries of blameworthiness – which she associated with moral luck (Wolf 2001; Enoch 2012). Williams identified a form of ‘agent-regret’ which arises even where we have done nothing wrong, or beyond the extent of our wrongdoing, because of the way things have turned out. This is argued to reflect our sense of agency, which extends beyond that which we have done ‘wrong’. Wolf (2001) added to this that there is a nameless virtue of taking responsibility generously for the effects of our actions. While this is clearly directed to ‘taking’ (not imposing) responsibility, it is an influential step in the argument towards vicarious responsibility. Wolf associates the virtue with willingness to be accountable for ‘what one does’, the latter to be understood in an expansive way. Indeed she included some vicarious responsibilities in her examples – the willingness to offer money when one’s child breaks a vase, or to take blame for harms done through the practices of an ‘agency’ of which one is the head, though not without limitation.7 Equally, we have expectations that agents will accept responsibility in these ways. This virtue and these expectations are linked by Wolf with outcomes in the world that lie beyond the sphere of our will. In defining ourselves, we should not minimize the significance of contingency. These general lessons could in fact be applied directly to the responsibilities of organizations, whose responsibility in the world might be modelled on the responsibility of human agents. We can go further, to suggest that acceptance of contingencies is, arguably, a substantial feature of interactions between the ‘parties’ with which private law deals. Dealing with those contingencies is a major aspect of private law, and possibly a key element of its role which distinguishes it from criminal law. From these starting points, it has been suggested that we can amplify the argument made by Wolf concerning the consequences of actions that are not our own. Enoch (2012) suggested that this may occur in the ‘penumbra’ of our agency where wrongs are committed by people, or collectives, to whom we are connected. One way of putting this link is that while we can rightly say we are not at fault, there may be thought to be something wrong with people who simply reject all responsibility in such cases.8 Examples – developed from Wolf’s account – include harms that are done by our children, or by our country, in which we have no personal involvement. Enoch, as we have seen, agreed that sometimes there may be a virtue of taking responsibility where harm has been done by an individual or a group to which we are connected. However, he argued that we are not responsible per se; rather, we are responsible only to the extent that we ‘take responsibility’. This may seem quite far from justifying the kinds of responsibility that underlie legal liability, since legal liabilities of most kinds are imposed. But even so, Enoch does argue that sometimes, we may have a duty to take responsibility in this sense. This would seem to go further than Wolf’s suggestion that there is an ‘expectation’ that responsibility will be taken to an appropriate extent. Goetze (2021) has gone further still, maintaining that the stage of ‘taking responsibility’ may be unnecessary in some contexts (Goetze 2021). Where there is a significant degree of ‘moral entanglement’, in his expression, we may acquire obligations to ‘make things right’, whether or not we had an obligation in relation to the initial event, and without a ‘taking’ of responsibility. This 505

Jenny Steele

requires some focus on the nature of the obligations that are in play. These may simply be obligations to do something in the present, following the harm, and not obligations to protect. There is therefore some philosophical pedigree for the idea of obligations to make good, which are separate from obligations to keep safe, and may be either taken (assumed), or imposed because they ought to be taken, or simply arise because of our link to the harm. These ideas may be very useful in the different context of considering an organization’s responsibility for the wrongdoing of employees and of others with whom it is connected. The organization may not breach its own duties, and the actions of the harm-doer may not be attributed to it. When, then, is it responsible for the actions of these others? And is there a separate notion of being under obligations to act that are separate from obligations to keep safe? Some of the complexity of the answers is considered shortly. But in broad terms, can these ideas about attribution, penumbral agency, duties or virtues of taking responsibility, and obligations to make things right offer any assistance in the context of legal liability of – predominantly – organizations? The links may be numerous, and quite subtle. Only some of them can be addressed here, and this examination is not intended to be exhaustive. The links between legal liability and philosophical accounts of responsibility for others are, like those accounts themselves, complex. How and to what extent can we make a link between responsibilities that are taken in conditions of penumbral agency, or that are thought to arise in conditions of ‘moral entanglement’, and legal liability? Legal questions arise in a specific context. Here, they arise in respect of a claim for compensation. Perhaps these legal consequences appear too heavy handed. There is a gulf between the suggestion that we should sometimes do something to accept responsibility or take steps to ‘make things right’ in the absence of fault, and the suggestion that we should in fact make good the losses.9 At the same time law’s response appears impersonal. There is no duty to send flowers personally or to turn up at the hospital. These issues can be addressed in more than one way. One way is to argue that private law simply deals with corrective justice. Its available responses are limited by this function. That is not to say that there is an imperative to do corrective justice in all situations (Gardner 2014). But it means that law is choosing between liability, or no liability, and (in most cases) nothing more or less. But we can also dig further for signs of responsibility in law. We can recognize that the law incorporates many structures which make it possible to make liabilities effective and to some degree proportionate. After all, the resulting liability is unlikely to be paid for by any individual responsible party. Where it is the responsibility of an organization, its calculation in terms of loss to the claimant means there is no attempt to bring home a wrong (contrast, for example, competition law fines designed to be imposed on organizations and to influence their activities). This impersonal character is not necessarily a dilution, or even a contradiction. Rather than seeing this as a separate issue from responsibility, it could be seen as an achievement in responsibility terms.10 Depersonalising responsibility makes it possible to address the consequences of harms, including harms that are widespread and significant, in the sense that is relevant to private law. We don’t have to let responsibility wreck lives or destroy organizations. Both vicarious liability, and insurance, are elements of this. This impersonal character does not necessarily snuff out the sense of responsibility. Rather, responsibilities are varied, so that for example taking responsibility vicariously, or taking responsibility to insure, are significant in their own right. In the driving context (where the notion of ‘moral luck’ was first brought close to the law) the duty to insure has indeed taken on the 506

Responsibility for others

guise of a more significant duty than the duty to take care itself. Failure to insure can leave a driver open to the full weight of legal responsibility, from which most individuals are protected.11 The point is not therefore that these structures or ‘realities’ have the effect of removing responsibility from legal thinking. Rather, they alter the impact of responsibility and in some sense, make it possible (tolerable) for responsibility to have the kinds of consequences that are involved in private law. To suppose that responsibility is negated by the shifting of financial burdens to insurers, as some lawyers have argued, is to misinterpret ‘responsibility’ as confined to the legal liability to make good. This is far from the case. These structures supporting liability in fact enable responsibility, channel responsibility, and perhaps shift the forms of responsibility, introducing new variants. They allow for the heavy-handed consequence of responsibility which takes the form of liability, by avoiding disproportion. Lawyers need to be alert to these structures and their flaws, as they may be more significant in supporting responsibility than has been appreciated. But the key point which lawyers sometimes overlook is that the idea of proportionate response is present in the writing about moral luck and moral entanglement. What is there a responsibility to do? Because law has fixed responses, it can offer this element of proportion only through its own mechanisms. Those mechanisms  – organizational responsibility, insurance, and so on – have significance in understanding responsibility. They are part of the subject for consideration. There is promise therefore in the notion that philosophical approaches may assist, so that vicarious liability does not appear to contradict moral accounts, or to be explicable only in practical terms (as though these were not part of any analysis of responsibility). This diversification of responsibility may itself be of interest to philosophy. But at the same time, it is worth considering whether the law’s approach to vicarious liability itself, developed over time, gives any further hints.

Vicarious liability in law: questions of connection Law has to be usable. Therefore, it generally offers clear structures for argument, even if there is less agreement on the fundamental reasons behind them. The current approach to vicarious liability in English law is a good example of this, leaving plenty of room for disagreement as to which of the competing justifications is most significant but at the same time providing some clear structure for analysis. As law currently stands, two ‘connections’ are sought. The first is the connection between the ‘tortfeasor’ (whose actions or omissions caused the harm), and the vicariously liable party. This seems familiar from discussions of moral entanglement and penumbral agency, since it focuses on the potentially liable party’s connection to the wrongdoer. At present, the courts ask whether this relationship is sufficiently ‘akin to employment’ to justify liability. Of course, this is a narrower version of the question than the wide variations that are considered in philosophical debate. This reflects the extension from specific origins in employment. But it also reflects the fact that law cannot possibly attach liability to every example of responsibility for harm. As previously, there needs to be some form of workable solution to the possibility of disproportion, and law will move to new areas with caution.12 The second connection is between the tort (or harm), and the relationship just considered. How does the relationship stand in relation to this tort? Now, this could be a proxy for asking ‘was this really something done for your benefit or in pursuit of your goals?’ That would appear to narrow the range of vicarious liability considerably compared to the 507

Jenny Steele

philosophical debates. But at present, the second limb does not operate in this way. It has become recognized that it is much broader, because it includes liability for acts expressly forbidden and in contradiction to the interests of the ‘employer’. It includes liability for bad apples, up to a point.13 It is worth considering the reasons for this. I do not think the law has gone astray here. Those who do think it has gone astray tend to believe that it is beyond the bounds of responsibility or fairness to hold organizations responsible for deliberate wrongdoing or rule breaking from those relevantly connected to them. But the philosophical literature contains no such limit. If I am responsible, or am expected to take responsibility, for the wrongdoing of my child for example, there is no hint that this responsibility is confined to negligent harms, rather than deliberate ones. In fact the reverse may even be true, and if the deliberate harm is done in order to spite the parent, then my ‘moral entanglement’ may be all the stronger. I may be the innocent party, but the harm that has been done is related to my relationship with my child. Deliberate harms by my government with which I disagree may still give rise to some sense that I should take responsibility of some sort – even if that is responsibility to act now. So, a simple suggestion that liability here is beyond the bounds of responsibility in philosophical terms does not stand up. However, we have the problem of law’s clumsy response. To repeat the points made previously, the existence of some responsibility does not mean that I am just as responsible as if I was a deliberate wrongdoer. Is tort too heavy handed to deal with this? It is here that it is appropriate to consider institutional and practical solutions to support responsibility judgments. There are institutional ways of reflecting the type or degree of responsibility involved. For example, it is possible to insure a deliberate harm done by another (provided they are not too closely ‘associated with’ the organization!),14 whereas one cannot insure against the products of one’s own deliberate wrongs.15

Vicarious liability in law and philosophy Organizations have constitutions and reasons for existence expressed in terms of goals. Their members even adopt institutional statements about values for the organization. They have power linked to the pursuit of these, as well as benefiting from legal personality and human rights. The more satisfying accounts of vicarious liability do not just suggest that organizations should pay ‘because they can’. They should pay because they should bear responsibility; the implications of that liability are pertinent to its proportionality as a response. The two present ‘limbs’ of vicarious liability explored previously give expression to when this is the case, based on the connection between the organization or employing individual, and the harm that is suffered. It is not as narrow as doing work to the benefit of the organization, but not as broad as ‘penumbral agency’ in general. It expresses a relevant link in two dimensions. If we look at the law’s approach to vicarious liability in light of philosophical discussions of vicarious responsibility therefore, we can see a complex relationship between the two. Vicarious liability is not only a necessity for the legal system, but can be expressed in terms of responsibility without collapsing into a much narrower version that is dependent on the party at fault working for the benefit of the vicariously liable party, for example. Departure from attribution of the tort to the ‘master’ does not destroy the possibility of a responsibility-based account. Asking about the link between organizations as agents, and the acts of those associated with them, does not require any particular stretching of responsibility as ordinarily 508

Responsibility for others

understood, contrary to what some lawyers may suspect. If anything, the limitations that are in place are required partly because the law is concerned with certain responses, which are necessarily more onerous than general regret or duties to acknowledge one’s agency or entanglement. Tort law is naturally concerned with ‘how things turn out’ (to use Wolf’s expression).16 But it has been suggested previously that institutional structures which allow for the depersonalisation of responsibility through law operate in order to modify the impact of responsibility judgments. Responsibility continues to attach, but it is attached with awareness of the implications. This is, in fact, typical of responsibility (and also typical of law). A  nuanced approach both to responsibility and to appropriate actions and expectations are inherent in philosophical discussions of responsibility for others, including Wolf’s discussion of the virtue of taking responsibility, and its extension to harms done by those connected to us (Wolf 2001). The implications of legal remedies are an important element in determining how far the law can push responsibility, and amount to a reason for it to advance cautiously. But the law is not operating beyond the bounds of responsibility. Rather, even in its current form, it is asking additional questions beyond those which might be used to establish a proper basis for being responsible, or taking responsibility. Hopefully, with these points in mind, the way that law has advanced to consider two kinds of connections when determining vicarious responsibility may also be of help to philosophers – not because of the limits that are set by the law, but because of the kinds of issues that have appeared to be important in legal development.

Third-party acts in private law – more on responsibility It remains to consider, in the light of all this, whether we can learn any more from considering the distinct legal position on liability for third-party acts. This kind of liability, reflecting the continuing strength of the idea of separate agency, operates in limited circumstances. In fact, the importance of the distinction between causing harm and failing to prevent it from being caused by someone else has been given considerable emphasis by the UK Supreme Court in recent years.17 Two of the recognized circumstances where there may be liability for harm that is caused by another’s agency are creation of a danger, and failure to control the third party. These are based on a connection between the liable party and the event. As such, they don’t seem to raise particularly challenging responsibility questions – if anything, it is their relative narrowness that is noticeable. More significant for present purposes is the third exception, which is known as the ‘assumption of responsibility’. It should be of immediate interest that ‘responsibility’ is evident in the very test for liability here. But it will also be noticed that the implication of this head of liability for the consequences of another’s acts is that the responsibility at work appears to be assumed or taken. Does this mean that the law in this instance reflects the idea, present in some philosophical accounts, that responsibility for consequences of another’s act might sometimes exist because it is taken, and not because it exists in any case? It is interesting to notice that what the law does here is not to look for an actual assumption in the sense that would apply in a neighbouring area of law, namely the law of contract. Rather, it looks for circumstances which suggest that responsibility can be treated as having been taken. This notion has offended some lawyers, as it appears to conflate contractual and tortious ideas, and to introduce a contradictory idea of liability that is entirely distinguished and justified by the fact that it is assumed, when in reality it is not. The previous discussion may help us to argue that what we see here is not such a conflation, but a legal 509

Jenny Steele

obligation to take responsibility, in these circumstances, in the sense that the party’s actions (a stronger idea than their activities in any general sense) put them in a position where they are strongly connected to the harm suffered. A party with agency (which as we have seen includes organizations) ought to take responsibility in such cases, and we ought to understand this responsibility in a particular way, relating to obligations to make good. We therefore manage to skip over the issues connected to liability for others, by dealing directly with the obligation to take responsibility given particular actions or relationships. The law is quite specific about this (as with our observations previously, it is specific, but not completely clear), and certainly does not give effect to this idea in all those circumstances in which it could. Structurally, instances where these considerations apply are cases where the defendant takes on responsibility to the claimant (not, for the acts of the party who caused the harm). It is not a case of connection with the wrongdoer in the same way as vicarious liability. The analysis is differently focused. Naturally, this has been explored within the law as a different possible route to overcoming the problem of organizational responsibility and its justification.18 But it has not been used in this way very often in English law (though assumptions of responsibility can be observed in other contexts, such as the liability of local authorities for children who have been taken into their care). It has been used to establish the liability of a school to a pupil injured during a swimming lesson, through the negligence of third-party swimming teachers,19 and to establish the liability of a dentist’s practice for harm caused by independently contracted dentists operating through the same surgery.20 But so far, it has not expanded to take the role generally performed by vicarious liability. What is meant by responsibility here? It appears to be that the defendant identifies themselves as someone who will protect your interests. They may say so (the classic ‘leave it to me’), or their relationship may be such that they are in a protective relationship without saying anything or even choosing anything (say, a school and its pupils). Now they are identified as the ones to take care of the injured party. But in what sense? Arguably, what is in play in the case of an ‘assumption of responsibility’ is the taking on of a particular responsibility, to make good harm that arises. In other words, as with vicarious liability, the responsibility does not depend on what they have done wrong, and is not coextensive with a duty to avoid (though such a duty may also exist). But this time, it depends on their relationship with the party who has suffered harm rather than their relationship with the wrongdoer or the relationship between the wrong and their own purposes. Here it could be argued that law has intuitively recognized a responsibility to respond or make good, which is not purely ‘secondary’ to the agent’s own failure to protect. Their failure to protect may not be capable of being identified. So this is expressed by the law as a responsibility to make good where there was a responsibility to protect even where the only failure is that of another party. And perhaps it is this very responsibility to make good, which is ‘assumed’ – or treated as assumed, because it should be assumed. Vicarious liability did indeed turn on the connection with the harm, according to the previous analysis. But the implications of liability were important factors in limiting its expansion. Assumption of responsibility in the context of third-party acts may be approached as much more directly related to the duty to make good. And here I think there is something of interest in the comparison with philosophical literature to the extent that it focuses on the ‘taking’ of responsibility. Because some of those responsibilities too, in essence, are about the consequences of responsibility, and about how to respond to harm, rather than about avoidance of harm. In this instance, unlike perhaps the case of vicarious liability, the nature of legal liability (making good through compensation) is part of the rationalization 510

Responsibility for others

for responsibility. The responsibility that is assumed is in broad terms a responsibility to make good. At the same time, the structure of legal reasoning here appears to support the notion that sometimes there are indeed obligations to take responsibility.

Conclusions Have law and philosophy helped each other here? I hope so, with important caveats about the limited range of law’s responses, and (on the other hand), the extra resources that legal institutions may bring to bear to resolve tensions around the extent of responsibility without blame. Both may gain from attempting to appreciate the complexity of the other’s position. It would be wrong to write off law’s responses as overly broad and inconsistent with moral readings of responsibility, or even as confusing the imposition of responsibility with the taking of responsibility (‘assumption of responsibility’), as many lawyers do. If anything, law selects a range of possibilities within the broad reach of responsibility, treating organizations as agents capable of entanglement in much the same way as individuals are, but adding the proviso that the relationship with the individuals who cause harm must itself be related to the harm. Law’s potential responses are relatively inflexible, and care is needed in comparing with more nuanced styles of responsibility; in important respects, law has the resources to answer this problem. A sense of proportion is not foreign to thinking about responsibility, and in law that sense of proportion may be satisfied by institutions which depersonalize the impact of liability. In fact, ‘personalising’ responsibility is very much the exception. At the same time, law’s intuitive and practical development of patterns of responsibility can offer clues to understanding the complexity of philosophical positions in relation to responsibility for what others do.

Notes 1 Liability imposed without fault is described as ‘strict’. 2 Recent consideration by the UK Supreme Court of the limited circumstances in which ‘the acts and intentions of real human beings were to be treated as the acts and intentions of the company’ can be found in Singularis Holdings Ltd v Daiwa Capital Markets Europe Ltd [2019] UKSC 50. There is no presumption that the acts and intentions of even a sole director can necessarily be attributed to the company, and whether there is attribution depends on the reason why the question arises. 3 Cane (2002) gives as examples the liability to make restitution where there has been an unjust enrichment (which need not be through the wrongdoing of the liable party), or where there has been a fraud (which may be on the part of another). Liability here is subject to defences where it may be unjust to require a response, for example where the defendant has spent the money or made a gift of the assets received, and has done so in good faith. Here it could be said there are reasons why there is not a responsibility to make good. 4 In English tort law, a formal role for apologies can be found in relation to defamation, where a defendant can offer to ‘make amends’, including an apology and suitable damages. The aim is to achieve vindication of the reputation of the defamed party – that is, to make things right. For innocent defamers (those who had no grounds to think they were defaming the claimant), the offer has the additional benefit of acting as a defence (Defamation Act 1996, ss 2–4). Notice that responsibility – whether to apologize and correct, or to compensate – is expressly accepted by the law in this case where the harm is innocently caused, and the response varies according to awareness (roughly speaking, fault). 5 See also the discussion by Cane (2002: Chapter 5), arguing that it would be to overlook the social function of responsibility to treat it as purely a property of human agents and not groups or corporations.

511

Jenny Steele 6 Note for example use of human rights to possessions in ECHR art 1, Protocol 1 by corporations, which have also used art 10 (for example in relation to defamation and privacy) and art 8 quite liberally. 7 For example, Wolf suggested that it would not be appropriate for all German citizens to take responsibility for the actions of the Third Reich. 8 Wolf (2001) argues that it is a moral virtue to ‘take responsibility’ liberally in this sense, but not without limit. 9 Or at least, one may think so. Wolf (2001) appears to argue that taking the economic consequences – paying for the broken vase – is something to be expected. Perhaps then taking responsibility in the sense of making good the loss is not perceived to be the most onerous response. 10 See also Voyiakis (2021), proposing that responsibility may be found in a distinct form in no- fault compensations systems like the one operating in New Zealand. 11 This occurs because the Motor Insurers’ Bureau (MIB) (who exist to compensate the victims of uninsured and untraced drivers) are able to – and do – proceed against those uninsured drivers for a contribution, sometimes seeking bankruptcy (Merkin & Steele 2014). 12 This is currently captured in terms of ‘incremental development’ – moving on the basis of what has been tried, rather than extending general principles into wholly new areas. 13 There has been vicarious liability in many instances of child abuse perpetrated by employees and those similar to employees where their ‘employment’ is connected to the abuse, amongst other examples. Some limits were set by the UK Supreme Court in Morrison Supermarket v Various Claimants [2020] UKSC 12, where the wrongdoer’s motive was to ham the employer by effecting a major breach of data protection rules. 14 This is another experiment in attribution. 15 See also the different treatment of ‘innocent’ defamers, in fn 2. 16 Wolf 2001: 3. 17 Poole v GN [2019] UKSC 25. 18 An interpretation of the key case of historic sexual abuse of children in the UK, but not the majority judgment: Various Claimants v Catholic Child Welfare Society [2012] UKSC 56. 19 Woodland v Swimming Teachers’ Association [2013] UKSC 66. 20 Hughes v Rattan [2021] EWHC 2032 (QB).

Further reading S. Wolf, ‘The Moral of Moral Luck’, Philosophic Exchange 31(1) approaches ‘moral luck’ in terms of responsibility and although it is not the main focus of the essay, it importantly includes the question of responsibility where harm has been done by others with a connection to ourselves. Enoch, D. (2012). “Being Responsible, Taking Responsibility, and Penumbral Agency,” in Ulrike Heuer & Gerald Lang (eds.), Luck, Value and Commitment: Themes from the Ethics of Bernard Williams (pp. 95–131). Oxford: Oxford University Press, more directly considers how to approach responsibility for others, building on the insights of Susan Wolf, and proposing a duty to take responsibility. Goetze, Trystan S. (2021). “Moral Entanglement: Taking Responsibility and Vicarious Responsibility,” The Monist, 104, 210–223, takes a more direct approach, arguing that Wolf’s analysis helps us to see how responsibility for others may be established without being ‘taken’ by the relevant party. Cane, P. (2002). Responsibility in Law and Morality. Oxford: Hart Publishing, includes an important chapter (Chapter 5) which explores the basis of the responsibility of corporations and groups. Despite its age, the leading work on the basis of vicarious liability in the law of tort remains Atiyah, P. S. (1967). Vicarious Liability in the Law of Torts. London: Butterworths. Another recent article which debates vicarious responsibility in terms of moral luck is Kuan, T. (2021). “Vicarious Responsibility and the Problem of ‘Too Much’: Moral Luck from the Perspective of Ordinary Ethics,” The Monist, 104, 168–181.

References Cane, Peter. (2002). Responsibility in Law and Morality. Oxford: Hart Publishing. Cohen, Andrew I. (2017). “Vicarious Apologies as Moral Repair,” Ratio XXX, 359–373.

512

Responsibility for others Enoch, David. (2012). “Being Responsible, Taking Responsibility, and Penumbral Agency,” in Ulrike Heuer & Gerald Lang (eds.), Luck, Value and Commitment: Themes from the Ethics of Bernard Williams (pp. 95–131). Oxford: Oxford University Press. Gardner, John. (2014). “What is Tort Law For? Part 2: The Place of Distributive Justice,” in John Oberdeik (ed.), Philosophical Foundations of the Law of Torts (pp. 335–353). Oxford: Oxford University Press. Goetze, Trystan S. (2021). “Moral Entanglement: Taking Responsibility and Vicarious Responsibility,” The Monist, 104, 210–223. May, Larry. (1983). “Vicarious Agency and Corporate Responsibility,” Philosophical Studies, 43, 69–82. Merkin, Rob & Steele, Jenny. (2014). ‘Policing Tort and Crime with the MIB: Remedies, Penalties and the Duty to Insure,’ in Matthew Dyson (ed.), Unravelling Tort and Crime (pp. 22–57). Cambridge: Cambridge University Press. Voyiakis, Emmanuel. (2021). “Moral Luck, Responsibility, and Systems of Tort Liability,” Res Publica, 21, 271–286. Williams, Bernard. (1981). “Moral Luck,” in Moral Luck: Philosophical Papers 1973–1980 (pp. 20–39). Cambridge University Press. Wolf, Susan. (2001). “The Moral of Moral Luck,” Philosophic Exchange, 31(1).

513

41 LEGITIMATE DIVERGENCE BETWEEN MORAL AND CRIMINAL BLAME Alexander Sarch

Legal moralists think the criminal law should mirror morality absent overriding reasons to the contrary (Husak 2016: 34; Duff 2014: 229–230).1 One argument for this view is that criminal law aims to regulate conduct by speaking in a moral voice. Through criminalization, conviction and punishment, it condemns conduct it views as wrongful. We do not want these condemnatory messages to be false. So conduct should not be criminalized unless it really merits such condemnation thanks to the offender’s culpable wrongdoing (Simester 2021: 4).2 This legal moralist view is compatible with many accounts of the justification of criminalization and punishment,3 but fits especially well with influential forms of desert-constrained pluralism, which view culpable wrongdoing both as a reason in favor of criminalization and a necessary condition therefor (Berman forthcoming: 2). While many think criminal law should mirror morality, it’s also widely acknowledged that it inevitably won’t do so perfectly (Husak 2016: 41). This is not only because of immoral criminal laws, which leave no choice but to abolish them. More interestingly, there may be legitimate ways the criminal law might diverge from morality. Some of these might be isolated practical compromises due to efficiency or resource constraints. Legal moralists recognize this possibility. Husak suggests that “deviations [from the moral ideal] should be regarded as occasions for regret, as invitations to try to do better,” while accepting that “retreat [from the ideal] is almost certainly needed” (Husak 2016: 44). But is it always a matter of deep regret when the content of criminal law diverges from morality? Might this not also be justifiable based on our principled commitments? If so, perhaps divergence from morality isn’t always an invitation to do better. There might be principled reasons, not always consequentialist in nature, for the criminal law to eschew some moral nuance. Consider the offense of statutory rape: having sex with anyone under the age of consent, often set at 16. The offense uses a brightline age-cutoff to target the underlying wrong of exploitative sex. But not every instance of the offense will involve exploitation, as in cases of sex between two mature youngsters in a loving, committed relationship. The offense thus is overinclusive relative to morality. Nonetheless, as Cornford argues, conditioning liability on actual exploitation  – either by making exploitation an offense-element or recognizing a non-exploitation defense – would undermine the values DOI: 10.4324/9781003282242-54 514

Legitimate divergence between moral and criminal blame

promoted by using a brightline age limit to define the offense (Cornford 2017: 640–643). It not only becomes harder for citizens to know what conduct is prohibited, but also threatens privacy and autonomy. It would tend to put complainants “on trial” by forcing them to rebut defendants’ predictable arguments that the sex in question was not exploitative. Moreover, it would grant power to law enforcement to examine minute details of citizens’ sexual activities to determine which sexual encounters are exploitative, which carries serious privacy implications. Accordingly, the brightline rule has advantages grounded in institutional design and moral principle, so the over-inclusiveness of the brightline rule is perhaps justifiable. This chapter focuses on such apparently legitimate divergences between criminal law and morality. Institutional design considerations tend to be undertheorized and are too often mentioned as an afterthought once the moral ideal has been characterized. This is what I hope to help remedy here. Section I sketches an analytical framework for evaluating divergences between criminal law and morality within the orthodox legal moralist view. Section II sketches some core cases of plausibly legitimate divergence between criminal liability rules and morality. Section III formulates several mid-level principles, not all of which are directly consequentialist in flavor, that can help justify divergences of this kind. Section IV applies the previous framework and principles to the examples in Section II to reach provisional conclusions.

I  Analytical framework A)  Legal moralism and the criminal law The most basic commitment of legal moralism is the desert-constraint: conduct may permissibly be criminalized and punished only if and to the extent this is deserved. Desert is typically understood in terms of the action’s wrongness and/or the actor’s culpability for doing it. The orthodox view among legal moralists assumes that both wrongness and culpability are required for permissibly criminalizing an action and punishing someone convicted of doing it. For simplicity, I treat the two requirements together and talk simply of a culpable wrongdoing constraint.4 There is debate about whether this is a strict constraint or a defeasible (if weighty) one.5 I assume the latter view because all legal moralists agree that at least it holds. Beyond this negative constraint, many legal moralists also think wrongness and culpability entail affirmative reasons in favor of criminalization (perhaps only within the public sphere). There is more debate about this claim (Duff 2014: 229–230), but it fits well with the pluralist views that, with the culpable wrongdoing constraint, represent the dominant account of when and why criminalization and punishment are justified. Here I assume that culpable wrongdoing is both a negative constraint on and a positive reason for criminalization and the resulting punishments. Wrongness and culpability can be understood in terms of reasons. An action is wrong when the applicable normative reasons bearing on what to do decisively (or in some other binding sense) disfavor doing this action (Edwards & Simester 2019: 55). One is culpable (equivalently, blameworthy) for a wrong to the extent that it manifests insufficient regard for the properly protected interests and rights of others (Edwards & Simester 2019: 55; Alexander & Ferzan 2009: 67–68; Tadros 2005: 250; Yaffe 2011: 38). Insufficient regard, in turn, is commonly understood as defective engagement with – i.e., failure to properly 515

Alexander Sarch

recognize, weigh, and be duly motivated by – the applicable normative reasons in virtue of which the action is wrong (or appropriate proxies therefor). On this reasons-based picture, the content of the criminal law can diverge from morality insofar as the legally recognized reasons come apart from the moral ones. The law might recognize reasons morality doesn’t, or it might take them to have different weights than morality does. This could lead to criminal law recognizing different wrongs than morality or differences in the degree of wrongness attributed to them. Likewise, if the reasons the criminal law demands, on pain of culpability, that we not manifest disregard for were to diverge from the moral reasons, then this could lead to differing judgments in law and morality regarding which actions are culpable and how culpable they are.

B)  The framework This reasons-based sketch of how criminal law might diverge from morality does not answer evaluative questions about criminal law overbreadth: whether it’s justified, when it’s regrettable. Here is a framework for how legal moralists can address such questions. The framework maps the normative features of four central cases. 1) Perfectly ideal criminalization. In the ideal case, the legislature passes a law declaring some conduct, C, to be an offense and this not only substantially reduces harm, but the culpable wrongdoing constraint is met: every token of C (absent a defense) is culpably wrong. This is permissible for the legislature and there’s nothing regrettable about it. Citizens have a moral obligation to obey and are culpable if they do C (absent a defense that should be recognized). 2) Prima facie overbroad offenses. The remaining three cases involve apparent overbreadth – offenses that prima facie conflict with culpable wrongdoing constraint. Here, a direct application of moral theory doesn’t straightforwardly reveal culpable wrongdoing. These cases may or may not really conflict with the constraint, depending on whether institutional facts or the operation of law changes the moral status of the conduct and make it culpably wrong. Category 2) offenses are likely to be not mala in se, but malum prohibitum (culpably wrong because prohibited). But this category must be further sub-divided. 2.a) Prima facie, but not in fact, overbroad. Some offenses initially seem overbroad but on closer inspection actually involve culpable wrongdoing. Suppose it’s an offense to do some risky conduct (like dispose of hazardous waste) without a license even when one knows one can do it safely. Or maybe it’s an offense not to report certain transactions carrying a high risk of money laundering (perhaps a cash transaction over $10,000) even when one knows the transaction is entirely legitimate. Such offenses may seem overbroad, but there are strategies from the mala prohibita literature for arguing that they in fact do involve culpable wrongs. First, Rawls famously suggested we have a duty to uphold just institutions (1971: 115). Assuming the licensing or anti-money-laundering regime is just, one incurs a moral obligation to do one’s part to support these institutions by acting consistently with their requirements and is culpable for failing to do so (without justification or excuse). Second, some have offered similar arguments based on fair-play (Simmons 1979) or our civic responsibilities of mutual assurance to other citizens (Duff 2018: 319–320, 324–332). The idea is that it amounts to not playing fair, or a breach of one’s civic responsibilities, to violate the duly enacted rules of legitimate and beneficial regulatory frameworks. Third, to extend an argument by Simester (2021: 240–242), perhaps we must be able to publicly defend our 516

Legitimate divergence between moral and criminal blame

actions or else they can be deemed culpable wrongs.6 Suppose one cannot reasonably justify to the public that one will grant oneself an exception to the relevant licensing or disclosure requirement – perhaps because it would be too difficult to adjudicate similar claims to an exception as part of a general scheme. If so, this makes the prohibited conduct a culpable wrong, as it’s publicly indefensible. If some such argument succeeds, the offense at issue would not actually violate the culpable wrongdoing constraint. On closer inspection they would involve no injustice and individual citizens would have a moral obligation to obey – even when one knows one can safely do the prohibited conduct. Again, there is nothing to regret in such offenses. 2.b) Genuinely overbroad but justifiable. A related scenario involves offenses that genuinely are overbroad – where no strategy like the above succeeds – but the overbreadth is justifiable. Perhaps violating the culpable wrongdoing constraint is justifiable on necessity grounds as the only way to prevent greater harms. Or perhaps violating the constraint is justifiable on non-consequentialist grounds. The statutory rape example might be justified this way. If the law didn’t use a brightline age cutoff, but only criminalized the underlying wrong of exploitative sex, this would give a grant of power to law enforcement to broadly investigate our sexual relations to decide which sex is exploitative. Perhaps this intolerably interferes with our rights to privacy and autonomy (beyond the well-bring reduction it might bring). Perhaps the effect on complainants, putting them on trial in effect, is similarly intolerable. Such considerations might help justify using a brightline age cutoff in the offense definition even though this criminalizes some non-culpable conduct. Categories 2.a) and 2.b) differ mainly in that the prohibited conduct is culpable in the former but not the latter. For both categories, it’s all-things-considered permissible for the legislature to create the relevant offense. But while there is an obligation to obey the 2.a) prohibitions (generating moral culpability if breached), it’s doubtful this holds for 2.b) offenses. If there were such an obligation, it would be a 2.a) offense. Thus, while 2.b) offenses are justifiably created and enforced, one isn’t necessarily morally culpable for doing them (say, if one knows one can safely do the conduct). How regrettable are 2.b) offenses? One might think there is no cause for regret here because the creation of the offense is all-things-considered justified and thus at bottom permissible. Harry Truman reportedly believed this about his decision to use the atom bomb, repeatedly asserting he “never lost any sleep over that decision” (Kuznick 2007). Nonetheless, this misses something normatively important. 2.b) offenses genuinely conflict with the culpable wrongdoing constraint, so they allow undeserved convictions and punishments. There does seem to be an important respect in which the offense has tragic effects that are to be regretted. The enactment of such offenses remains a pro tanto wrong, even if still all-things-considered permissible. To capture this, let’s distinguish two types of regret: an act might be regrettable in one respect or regrettable all-things-considered. Acts of criminalization in the 2.b) category are regrettable in the former sense (as there is indeed something deeply unfortunate about them), but not in the latter sense because they are the best the legislature could do.7 2.c) Genuinely overbroad but not justifiable. Finally, there are cases where the overbroad offense really does violate the culpable wrongdoing constraint (like in 2.b) but there is no sufficient justification for it. Neither principles of political morality (such as our right to minimum levels of privacy and autonomy) nor the good consequences of creating the offense suffice to justify the fact that the offense sweeps in some conduct that is not culpably wrong. Accordingly, it is simply wrong for the legislature to create such offenses. These are 517

Alexander Sarch

immoral laws – clearly regrettable. There is no moral obligation to obey them, meaning that one can do the prohibited conduct without necessarily incurring culpability.

II  Core cases of legitimate divergence Given this analytical framework, let’s examine some paradigmatic cases where criminal law may legitimately diverge from morality. I don’t endorse or evaluate the moral views these examples are based on. They are merely to illustrate. First, set aside one type of divergence involving under-inclusiveness relative to morality. Many legal moralists, like Duff, think criminal law should only be concerned with moral wrongs that are properly a matter of public concern, not private immorality (Duff 2018: 75–80). This is one way in which criminal law might fail to fully mirror morality, but a non-problematic one. Restricting criminal liability so it’s narrower than the full sweep of morally blameworthy conduct entails no excessive punishment or conflict with the culpable wrongdoing constraint.8 Here, then, are three broad categories of divergence that raise the specter of over-inclusiveness. They may overlap, and the list is non-exhaustive. a) Simplifications of the moral rule. One type of divergence involves simplifying a complex moral rule. These tend to be proxy crimes, which combat a target wrong with complex or mushy contours using a clearer proxy for it (Lee 2022). Some might involve a brightline cut-off within a degreed or vague moral property. For example, “having sex with someone below [a specified age]” is a simple proxy for the culpable wrong of exploitative sex. Likewise, the failure to report cash transactions over $10,000 is a proxy for facilitating money laundering (31 U.S.C. § 5331). But not all proxy crimes involve cut-offs. The prohibition on “wash trades” (7 U.S.C. § 6c(a)(2)(A)(i)) – buy-orders for derivatives that one immediately cancels (perhaps giving the false impression of high trading volume) – is a proxy for manipulative trading, though no brightline cut-off is used. Proxy offenses will genuinely conflict with the culpable wrongdoing constraint if no deeper strategy can show that the prohibited conduct really is a culpable wrong.9 Even if the conflict proves genuine, however, this might still be justified by countervailing benefits like clearer guidance for citizens, greater protection of privacy, or harm reduction through greater deterrence. If so, this would put the offense in category 2.b) from before. b) Declining to recognize moral justifications or excuses. Legitimate divergence may arise where the law has reason not to recognize all the justifications or excuses that morality does. Take the offense of disposing of hazardous waste without a license. Morally, if you know you can safely dispose of the waste without any risk, you plausibly wouldn’t face moral blame for doing so without the license. The law, however, may have good reasons not to recognize an affirmative defense or other exception in such cases. Section III explores some reasons why, like the risk of over-estimating one’s competence and engaging in selfserving rationalization. Similarly, consider Judge Bazelon’s proposed “rotten social background” defense (Morse 1976: 1252). Morality plausibly would excuse one (perhaps partially) for certain crimes (like drug or property offenses) if one faced extremely difficult social and economic conditions. After all, deprivation may have neurological effects like reducing impulse-control, not to mention unfairly limiting one’s lawful opportunities for economic advancement. Even if this morally excuses, legislatures did not take up Bazelon’s suggestion  – perhaps legitimately. Beyond the worries scholars raised about whether this would unduly weaken 518

Legitimate divergence between moral and criminal blame

the deterrent effect of the law and fail to see individual defendants as having full agency (Morse 1976: 1267), the legal excuse might also prove difficult to apply – especially given the sweeping look at one’s upbringing it would require. Another example concerns the principle that ignorance of law does not exculpate. Theorists like Husak argue that normative ignorance can morally excuse (2016: 150–156). Common law jurisdictions don’t conform to Husak’s exhortation, instead presuming citizens know the law – or should. This seems sensible for the worst wrongdoing: mala in se behaviors like murder and rape whose wrong-making features will be blindingly obvious to competent adults. We can see why ignorance of the law needn’t exculpate here. Still, it’s plausible that it should when the ignorance is reasonable, as in technical regulatory areas or when traveling in foreign jurisdictions with surprisingly different rules. Still, the law typically does not let even reasonable ignorance exculpate. This may not be simple injustice. There arguably are practical benefits (2.b-style outweighing considerations) that help justify not allowing ignorance of law, even when reasonable, to exculpate. Standard justifications include increasing the deterrent effect of the law and boosting judicial efficiency by not requiring prosecutors to prove knowledge of law. Furthermore, the alternative may prove unworkable: a “reasonable ignorance of law” defense likely would be opportunistically invoked by many defendants. The question remains open. Personally, I suspect a narrow “reasonable ignorance of law” defense for non-malum in se offenses could be a workable way to reduce injustice. But it’s not an unreasonable debate. There are other cases where the law might legitimately decline to recognize certain moral justifications or excuses. The legal conception of insanity plausibly is narrower than the analogous moral excuse (Litton 2018). Moreover, for hard incompatibilists who think we lack free will and moral responsibility, criminal liability in general might be seen as an instance of the law declining to recognize a particularly wide-ranging moral excuse – perhaps legitimately. What one says about such prima facie overbreadth depends on whether the legally prohibited conduct turns out, on closer inspection, to be culpably wrong (Category 2.a)), or if not, whether there are sufficient justifying reasons for the law’s divergence from morality that outweigh the injustice of violating the culpable wrongdoing constraint (Category 2.b)). If not, the overinclusive offense just is an immoral law in need of reform (Category 2.c)). c) Ignoring morally relevant factors in the offense definition. The legislature may have reasons to omit certain complex, morally relevant factors from the definition of the corresponding legal offense. Consider the tendency for criminal offenses to be defined only using a few narrow mental states (paradigmatically purpose, knowledge, recklessness and negligence), but beyond this to ignore motives – i.e., what one hoped to accomplish by doing the offense (LaFave 2003: § 5.3).10 Of course, motives affect sentencing (Hessick 2006: 90) and are elements of isolated offenses like hate crimes (18 U.S.C. § 249). But generally, criminal law is unconcerned with motives in offense definitions. Perhaps this makes criminal liability overinclusive relative to morality, as when a poor mother is technically guilty of theft when she steals to feed her children (assuming she lacks a necessity defense because the harm she acts to prevent is not imminent). Even if her sympathetic motives matter at sentencing, she still can be convicted even though she plausibly is not morally blameworthy for her conduct. Some argue the general irrelevance of motive to criminal liability is justified for practical reasons (including the need for simple and uniform rules).11 Section III proposes more principled explanations as well, including reducing privacy invasions by law enforcement.12 519

Alexander Sarch

Among other possible examples,13 strict liability offenses arguably also fit here (depending on the details). Such offenses impose liability for causing specified harms, perhaps when engaged in specified activities, regardless of one’s mental state or fault as to the harm. One can be convicted even if one could not have known of the risk of harm, took all possible precautions and was not at fault for causing it. Such offenses disregard complex moral factors  – namely, the typically important requirement of mens rea (“guilty mind”)  – in defining offenses. Defenders of strict liability offenses appeal to practical considerations like efficiency and increased deterrence (Hamdani 2007). However, as an offense type, they unavoidably conflict with the culpable wrongdoing constraint.14 Thus, the question is whether they belong in category 2.b) due to the justifications that might be adduced for them or 2.c) if they simply are normatively unjustifiable.15

III  Grounds for divergence Whether apparent criminal law overbreadth is justifiable (2.a) or 2.b)) or simply immoral (2.c)) depends on the considerations that support divergence from morality. However, the role of practical and institutional design considerations in determining the proper content of criminal law is undertheorized. Perhaps this is because it’s assumed that such considerations matter only through consequentialist reasoning applied on a case-by-case basis. But this is too simple. Divergences from morality needn’t only be one-off practical compromises grounded in a highly changeable consequentialist calculus, but could also be supported by stable principles applicable across time and place. Thus, rather than leaving the institutionality of criminal law as an afterthought, legislators and theorists would benefit from mid-level principles to be guided by. I propose some such principles here. Crucially, not all of them directly involve or are based on the expectation of good consequences. I can’t defend these principles here, nor is my list exhaustive. The aim is to highlight some principled ways we should want and expect criminal liability rules to diverge16 from morality and then we can evaluate whether the divergences these principles ground are ultimately defensible.17 I formulate them as generalizations with an implied ceteris paribus clause. For each principle, it’s not clear it would similarly affect the content of morality. These principles concern the institutional setting in which criminal law norms are applied, but moral norms are not wielded this way and so need not be similarly constrained by institutional or implementation considerations. To start, because it’s been widely discussed elsewhere (Duff 2018: 75–80; Husak 2011: 83), I’ll just briefly mention the following: Principle I: Criminal law rules should focus on public wrongdoing not private wrongdoing. This principle serves to limit the content of criminal law relative to morality but this does not produce overbreadth. Principle II: Criminal law rules (offenses and defenses) should not be unduly complex or employ overly fine-grained distinctions; they should instead attain appropriate levels of simplicity and generality.

520

Legitimate divergence between moral and criminal blame

This principle is supported by deeper considerations, which draw out the benefits of simplicity: Principle II.a: Criminal law rules should be capable of providing effective guidance to most citizens, not just those with special knowledge or skill. This principle, which lends support to the simplicity principle, applies because rules that effectively guide conduct tend to produce better compliance – whether through deterrence or voluntary compliance by those who think the law worthy of adherence. Principle II.b: Criminal law rules should promote, and reasonably be seen to promote, uniformity in how cases are treated (not merely in practice but in the law on the books). It is important not only for the criminal law to treat like cases alike, but also for this uniformity to be reasonably visible to the public (assuming it’s passably informed and engaging in good faith). For one, this promotes the perception of legitimacy by avoiding the (reasonable) appearance of arbitrariness. Perceived legitimacy can, of course, boost compliance (Robinson 2013: 161–162). Moreover, broad uniformity can give the law greater predictability, which is itself important given that we can assume the law will not be known in perfect detail by all. Crafting laws whose implications can be extrapolated to novel scenarios not only by trained lawyers but by most citizens is important for making the law predictable and uniform, so regular citizens can plan their affairs in light of it. In some cases, however, uniformity may mean the law treats like cases alike at a higher level of generality, even when more fine-grained distinctions might matter to the proper moral assessment. Principle III: Criminal law rules should be formulated to reasonably reduce (to the extent compatible with the other aims of the law) exceptions and the use of malleable concepts, which can be easily taken advantage of to self-servingly justify or rationalize away the risks or harms of one’s conduct. The idea is to avoid, where possible, rules and exceptions that are too easily twistable – ones that give citizens wide latitude to opportunistically rationalize their desired risky or bad actions. Suppose it’s an offense to dispose of hazardous waste without a license unless one knows one can do so safely in the circumstances. This qualification (either as a defense or part of the offense definition) carries risks given human psychology. We tend to over-estimate our own talents and underestimate our limitations (Dunning 2003). Moreover, we tend to self-servingly rationalize misconduct we wish to perform (Ashforth & Joshi 2005: 10–14). Thus, even if it’s not morally culpable to dispose of waste without a license if one knows one can do so safely, there are dangers in codifying this exception, as it can easily be used opportunistically. However, Principle III must not be taken too far. We would not want the fear of a few individuals taking advantage to prevent us from codifying important exceptions whose absence would produce far greater injustice. Striking this balance in particular cases is difficult but vital. Principle IV: Criminal law rules should protect and where feasible promote individual freedom.

521

Alexander Sarch

The importance of “maximize[ing] individual freedom within the coercive framework of law” (Hart 2008: 23) is familiar, though the same doesn’t obviously apply to moral norms (or not in the same way). Promoting freedom is important on consequentialist grounds as freedom tends to promote well-being but can also be justified on non-consequentialist grounds of respecting autonomy rights. Criminal law can promote individual freedom in numerous ways. First, clear guidance is needed for fair notice and a reasonable opportunity to avoid entanglement in harsh criminal law processes (Hart 2008: 23). Furthermore, every offense is a grant of power to law enforcement to investigate – perhaps invasively – related conduct that goes beyond the actions criminalized. After all, police investigations concern not only actions that do violate the relevant statute, but also nearby conduct (by the primary suspect or others) that may not be illegal. All this will have the effect of chilling legitimate conduct. Thus, criminalizing conduct even when we think it morally wrong and within the public sphere still carries costs to privacy and autonomy. Creating new offenses inevitably increases the risk of innocent citizens being (perhaps reasonably) investigated or prosecuted. Even when they can prove their innocence, the experience will surely be traumatic. Accordingly, free societies should criminalize only with reluctance and only for conduct widely agreed to be serious moral wrongs of proper public concern. Principle V: Criminal law rules should (where reasonably possible) seek to reduce social conflicts and division by providing normative clarity. The idea is to reduce conflict by giving content to broad moral principles, thus clarifying expectations and preempting quarrels arising from competing good faith understandings of the underlying moral norm. For example, suppose we’re in the cycle lane. You start to overtake me at a high speed. I feel threated and I use force to defend myself: I think you’re about to knock me over so I shout “don’t hit me!” and I push you first to protect myself. In fact, you’re an expert biker: you’re speedily but safely dashing around slower bikers. If you perceive my push to be unjustified – perhaps you think it not necessary, as you could have avoided harming me – you may conclude I’m the aggressor, thus prompting you to use defensive force in response and a fight ensues. If there is a lack of clarity about when and how much defensive force is permissible, escalatory misunderstandings are more likely. By contrast, if the criminal law sets a clear standard for the use of defensive force, it can help prevent escalatory miscalculations. Suppose the rule is that a mere honest belief in the necessity of force is required for self-defense. You may know my defensive force is not really necessary, given your expertise. But if you’re aware that the law requires only an honest belief in the necessity of force, and you recognize I believed I was in danger, then you’re less likely to think me an aggressor. Thus, even if you disagree with the legal standard, the fact of a clear public rule can help prevent escalation in such settings. Indeed, the law may have reason to set clear and simple standards in such cases even if this loses some of the nuance that may apply to attributions of moral blameworthiness.18 This can further support criminal law diverging from morality. (Other principles might also be added to this list.19)

IV Applications Revisit our earlier examples of overbreadth. I  suspect some really do involve culpable conduct, thus falling in 2.a). Plausibly some proxy crimes  – like failing to report cash 522

Legitimate divergence between moral and criminal blame

transactions over $10,000 and disposing of hazardous waste without a license  – simply give content to pre-existing moral obligations to give reasonable assurances to fellow citizens. These laws specify how we are to officially assure others we’re not facilitating money laundering or handling hazardous material unsafely. Breaching the duty to give the required assurances is a culpable wrong. Even if this argument fails, though, these proxy offenses could be supported by Principles II (simplicity), II.a (guidance), II.b (uniformity), III (no twistable exceptions) and perhaps IV (freedom). Other overbroad offenses fit better within 2.b). These include the brightline age cutoff in defining statutory rape, declining to recognize a “rotten social background” defense, not letting ignorance of law exculpate and the irrelevance of motives for criminal liability (though this might vary between offenses). These features of the criminal law genuinely conflict with the culpable wrongdoing constraint but may well be all-things-considered justified. These doctrines derive support from Principles II (simplicity), II.a (guidance), II.b (uniformity) and IV (freedom). Still, I don’t put these offenses in category 2.a) because the conflict with the culpable wrongdoing constraint seems to persist: genuinely non-culpable conduct is criminalized. Thus, these offenses seem pro tanto regrettable, if not all-thingsconsidered regrettable, as would entail a call to do better. Finally, some offenses fall within 2.c), especially strict liability offenses. Even if there might be a consequentialist case for some, I’d maintain in general they conflict too starkly with the culpable wrongdoing constraint and have too many available alternatives (adding mens rea, carving out defenses for adequate precautions or using non-criminal regulation) to think that strict liability offenses generally are all-things-considered justified. Perhaps they are justifiably directed at actors who are very well informed and sensitive to cost-benefit analysis in their conduct. So I accept that strict liability may be all-things-considered justified in places – if one could identify the relevant actors clearly (perhaps just corporations). But generally I suspect strict liability offenses are too morally problematic to be legitimate. Here the basis for regret is palpable. Even if one disagrees with these conclusions, I hope the evaluative framework from Section I and the justifying principles from Section III help provide legislators and theorists with better tools for assessing legislative acts of criminalization. Moreover, I hope to have dispelled the idea that the criminal law’s divergences from morality always sacrifice our principles to the messy necessities of practice. Criminal law’s divergence from morality can sometimes itself be a matter of principle.

Notes 1 Husak, for instance, endorses “a presumption that the criminal law should . . . conform to, or mirror[,] critical morality” (2016: 34). 2 As Simester argues, since convictions make “a public, denunciatory statement about the defendant,” “it is imperative that convictions . . . be deserved[,] that [they] speak the truth” (2021: 4). 3 The justifications for criminalization and punishment can come apart. There can be benefits from creating an offense even if no one is punished under it. But set that aside here. Normally, the justifications for criminalization and punishment dovetail. The same types of justification generally apply to both, including harm-reduction through deterrence, expressive benefits and retributive benefits. 4 As Simester observes, “[w]hat the law criminalizes, and what the criminal law condemns, is not wrongdoing per se; but culpable wrongdoing” (2021: 32). 5 Cornford 2017 defends the defeasible version. Duff 2018 replies. 6 Scanlon (1998: 153) defends a contractualist account of wrongness along similar lines.

523

Alexander Sarch 7 It’s also possible (if unlikely) for there to be offenses whose overbreadth is not justified, but the legislature is excused in enacting it. Maybe this is due to exigent practical circumstances (e.g., unfair budget limits or legacy technology) that create intense pressure on the legislature that cannot reasonably be resisted, but the benefits of the overbreadth are insufficient to fully justify it. (Crucially, merely the threat of being voted out of office would not suffice for such an excuse.) Thus, it would remain all-things-considered wrong (and regrettable) to create the offense, but legislators would not be blameworthy due to their excuse. 8 But underinclusive offenses can face other objections like arbitrariness  – say, if fraud were not illegal on Thursdays. 9 This assumes no legislative fix is possible. Another way to avoid overbreadth is to alter the offense definition or add an affirmative defense so only culpable conduct is subject to criminal liability. The discussion assumes such amendments are not feasible. 10 “[M]otive [as distinct from] the ‘specific intent’ requirements of some crimes, is not relevant on the substantive side of the criminal law” (LaFave 2003: § 5.3). 11 Martin Gardner argues the law shouldn’t require proof of “evil motives” because they “are difficult to evaluate,” and it “would require trial courts to consider detailed case histories of each defendant,” which often are “unavailable to prosecutors at the time necessary to charge the crime” (1993: 685–686). 12 For another such argument, see Sarch 2019: 75–79. 13 Another might be the lack of an “unjustified” element in the definition of criminal law knowledge, although such an element is included in the definition of recklessness. Model Penal Code § 2.02(b)-(c). 14 At least, this is so for substantive not merely formal strict liability (Simester 2021: 291, 312). 15 Negligence liability might be another example if one thinks acting in unreasonable unawareness of the risks of one’s conduct is not morally culpable (unless traceable to prior culpable acts or decisions) (Alexander & Ferzan 2009: ch. 3). Others view negligence as independently culpable (Stark 2016: 243–252). They won’t see negligence liability as involving significant criminal law overbreadth. 16 These principles can affect the legally recognized reasons we must not manifest insufficient regard for. 17 Some conflicts between law and morality could be resolved through discretion or restraint in enforcement, thus making criminal law more justifiable in practice. (Thanks to Antony Duff for this point.) Still, my concern here is criminal law “on the books.” If that is defective, communicative problems, plus greater risks of unjustified outcomes, would persist. 18 For excellent discussion of similar themes, see Stone 2023. 19 Yaffe argues the law should charitably presume we’re rational (2018: 27). Perhaps law might also presume our competence to recognize reasonably available information, like the risks of our conduct or the law’s requirements (Sarch 2021: 322–327). Perhaps this explains why criminal law imposes negligence liability and does not treat ignorance of law as exculpatory, even if morality disagrees.

Further reading For discussion of how institutional design constraints and non-desert-focused values like privacy and autonomy can legitimately impact the definition of criminal law offenses and defenses, see Cornford, A. (2017). “Rethinking the Wrongness Constraint on Criminalisation,” Law and Philosophy, 36, 615–649; Lee, Y. (2022). “Proxy Crimes and Overcriminalization,” Criminal Law and Philosophy, 16, 469–484; Sarch, A. (2021). “Should Criminal Law Mirror Moral Blameworthiness or Criminal Culpability? A  Reply to Husak,” Law  & Philosophy, 41, 305 and Chapter  3 of Simester, A. (2021). Fundamentals of Criminal Law. Oxford: Oxford University Press. For a terrific exploration of similar themes in the private law context, see Stone, R. (2023). “Private Law Without Wrongdoing,” University of Toronto Law Journal, 73, 53–87. For an excellent and comprehensive discussion of the issues arising in this chapter, see generally Duff, A. (2018). The Realm of Criminal Law. Oxford: Oxford University Press.

524

Legitimate divergence between moral and criminal blame

References Alexander, L., & Ferzan, K. (2009). Crime and Culpability. New York: Cambridge University Press. Ashforth, A., & Joshi, B. (2005). “Business as Usual: The Acceptance and Perpetuation of Corruption in Organizations,” Academy of Management Executive, 19(9). Berman, M. (forthcoming). “Proportionality, Constraint, and Culpability,” Criminal Law  & Philosophy. Cornford, A. (2017). “Rethinking the Wrongness Constraint on Criminalisation,” Law and Philosophy, 36, 615–649. Duff, R. A. (2014). “Towards a Modest Legal Moralism,” Criminal Law & Philosophy, 8, 217. ———. (2018). The Realm of Criminal Law. Oxford: Oxford University Press. Dunning, D. et al. (2003). “Why People Fail to Recognize Their Own Incompetence,” Current Directions in Psychological Science, 12, 83–87. Edwards, J., & Simester, A. (2019). “Crime, Blameworthiness, and Outcomes,” Oxford Journal of Legal Studies, 39, 50. Gardner, M. (1993). “The Mens Rea Enigma: Observations on the Role of Motive in the Criminal Law Past and Present,” Utah Law Review, 635. Hamdani, A. (2007). “Mens Rea and the Cost of Ignorance,” Virginia Law Review, 93, 415. Hart, H. L. A. (2008). Punishment and Responsibility (2nd ed.). Oxford: Oxford University Press. Hessick, C. B. (2006). “Motive’s Role in Criminal Punishment,” Southern California Law Review, 80, 89. Husak, D. (2011). Overcriminalization. New York: Oxford University Press. ———. (2016). Ignorance of Law. New York: Oxford University Press. Kuznick, P. (2007). “The Decision to Risk the Future,” Asia-Pacific Journal, 5, Article ID 2479. LaFave, W. (2003). Substantive Criminal Law, vol. 1, § 5.3 (2nd ed.). WestLaw (Thomson Reuters). Lee, Y. (2022). “Proxy Crimes and Overcriminalization,” Criminal Law and Philosophy, 16, 469–484. Litton, P. (2018). “Traumatic Brain Injury and a Divergence Between Moral and Criminal Responsibility,” Duquesne Law Review, 56, 35. Morse, S. (1976). “The Twilight of Welfare Criminology: A Reply to Judge Bazelon,” Southern California Law Review, 49, 1247. Rawls, J. (1971). A Theory of Justice. Cambridge: Belknap Press and Harvard University Press. Robinson, P. (2013). Intuitions of Justice and the Utility of Desert. New York: Oxford University Press. Sarch, A. (2019). Criminally Ignorant: Why the Law Pretends We Know What We Don’t. New York: Oxford University Press. ———. (2021). “Should Criminal Law Mirror Moral Blameworthiness or Criminal Culpability? A Reply to Husak,” Law & Philosophy, 41, 305. Scanlon, T. M. (1998). What We Owe to Each Other. Cambridge: Belknap Press and Harvard University Press. Simester, A. P. (2021). Fundamentals of Criminal Law. Oxford: Oxford University Press. Simmons, A. J. (1979). “The Principle of Fair Play,” Philosophy & Public Affairs, 8, 307. Stark, F. (2016). Culpable Carelessness. Cambridge: Cambridge University Press. Stone, R. (2023). “Private Law Without Wrongdoing,” University of Toronto Law Journal, 73, 53–87. Tadros, V. (2005). Criminal Responsibility. Oxford: Oxford University Press. Yaffe, G. (2011). Attempts. New York: Oxford University Press. ———. (2018). “The Point of Mens Rea: The Case of Willful Ignorance,” Criminal Law & Philosophy, 12, 27.

525

INDEX

absolute responsibility 110 – 111, 114 – 115; forced to be free 118 – 120; morality and the “empirical world” 112 – 113; morality and the “intelligible world” 111 – 112; predispositions to goodness 115 – 118; responsibility for the whole world 113 – 114; transcendental freedom regained 120 – 121; see also responsibility accountability 254 – 255 action, eliminative 491 – 492, 495 – 497 action, fatedness of: Chrysippus 53 – 54 action, voluntary 472 – 474; Aristotle 100 – 101; see also Voluntary Act Requirement action from forethought: Aristotle 100 – 101 actual awareness 278 – 279 African tradition 380 – 381; an interpretation of reconciliation 381 – 383; reconciliatory civil justice 385 – 388; reconciliatory criminal justice 383 – 385; reconciliatory trials 388 – 390 agency 153 – 161, 290 – 296 agent/agents 427 – 428; Stoics 51 – 52 agential power 397 – 400 anger: Aquinas 138 – 142; Stoics 52 – 54 answerability: dialogical 211 – 215; and the reflection of judgment 204 – 208; strict answerability in the law 208 – 211 apology 346 – 354 appearance: Epictetus 52 – 53 Aquinas 137 – 145 Aristotle 104 – 108; arguments about forethought 102 – 104; conditions for forethought 101 – 102; legal responsibility 97 – 99; voluntary action

and action from forethought 100 – 101; voluntary and involuntary 99 – 100 artificial intelligence (AI) 395 – 396; agential power of 397 – 400; governance as social care 402 – 405; responsibility as social care 400 – 402 assent: Epictetus 52 – 53; Stoics 51 – 53 attributability 254 – 255 attributivism 371 – 372 autism 290 – 294 autonomy 211 – 215 awareness 278 – 279 bail 498 – 499 banks see international banks belief in a just world 374 – 377 bipolar disorder 290 – 294 blame/blaming 192 – 194, 315 – 323; Aquinas 137 – 145; legitimate divergence between moral and criminal blame 514 – 523; the standing to blame 337 – 343; without blame 368 – 377 blameworthiness 277 – 281 Bramhall 57 – 66 business ethics 421 – 423 capacitarianism 280 – 281 capacity 299 – 306 care see social care causation 164 – 173; Hobbes 60 – 61; Hume 68 – 70; Stoics 46 – 48 chance: Hume 72 – 73 character: Chrysippus 53 – 54 Chyrsippus 53 – 54 civil justice 385 – 388

526

Index climate 431 – 439 coercion 261 – 270 communication 326 – 333 compatibilism 230 – 238 compensation 380 – 381; an interpretation of reconciliation 381 – 383; reconciliatory civil justice 385 – 388; reconciliatory criminal justice 383 – 385; reconciliatory trials 388 – 390 competence see moral competence complexity 501 – 503 complicity see reckless complicity Confucian thought: modeling and teaching from others 129 – 132, 134 – 135; responsibility and control 133 – 134; and the responsibility relation 126 – 127; responsibility to self 132 – 133; responsibility as shared 127 – 129; and the welfare of others 125 – 126 contempt 198 – 200 contingent causation: Hobbes 60 – 61; see also causation control 155 – 160, 194 – 195; Confucian thought 133 – 134 convention 299 – 306 conversation 346 – 354 corporate digital responsibility (CDR) 419 – 421, 428 – 429; agents and their areas of responsibility 427 – 428; business ethics as ethics of responsibility 421 – 423; concept of 424 – 425; digital ethics 425 – 426; normative requirements 427; practical implementation 428; as a response to digitalization 426; working definition 423 – 424 crime 447 – 449, 493 – 494 criminal blame 514 – 523; see also blame/blaming criminal justice 383 – 385 criminal law 515 – 516 criminal liability 483 – 485 criminal responsibility 485 – 488 culpable threat 494 – 495 cultural ignorance 412 – 415; see also ignorance dangerousness 492 – 497 deep self 177 – 178, 186 – 187; and ascriptions of responsibility 183 – 186; in defense of Deep Self Views 178 – 180; the variety of Deep Self Views 180 – 183 defensive harm, liability to 255 – 258 deliberation 239 – 247 derivative blameworthiness 277 – 281 detention see pretrial detention determinism: Hume 72 – 73; Sidgwick 89 – 92; Stoics 52 – 54

deterrence 447 – 449 dialogical answerability 211 – 215; see also answerability digital ethics 425 – 426 digitalization 426 digital responsibility see corporate digital responsibility direct blameworthiness 277 – 281 disappointment 198 – 200 divergence see legitimate divergence divine providence 45 – 54 doctrine of necessity: Hume 70 – 72; see also necessity domination 266 – 268 duties 449 – 451, 491 – 492 duty to appear 497 eliminative actions 491 – 492, 495 – 497 emotions 190 – 191, 198 – 200; and blame 192 – 194; fair and deserved emotions 194 – 195; fitting emotions 195 – 198 empirical world: Kant 112 – 114 Epictetus 52 – 53 epistemic condition 274 – 283 ethics 82 – 88; business ethics 421 – 423; digital ethics 425 – 426; libertarian and deterministic conceptions 89 – 92 exception 501 – 503 exculpation 157 – 160 excuse 157 – 160, 299 – 306 exemptions 157 – 160 expectation see reasonable expectation fate: Stoics 46 – 52 first-person narratives 290 – 294 flight 497 forethought: Aristotle 100 – 108 forfeitures 491 – 492 forgiveness 346 – 354 fossil fuels 432 – 438 fraternal correction: Aquinas 142 – 144 freedom 110 – 111, 114 – 115; forced to be free 118 – 120; morality and the “empirical world” 112 – 113; morality and the “intelligible world” 111 – 112; predispositions to goodness 115 – 118; responsibility for the whole world 113 – 114; transcendental freedom regained 120 – 121 free will 82 – 88; and compatibilism 230 – 238; Hobbes and Bramhall 57 – 66; Hume 68 – 79 free will skepticism 241 – 243; libertarian and deterministic conceptions 89 – 92; see also skepticism

527

Index gender 442 – 451 goodness: Kant 115 – 118 governance 395 – 398, 402 – 405 guilt 31 – 39 Hegel 31 – 39 history/historical context: AI governance 402 – 405; Hegel 34 – 35; historical injustice 407 – 415 Hobbes 57 – 66 holding responsible 198 – 200, 380 – 381; an interpretation of reconciliation 381 – 383; reconciliatory civil justice 385 – 388; reconciliatory criminal justice 383 – 385; reconciliatory trials 388 – 390; see also responsibility holism world view: Confucian thought 126 – 127 Hume: determinism and chance 72 – 73; doctrine of necessity 70 – 72; hypothetical liberty 73 – 75; liberty of spontaneity 75 – 76; moral responsibility 77 – 79; theory of causation 68 – 70; a traditional misreading 76 – 77; two-way liberty 77 hurt feelings 198 – 200; see also emotions hyper-agency 290 – 294 hypo-agency 290 – 294 hypothetical liberty: Hume 73 – 75; see also liberty ignorance 274 – 277, 412 – 415; Plato 20 – 22, 26 – 28 imputation model: Hobbes 60 – 61 incapacity 302 – 304 incentives 449 – 451 incompatibilism 219 – 227 indicatives 220 – 222 individual, the: Confucian thought 126 – 127 injustice, historical 407 – 415 innocent incompatibilism 222 – 227 intelligible world: Kant 111 – 112, 113 – 114 international banks 431 – 439 interpersonal argument for control 194 – 195 intrapersonal argument for control 195 involuntariness: Aristotle 99 – 100; Plato 22 – 23, 26 – 28 joint narrative 346 – 354 judgment 204 – 208 justice, civil 385 – 388 justice, criminal 383 – 385 justifications 157 – 160, 483 – 488, 491 – 492 just world, belief in 374 – 377 Kant 110 – 111; forced to be free 118 – 120; morality and the “empirical world” 112 – 113; morality and the “intelligible world”

111 – 112; predispositions to goodness 115 – 118; rescue attempts 114 – 115; responsibility for the whole world 113 – 114; transcendental freedom regained 120 – 121 knowledge: Plato 20 – 22 law 208 – 211, 457 – 459, 463 – 464, 507 – 511, 515 – 516 least exceptional exception 501 – 503 leeway 234 – 238 legal moralism 515 – 516 legal responsibility 97 – 99, 104 – 108, 457 – 464; arguments about forethought 102 – 104; conditions for forethought 101 – 102; voluntary action and action from forethought 100 – 101; voluntary and involuntary 99 – 100; see also responsibility legitimate divergence 514 – 523 lesser-evil justification 491 – 492 liability see liability to defensive harm; strict liability; vicarious liability liability to defensive harm 255 – 258 libertarian conceptions of punishment: Sidgwick 89 – 92 liberty: Hume 73 – 77 logical context: Hegel 32 – 34 manipulation 250 – 258 Mengzi 127 – 129 mental disorder 287 – 290, 297; hypo- and hyper-agency in first-person narratives 290 – 294; scaffolding of one’s own agency 294 – 296 mind, character of: Epictetus 52 – 53 modeling: Confucian thought 129 – 132 moral blame 514 – 523; see also blame/blaming moral competence 287 – 290, 297; hypo- and hyper-agency in first-person narratives 290 – 294; scaffolding of one’s own agency 294 – 296 moral ignorance 276 – 277; see also ignorance moralism, legal 515 – 516 morality 463 – 464; Kant 111 – 113 moral philosophy 459 – 462 moral responsibility 57 – 60, 104 – 108, 372 – 374, 457 – 464; arguments about forethought 102 – 104; conditions for forethought 101 – 102; the epistemic condition 274 – 283; for historical injustice 407 – 415; Hume 68 – 79; legal responsibility 97 – 99; Stoics 48 – 50; voluntary action and action from forethought 100 – 101; voluntary and involuntary 99 – 100; see also responsibility moral virtue: Confucian thought 129 – 132; see also virtue

528

Index narrative see first-person narratives; joint narrative necessity: Hume 70 – 72; Stoics 46 – 52 neuroscience 470 – 472 non-necessity: Stoics 50 – 51 normative power: Hobbes 64 – 66; see also power obstruction 497 – 498 offers, coercive 268 – 269; see also coercion organizational responsibility 503 – 504; see also responsibility others, responsibility for 501 – 503; philosophy and vicarious responsibility 504 – 507; third-party acts in private law 509 – 511; vicarious liability in law 507 – 509; vicarious liability and organizational responsibility 503 – 504; vicarious liability in philosophy 508 – 509; see also responsibility others, welfare of: Confucian thought 125 – 126 partial re-scaffolding 294 – 296 Philosophy of Right (Hegel) 35 – 38 physical coercion 262; see also coercion Plato 19 – 28 possibility: Stoics 50 – 51 power 57 – 60; Hobbes 64 – 66 power, agential 397 – 402 power to refuse 264 – 266 practical deliberation 241 – 243; see also deliberation practical reasoning 253 – 254 praise 326 – 333, 342 – 343 pretrial detention 490 – 499 private law 509 – 511; see also law providence see divine providence punishment 380 – 381; an interpretation of reconciliation 381 – 383; reconciliatory civil justice 385 – 388; reconciliatory criminal justice 383 – 385; reconciliatory trials 388 – 390; Sidgwick 89 – 92; Stoics 52 – 54 quality of will 277, 281 reasonable expectation (RE) 276 – 281 reasoning see practical reasoning reckless complicity 431 – 439 reconciliation 380 – 381; an interpretation of 381 – 383; reconciliatory civil justice 385 – 388; reconciliatory criminal justice 383 – 385; reconciliatory trials 388 – 390 refusal see power to refuse re-scaffolding 294 – 296 resistance 269 – 270 responsibility: absolute 110 – 121; African tradition 380 – 390; and agency

153 – 161; and artificial intelligence 395 – 405; without blame 368 – 377; and causation 164 – 173; and coercion 261 – 270; Confucian thought 125 – 135; corporate digital responsibility 419 – 429; and the deep self 177 – 187; and the emotions 190 – 200; and gender 442 – 451; Hegel 35 – 38; legal 97 – 108, 457 – 464; and manipulation 250 – 258; moral 48 – 50, 57 – 60, 68 – 79, 97 – 108, 274 – 283, 372 – 374, 407 – 415, 457 – 464; for others 501 – 511; Plato 26 – 28; and pretrial detention 490 – 499; Sidgwick 89 – 92; Stoics 45 – 54; strict responsibility 480 – 488; taking responsibility 357 – 365 retribution 447 – 449 rights 491 – 492 role-responsibility 368 – 369; see also responsibility scaffolding 294 – 296 schizophrenia 290 – 294 self, deep 177 – 178, 186 – 187; and ascriptions of responsibility 183 – 186; in defense of Deep Self Views 178 – 180; the variety of Deep Self Views 180 – 183 self, responsibility to: Confucian thought 132 – 133; see also responsibility self-determination: Hobbes 61 – 63 self-harm 444 – 447 shame 198 – 200 shortgevity 444 – 447 Sidgwick 82 – 88; libertarian and deterministic conceptions 89 – 92 skepticism 239 – 247, 341 – 342 social care 395 – 396; AI governance as 402 – 405; responsibility as 400 – 402 social environment: Confucian thought 126 – 127 social implications of digitalization 426 source compatibilism 234 – 238; see also compatibilism spontaneity, liberty of: Hume 75 – 76; see also liberty standing to blame 337 – 343; see also blame/blaming standing to praise 342 – 343 Stoics 45 – 54 Strawson 190 – 192, 197 – 200 strict liability 480 – 488 strict responsibility 480 – 488 take-charge responsibility 372; see also responsibility taking responsibility 357 – 365, 368 – 371; see also responsibility teaching: Confucian thought 129 – 132 third-party acts 509 – 511 threat, coercive 264; see also coercion

529

Index threat, culpable 494 – 495 tracing strategy 278 – 279 transaction 346 – 354 transcendental freedom 110 – 111, 114 – 115, 120 – 121; forced to be free 118 – 120; morality and the “empirical world” 112 – 113; morality and the “intelligible world” 111 – 112; predispositions to goodness 115 – 118; responsibility for the whole world 113 – 114; see also freedom trials 380 – 381, 385 – 386; reconciliatory 388 – 390 un-scaffolding 294 – 296 vengeful anger: Aquinas 138 – 142; see also anger vicarious liability 503 – 504, 507 – 509

vicarious responsibility 504 – 507; see also responsibility vice 19 – 28 virtue 211 – 215; Confucian thought 129 – 132; Plato 20 – 22 volitional coercion 262 – 264; see also coercion voluntariness: Aristotle 99 – 101 Voluntary Act Requirement (VAR) 468 – 477 welfare of others: Confucian thought 125 – 126 whole greater than the sum of its parts 498 will see free will; quality of will wrongdoing: Plato 22 – 23 Xunzi 129

530